who goes there: friend or foe? - unt digital library/67531/metadc... · who goes there: friend or...

Who Goes There: Friend or Foe?

June 1993

OTA-ISC-537NTIS order #PB93-218915

GPO stock #052-003-01328-2

-—

Recommended Citation:U.S. Congress, Office of lkchnology Assessment, Who Goes There: Friend or Foe?,OTA-ISC-537 (Washington, DC: U.S. Government Printing Office, June 1993).

TForeword

he recent Persian Gulf War focused attention on the problem of fratricide, or“friendly fire” casualties among combat units. Twenty-four percent of U.S.combat fatalities in that conflict were due to friendly fire. The House ArmedServices Committee requested that OTA assess the technology and techniques

available to reduce this tragic loss of life. Although friendly fire has often been dismissedin the military literature as an insignificant contributor to overall combat losses, in thosefew historical cases for which good data are available, estimated friendly fire losses haveaccounted for at least 10 percent of total losses. Friendly fire has been an important-andunder-appreciated-source of combat deaths.

Combat is confusing, complex, and deadly. Friendly fire casualties can probablynever be eliminated, but several measures can reduce them. Application of new and existingtechnology can make identification of friendly forces on the battlefield more reliable;improved communication can reduce confusion on the battlefield; and better training canhelp military personnel make crucial, rapid decisions under the extreme stress of combat.

Congress faces several decisions related to reducing friendly fire. These include:

the choice of best technical approaches to pursue;the allocation of resources between systems that are devoted exclusively toreducing friendly fire and other systems—for example, better navigation andcommunication devices--that may reduce friendly fire in indirect and lessvisible ways;the best mix of near-term deployments and longer term research anddevelopment; andthe need for cooperation and coordination among the military Services and withallies.

We wish to express our appreciation to individuals in the many government agenciesand other organizations that provided information essential to the assessment and to themany reviewers who helped ensure its accuracy.

Roger C. Herdman, Director

lit

Preject Staff

IVAN OELRICHProject Director

Peter BlairAssistant Director, OTAEnergy, Materials, and International

Security

Lionel S. JohnsAssistant Director, OTA(through May 5, 1993)Energy, Materials, and International

Security

Alan ShawInternational Security and

Commerce Program Manager

ADMINISTRATIVE STAFF

Jacqueline Robinson BoykinOffice Administrator

Louise StaleyAdministrative Secretary

Madeline GrossContractor

CONTRACTORS

Elizabeth SheleyEditor

Richard R. MullerAir UniversityAir Command and

Staff College

John C. LonnquestDepartment of HistoryDuke University

iv

1

2

3

Introduction, Findings, Issuesf o r C o n g r e s s 1F i n d i n g s 2Issues for Congress 5

Historical Review of the Causes ofF r a t r i c i d e 7Types of Fratricide 9

Tactical Constraints Due to Fear of Fratricide 1 8The Prevalence of Fratricide 20

The National Training Center 2 4Fratricides During the Persian Gulf War 2 6

S u m m a r y 2 8

Avoiding Fratricide:General Considerat ions 29Trends in the Frequency of Fratricide 3 0Measures of the Severity of the Problem 3 1

The Attack Sequence 3 2

The Combatant’s Sources of Identification

I n f o r m a t i o n 3 3

Identification: Friend or Foe 3 8

Making Choices About How To Avoid Fratricide 45

C o n c l u s i o n 4 8

(Continued on next page)

contents

--

v

4 Avoiding Fratricide of Air andSea Targets 49Battle Management 49Noncooperative IFF 50Cooperative Question-and-Answer IFF Systems 56

Coordination With Civilian Air Traffic Control 62

Identification of Ships 66

A Comprehensive Identification System 67

5 Avoiding Fratricide of Land SurfaceTargets 69Increasing Knowledge of the TacticalEnvironment 70Identification Systems 72Dismounted Infantry 76Noncooperative Identification 77Current Programs 78

Tactics, Doctrine, and Training 79Conclusions 7 9

INDEX 81

— —

Introduction,Findings,Issues forCongress 1

A merica’s recent combat in the Persian Gulf War broughtnew attention to an old problem: fratricide, or ‘‘fi-iendlyfire, ’ that is, casualties from U.S. or allied weapons firedat U.S. or allied military personnel. Twenty-four percent

of all U.S. combat fatalities in the war were caused by friendlyfire. This figure seemed much higher than in previous wars andcaused a sudden focus on avoiding fratricide in future wars.

The U.S. military and the American public are becomingincreasingly sensitive to the human costs of military involve-ment, especially for contests of less than national survival. TheUnited States has invested much in energy and equipment to keepcasualties low. The high fraction of deaths in the Persian GulfWar due to fratricide was much higher than the nominal twopercent rate frequently cited in the military literature. Broad-based data on fratricide rates are not available; but, a recentreview of long-extant casualty surveys from World War II andthe Vietnam War shows that fratricide estimates of 2 percent areunrealistic and 15 to 20 percent may be the norm, not theexception. Thus, one reason that fratricide seemed unusuallyhigh in the Persian Gulf War is that total U.S. casualties were lowbut another important reason is that past rates of fratricide havebeen systematically and substantially underestimated. If theserates are, indeed, typical, then reducing casualties from fratricidedeserves the same kind of attention as reducing casualties fromany other major source.

Beyond numbers of killed and wounded, fratricide has acompounding effect on combat effectiveness. Weapons aimed atfriends are not aimed at the enemy. Friends killed by friends arenot able to fight the enemy. Moreover, the psychological effectsof friendly fire are always greater than from similar enemy fire.Combatants expect to be shot at by the enemy, but being shot at

1

2 I Who Goes There: Friend or Foe?

Table l-l—Detrimental Effects of Fratricide

●

●

●

●

●

●

●

●

●

●

Hesitation to conduct limited visibility operationsLoss of confidence in unit’s leadershipIncrease of leader self-doubtHesitation to use supporting combat systemsOversupervision of unitsLoss of initiativeLoss of aggressiveness during fire and maneuverDisrupted operationsNeedless loss of combat powerGeneral degradation of cohesion and morale

SOURCE: Center for Army Lessons Learned.

by friends corrodes cooperation and morale;shooting at friends also destroys morale and cancause commanders to be overly cautious incombat. See table 1-1.

The Persian Gulf experience has concentratedthe attention of the military Services on theproblems of fratricide. Several new antifratricideprograms were started within the Department ofDefense, and existing programs have been accel-erated, reoriented, or brought under new manage-ment control. New emphasis has been given to thefratricidal implications of other programs-suchas those to improve communication-not explic-itly or primarily intended to reduce fratricide.Doctrine and training are also being reexaminedwith a view to minimizing the risk of fratricide.

FINDINGSFratricide may be a significant source of

casualties. Twenty-four percent of U.S. combatfatalities in the Persian Gulf War were caused byfriendly fire. This seemed extraordinarily highcompared to past conflicts. Several reasons havebeenbeen

1.

2.

presented for why the fraction should haveso large:

total U.S. losses were very low, thus thepercentage of fratricides appeared abnor-mally high;the war was so short that U.S. troops did nothave a chance to gain much experience,reduce fratricide, and get the average down;

3.

4.

near-absolute dominance of the battlefieldby the U.S. meant that only U.S. roundswere flying through the air and if a soldiergot hit by anything, it was likely to be froma U.S. weapon; andthe unique characteristics of many U.S.weapons, for example, the depleted ura-nium in the M-1 tank round, made thefratricide that did occur undeniable.

Some of these reasons will apply to a range ofpossible future military engagements. Whether ornot fratricide in the Persian Gulf was particularlyhigh compared to past wars, it may be representa-tive of future conflicts.

The fourth point above deserves careful con-sideration. While military historians have fre-quently used two percent as a notional fratriciderate, the figure has been higher in all of thosecases for which good data are available. A recentreview of medical records from World War II, theKorean War, and the Vietnam War show fratri-cides to account for 12 percent and more of totalcasualties in those cases for which data areavailable. Perhaps the Persian Gulf War was notso unusual.

Reducing fratricide is desirable and feasi-ble, but eliminating it is not. Although programsto reduce fratricide are certainly needed, setting agoal of eliminating fratricide is unrealistic andprobably even counterproductive. Overly restric-tive rules of engagement, for example, may soreduce combat effectiveness that casualties in-flicted by the enemy increase more than friendlyfire losses are reduced. See figure 1-1.

The new global military environment re-quires a reevaluation of antifratricide efforts.Any technical goals established during the daysof the Cold War need to be reexamined. Pastdirections will not be completely reversed, butpriorities among technical direction may change.For example, during the NATO-Warsaw Pactconfrontation, which type of weapon was onwhich side was clear-cut. Today, noncooperativeidentification based on type of weapon or plat-

Chapter I-Introduction, Findings, Issues for Congress 3

Figure l-l—Total Casualties andAntifratricide Measures

[

—— —————

, ). /

\ //

/

A ,< Total losses

‘ \ \~.. ——–—I\

Increasing losses/

\\ from enemy fire > / ’

\ /’

‘,<Declining losses /from fratricide

\ /\ /

‘ \ / “. /

‘A/

Az /c1 — — — ’r -- -— ----——.-.—.

No antifratricide Too stringent anti-measure fratricide measures

>More stringent anti-fratricide measures

SOURCE: Office of Technology Assessment, 1993.

form will be more difficult in a world in whichenemies and allies may own the same hardware.Also, future alliances, like that in the PersianGulf, may be more ad hoc, making plannin g forand sharing of identification technology moredifficult. On the other hand, some tasks should beeasier to accomplish since few potential enemieswill be as sophisticated as the former SovietUnion or possess the sheer numbers.

Fratricide results from multiple causes. Friendlyfire is often thought of as due primarily, orexclusively, to misidentification. Investigation ofparticular cases usually reveals that the fratricidewas in fact the last link in a chain of mistakes.That is, faulty navigation; poor communication,command, and planning; lack of fire discipline;and occasional malfunctioning equipment areresponsible at least as often as misidentification.See figure 1-2.

With multiple links in a chain of causes, thereare multiple solutions to the problem of fratricide

by strengthening any of the links. For example,outfitting tanks with compasses will improvenavigation, helping units to be where they aresupposed to be and not to stray into fields offriendly fire. Improved radios allow the transmis-sion of more of this navigation information.Improved displays within the tank would allowclearer representation of that information. And, ofcourse, better sensors would allow better recogni-tion of fiends seen through the sights of the gun.Each of these measures could reduce fratricide.

Some approaches to reducing fratricide haveother benefits. For example, improved communi-cation and navigation allow better command andcontrol of combat units, more flexible tactics, andmore efficient allocation of combat resources. Allof this together could improve combat capabilitysignificantly while also reducing fratricide. Animproved identification device will mostly re-duce fratricide with much smaller side benefits.Thus, only the appropriate fractions of the costs ofsystems should be compared when consideringtheir relative efficiency in reducing fratricide.

Figure 1-2—Causes of Fratricide: Direct FireFratricide in World War II, Korea, and Vietnam

Inexperience190!0

/--\ Unknown factors

Target 1 00/0

m(sldentlflcahon260/o

b“

% =

SOURCE: U.S. Army

Coordination45%

Incidents by category58 total incidents


Wreckage of a U.S. armored vehicle destroyed byfriendly fire in the Persian Gulf War. The effectivenessof modern weapons makes identification mistakesdeadly.

No single technical approach to target identifi-cation will be perfect. Identification techniquescan be roughly divided between cooperative andnoncooperative approaches. In general, coopera-tive techniques can provide positive identifica-tion of friends. Failure to respond to a cooperativeidentification query could be assumed to identifya putative target as an enemy, but in mostcircumstances, most shooters would be hesitant touse the lack of response as a justification to fire.Cooperative techniques could, however, catego-rize targets as either friends or unknowns and thennoncooperative techniques could identify the foesamong the unknowns.

The technology for avoiding fratricide ofland surface targets lags behind the technologyimportant to avoiding aircraft fratricide. Avoid-ing fratricide requires good navigation, communica-tion, and identification. Yet multimillion-dollarU.S. tanks do not have compasses. Simple mag-netic compasses will not work inside a 60-tonsteel box and tanks are only now being fitted withradio-navigation equipment. Question-and-answer IFF systems, developed for aircraft a halfcentury ago in World War II, are only now beingdeveloped for land combat vehicles. Programs toreduce ground combat fratricide will need special

support for several years just to get up to whereaircraft systems are today.

Coordination among the U.S. military Serv-ices and among U.S. allies is essential. Programsto develop technology to reduce fratricide must becoordinated among Services and allies from thebeginnin g. The U.S. military emphasizes “com-bined arms” operations where the strengths ofmany different types of weapons are brought tobear simultaneously against an enemy. Thisapproach presents many opportunities for friendlyfire among aircraft, artillery, land vehicles, surface-to-air missiles, and so on. Fratricide reduction, asmuch as any other DoD effort, needs some centralcoordination, either from Office of the Secretary,the Joint Chiefs of Staff, or other special joint-Service organization. The Services realize theimportance of coordination, and their ongoingmulti-Service efforts should be encouraged andmonitored.

Future conflicts are likely to be allied opera-tions—as much for political as military reasons—and coordination of antifratricide technologydevelopment with allies must be maintained fromthe beginning. The utility in allied operations isone criterion by which prospective technologymust be judged. For example, if technology beingpursued for identification of friend and foe is sosensitive that it cannot be released to allies,especially ad hoc allies as we had in the PersianGulf War, then the usefulness of the technologywill be limited. This does not mean that theseapproaches are worthless, but the need for alliedcooperation should be included as a program goal.

Nonmaterial changes will also reduce fratri-cide. Some cases of friendly fire in the PersianGulf War could have been avoided by differentpre-war training. For example, since the war theArmy intentionally includes occasional strayfriendly vehicles in training exercises and maneu-vers to let soldiers practice ‘ ‘don’t shoot’ situa-tions. Simulators are an increasingly importantpart of training. In the past, these have not fullyreproduced opportunities for fratricide; this lackis now being addressed.

Chapter I–Introduction, Findings, Issues for Congress 5

The rules of engagement strongly affect thelikelihood of fratricide. The Services train warri-ors and train them to be aggressive, But in manypotential future conflicts, the conditions of thePersian Gulf War may recur: the U.S. was able tooutrange, outsee, and outgun the enemy by asubstantial margin. This capability could allow(but not demand) different rules of engagement.Many Persian Gulf fratricide occurred when atarget was engaged quickly yet the shooter was inno imminent danger and could have been moredeliberate. In situations where U.S. forces haveclear weapons performance superiority, moreconservative rules of engagement may reduceoverall U.S. casualties. This is a very complexissue that is without a simple answer (e.g., if theUnited States had fought more slowly battle-by-battle, perhaps the overall envelopment of Iraqiforces would have failed), however, it at leastdeserves consideration.

ISSUES FOR CONGRESS

I Allocation of ResourcesReducing fratricide will require new technol-

ogy and equipment. That, in turn, requires fund-ing, which then requires allocations within a finitedefense budget. There is, as always, competitionbetween efforts to reduce fratricide and othermilitary needs. Having combat superiority helpsto keep casualties down, so even if minimizingcasualties is the goal, spending less on offensiveweapons and more on avoiding fratricide is notautomatically the answer. Two issues, however,suggest that the relative emphasis on fratricideprevention should increase: first, most militaryanalysts interviewed by OTA for this assessmentagree that antifratricide efforts, especially relatedto land combat, have not received sufficientattention in the recent past. Second, the experi-ence of the Persian Gulf suggests that fratricidemay be a relatively greater cause of casualties infuture conflicts than has been appreciated in thepast.

Each of the Services has IFF and antifratricideprograms, and Congress will be asked to allocateresources among these. One of the findings of thisreport is that technology to help prevent fratricideof land surface targets is least developed andCongress may consider giving relatively greaterweight for a few years to programs supportingthese technologies.

Resources must also be allocated among vari-ous technical approaches to reducing fratricide.When comparing these costs, Congress may wantto consider the potential multiple benefits ofmany approaches. Specifically, an IFF systemwill reduce fratricide by improving identification,but has only limited additional applications,while improvements in communication and navi-gation can reduce fratricide and have compound-ing benefits to overall combat effectiveness.

1 Short-Term v. Long-Term GoalsAfter the Persian Gulf friendly fire losses, the

Services decided—with some prompting fromCongress-that an accelerated antifratricide pro-gram was needed. The Army developed a plan forboth “near-term” (less than 5 years) and “far-term" (7 or more years) solutions. The generaltechnical approach for the near-term solution isfairly well determined to be a millimeter wavequestion-and-answer system. This has the advan-tage of being available to troops in the field withinjust a few years, although it is not the ideallong-term solution. The degree of pressure fromCongress is part of the calculus by which theServices determine their allocation of effortbetween near- and long-term solutions. Congressmay wish to make clear to the Army the extent ofits urgency:

●

●

should the Army get a less-than-ideal systemin the field quickly so soldiers have some-thing in the event of a new Persian Gulf-likeconflict, orshould the Army take a longer-term ap-proach to get a better system while risking


that a conflict within 5 years or so may resultin too many friendly fire losses?

M Cross-Service CoordinationThere is probably no better example of an effort

where inter-Service coordination is needed thanthe development of antifratricide technology andequipment. The Services now have a GeneralOfficers Steering Committee that seems quitesuccessful in assuring coordination among vari-ous Service antifratricide programs. Congressmay want to pay special attention to Servicecoordination in future years to ensure that it is

maintained at every level. Past experience is notuniformly encouraging.

One technical aspect of Service coordination isthe compatibility of various IFF devices. Notevery weapon can effectively fire at every otherweapon; for example, fighter/interceptor aircraftand tanks cannot fire at each other. Is it reallynecessary that they be able to query each other’sIFF devices? Yet fighters can fire at ground-attack aircraft and ground attack aircraft can fireat tanks. If they do not all have the same IFFsystems, will ground-attack aircraft need twosystems operating in parallel?

———

Review of theCauses ofFratricide 2

n November 12, 1758, during the French and IndianWar, Colonel George Washington of the British Armyled a detachment of infantry to take a hill near LoyallHannon (now Loyalhanna), Pennsylvania occupied by

French soldiers and their Indian scouts. The French fled after abrief exchange of fire. But hearing the firing, a second detach-ment—under the command of Lt. Colonel George Mercer—approached the hill to assist. They arrived at dusk and the day wasfoggy, making visibility poor. Each side seems to have mistakenthe other for French and an intense fire fight broke out, killingbetween 13 and 40.1

While the current high interest in combat fratricide is a directresult of U.S. experience in the Persian Gulf War, this tale showsclearly that fratricide is not a new problem. During the Persian GulfWar, incidents of fratricide received considerable press attentionand caused international political friction.2 There was bewilder-ment among the public and press about how fratricide could occur.After all, shouldn’t it be obvious who are friends and who are foes?In addition, to some the losses from friendly fire seem less accept-able as an inevitable cost of war than are losses from enemy fire.3

1 The Papers of George Washingto~ Colonial Series, vol. 6, W.W. Abbo4 ed.(Charlottesville, VA: University Press of Virgini% 1988), pp. 120-123. A “detachment”could consist of between 200 and a thousand infantry.

2 Just a sample of the popular press stories about friendly fme includes: Bruce vanVoorst, “They Didn’t Have to Die, ” Time, Aug. 26, 1991, p. 20; Joshua Hammer,“Risking Friendly Fire, ” Newsweek, Mar. 4, 1991, p. 33; David HackWorth, “Killed byTheir Comrades, ” Newsweek, Nov. 18, 1991, pp. 45-46; Glem Frankel, “In Britain,Fallout from Friendly Fire, ” Washington Post, May 18, 1992, p. Dl; and David MOrnSOrL“Foes Not the Only Threat in the Gulf War,” Narional Journaf, Feb. 9, 1991, pp. 335-336.

3 Sensitive to the possibility of a different reaction to friendly fire losses, the MarineCorps readily admitted occurrences of friendly fwe but was reluctant to identi& preciselywhich deaths it caused. For example, a Marine Corps spokesman, Lt. Col. Ron Stokes,was reported as saying: “We don’t want to start painting guys with a differentbrush-these guys were killed by the enemy and these guys by friendly fire. They wereall killed in a combat action. If you start breaking it down, we’re not certain that it benefitseither the public or the families. See, ‘‘Kitled by Friend or Foe, It’s All the Same, ” TheNew York Times, Feb. 14, 1991, p. B18.

7


The fratricide of the Persian Gulf War wasunusual in some regards compared to that of pastwars. Most striking was the apparently unprece-dented high fraction of U.S. casualties resultingfrom fratricide; this was due in large part, ofcourse, to the extremely low U.S. casualtiesinflicted by the enemy.

In addition, the types of fratricide were differ-ent from other large mechanized land battles, suchas those of World War II. In World War II, themost deadly reported individual incidents offratricide were the result of bombing of friendlytroops by friendly aircraft. Surface-to-surfacefratricide resulted most often from indirect-fireweapons, that is, artillery fired at a target that thecrews could not see. The Persian Gulf War had anunusually high fraction of fratricides from direct-fire weapons—for example, tanks-shooting mis-takenly at other land targets, which they could seebut misidentified.

This chapter, a historical review of fratricide,shows how serious a problem fratricide has beenin past wars and reveals patterns in the occurrenceof fratricide in past wars that might suggestlessons for the future.

There are difficulties with a historical ap-proach. The movements of armies are usuallywell recorded, but the record of particular actionsby front line soldiers that might lead to fratricideis spottier and less reliable. Thus, many casualtiesdue to fratricide are never realized to be such, andmany that are recognized as fratricide are proba-bly never recorded as such. Within the militaryhistorical record, the record of fratricide is partic-ularly suspect because fratricide is a mistake anda full airing can be embarrassing or traumatic andcan end careers.

Recording of fratricide has not been uniform.Casualty report forms, for example, have not

included fratricide as a cause. Thus, fratricidesduring the Vietnam War were cataloged undereither ‘‘accidental self-destruction” or “misad-venture.

Colonel Washington’s unfortunate “misad-venture” illustrates these problems well. Afterthe Loyalhanna incident, Washington was criti-cized by some of his officers for losing hiscustomary aplomb under fire, for which he wasjustly famous. What responsibility he felt after theaction we can never know, but he made nomention of the circumstance of his casualties inthe next day’s reports to his superior officers. Infact, he never mentioned the event in any of hiswritings until almost 30 years later when, inmarginal comments on a draft of his own biogra-phy, he related a version in which Colonel Mercerclearly fires the first shot.

The historical record does provide lessons.Many of the cases of fratricide include humanerrors, not just technical or tactical specifics.Because people change more slowly than ma-chines, history probably provides some usefullessons for today.

Very few works are devoted specifically tofratricide. One particular case of fratricide isprobably the most famous because a popular bookwas written about it, Friendly Fire, by C.D.B.Bryan (also serialized in the New Yorker and thesubject of a television series);4 this work dealsprimarily with a victim’s family and its dealingswith the U.S. Government. Lt. Colonel CharlesSchrader’s paper, Amicicide, contains far andaway the largest collection of historical ancedotesof fratricide of any single source and it is citedwidely in this chapter.5 OTA also contracted fortwo papers on fratricide and they are used freelyin this chapter.6

4 Courtlandt Dixon Barnes Bryan, Friendly Fire (New Yorkj NY: Putna.q 1976).5 Charles Schrader, Amicicide: The Problem of Friendly Fire in Modern War (Fort Leavenwo!@ KS: U.S. Army Comman d and General

Staff College, 1982).

c Richard R. Muller, “Fratricide and Aerial Warfare: An Historical Overview” and John C. Imnnquest and W. Blair Hayworth, Jr., ‘‘OTAFratricide Study.”

—. — .—

Chapter 2–Historical Review of the Causes of Fratricide 9

This chapter is organized not by chronologybut rather by the lessons important to findingtechnical and procedural solutions to the problemof fratricide. This approach is necessarily some-what arbitrary since fratricide almost alwaysresults from a complex and confused chain ofmistakes, making strict categorization impossi-ble. (A gunner may aim toward a friendly targetbecause he is disoriented but certainly will notpull the trigger unless he also fails to identify thetarget properly. Is the fratricide then due to hisdisorientation or his failure in identification?)

The chapter concentrates on, but does notrestrict itself to, American experience. Thisshould not, of course, imply that the U.S. militaryhas a particularly serious problem with fratricide;even a quick glance at military history shows thatevery army that has fired a shot has had to takeinto consideration hitting one of their own, or elsequickly learn hard lessons. Following the histori-cal anecdotes are some data from the NationalTraining Center—an instrumented, automatedfacility for combat manuevers--and finally asynopsis of the Persian Gulf incidents.

TYPES OF FRATRICIDEThere are no universally accepted definitions

of friendly fire’ or ‘‘fratricide. The broadest—and older-definitions include any case in whichanyone is hurt by a weapon from his own sideother than his own. Thus, if an artillery round isfaulty and falls short on friendly forces, that isfriendly free; but if it is faulty and blows up in thebreech and kills the artilleryman pulling thelanyard, that is an accident. More recently, themilitary has adopted definitions that exclude pureaccidents and grossly malfunctioning equipment.

The narrowest definition would include onlywillful, but mistaken, attacks on friendly forces.The current Army Training and Doctrine Com-mand (TRADOC) definition is: ‘‘The act of firingon friendly personnel or equipment, believingthat you are engaging the enemy. ’

This report uses a definition that excludespurely mechanical malfunctions but includes allother cases of friendly personnel receiving firefrom weapons operated by other friendly person-nel. Perhaps surprisingly, the material differencebetween the definitions is not great since fewfratricide result solely from equipment failure.

1 Fratricides Due to AccidentsMalfunctions always occur, of course, and

when dealing with weapons, they can be deadly.For example, in 1968, an F-4 flying to supporttroops engaged near Ban Me Thout, Vietnam,dropped a napalm canister on a church, killing 13civilians. The cause was determined to be simplya faulty bomb rack.8

More often, however, malfunctions are justpart of a chain of errors, sometimes compoundedby human actions. For example, in World War 11,the lead bomber of a group would determine thetarget and all others in the group would releaseupon seeing the leader’s release. During the attackon the Abbey of Monte Cassino in March 1944,a bomb rack malfunction resulted in the prema-ture release by a lead bomber, which resulted inits and others’ bombs being dropped on friendlypositions. Similarly, on July 24 of that year,during the preparation for the breakout toward St.Lo, when the bomb rack on one lead bomberprematurely released, the other 15 bombers in thegroup immediately released their loads; unfortu-

7 Briefing entitled, “TRADOC Fratricide Study” (undated but received 1992). Note that, in practice, the Army does not always stick withthis very strict definition, For example, some of the incidents from the Persian Gulf would have to be called “accidents” if using this wow

definition,8 Schrader, op. cit., footnote 5, p. 55.

. . . . . . .— ---- -- . -. --- . . - . . . - - - -


nately they landed on the U.S. 30th InfantryDivision, killing 25 and wounding 131.9

Ever since World War I, cases of rear gunnersdamaging their own aircraft have been commonand continue to the present. Helicopter gunshipsare equipped with traverse-limiting rods thatprevent the side-door machine-guns from swing-ing so far to either side that rounds could hit thehelicopter. When such a rod broke on onehelicopter during the Vietnam War, the gunner inthe heat of battle tracked a target so far forwardthat he fired into the cockpit and wounded thepilot.

Even when weapons operate properly, unfortu-nate circumstances can cause what, in the broad-est sense at least, might be classed as fratriciderather than accident. In the early morning hours of5 November 1942 during the second Battle ofGuadalcanal, the destroyer Walke was hit byJapanese naval guns and torpedoes. With theburning destroyer clearly sinking just a fewminutes after being hit, the battleship Washingtonpassed close by and launched life rafts for theWalke’s crew. However, when the destroyer wentdown soon after, the depth charges on its hullexploded—just as they were designed to whenreaching a certain depth-killing many of thecrew in the water above.10

Some accidents due to human error could beavoided by different equipment design. A U.S.bomber in World War H bombed an air base of theU.S. Ninth Tactical Air Force after the bomberwas hit accidentally by a packet of chaff; thesurprise caused the bombardier to hit mistakenlythe bomb release switch.11 During the Vietnam

War, an F-100 attacking a North VietnameseArmy Headquarters instead dropped bombs overa kilometer short on U.S. troops when the pilot hitthe bomb release while trying to adjust theaircraft’s trim button.12

S Fratricide Due to Command andControl Failures

Failure of command and control is a far morecommon cause of fratricide than simple failure ofequipment. Command and control includes tell-ing units where to be, having units know wherethey are, and keeping units properly informed ofthe locations of neighbors.

The nighttime Battle of Cape Esperance beganlate on the night of October 11, 1942, with thedestroyer Duncan breaking away from her groupand charging off in total darkness toward aformidable Japanese fleet. The Duncan closed ona Japanese cruiser and opened fire. The crews ofthe American cruisers, seeing gun flashes verynear known Japanese cruisers, assumed that theflashes came from a Japanese ship and attackedwith eight inch guns. The flashes stopped almostimmediately. Very likely the target had been thehapless Duncan, which was in flames, unable tofree, and sinking within ten minutes of leaving itsgroup. (Misidentification had been her downfallbut it might also have saved her from furtherattack. During the Japanese retreat, the Japaneseapparently also assumed that the ship within theirmidst was Japanese and did not attack, althoughheavy cruisers passing very close by could havedisintegrated the little destroyer with a singlesalvo.13)

9 Robert H. George, ‘‘The Battle of France,” in The Army Air Forces in World War II, Volume Three, Europe: Argument to V-E Day,January 1944 to May 1945, Wesley Frank Craven and James ha Cate, eds. (Washington, DC: U.S. Government Printing OffIce, 1983), pp.228-277.

10 ~5 repated a s~w ~agedy of a week ~lier when the des~oyer ~fey S* ~d her &p~ chmges killed several C)f those few (Xew

members that were able to abandon ship. C.W. Kilpatrick, The Naval Nigh Battles in the Solomons (Pompano Beack FL: Exposition Pressof ~Oli@ 1987), pp. 91, ] ] 8, and 121.

1 I by h Forces, op. cit., footnote 9, p. 230.

12 Sctiader, op. cit., foomote 5, p. 58.

13 Kflpa~&, op. cit., foomote 10, pp. 52-64.

.. — — .—— ..—

Chapter 2–Historical Review of the Causes of Fratricide! 11

The occasional misplaced shot is bad enough,but worse fratricide can occur when two friendlyunits start exchanging fire. One unit fires bymistake and the other unit-assuming fire to be apositive identification of enemy—returns fire. OnAugust 8, 1944 during the fighting on Guam, twobattalions of the 77th Infantry Division got into aprolonged fire fight. The exchange might havestarted as each side fried off several mortar roundsto calibrate the weapons’ emplacement positionsbefore settling down for the night. Rounds fromeach side fell near the other; both assumed that itwas Japanese fire and thus returned fire withsmall arms and more mortars. This firing, ofcourse, made it obvious to each unit that the otherwas enemy and then the accompanying tanks gotinvolved. A real firefight was under way. Finally,their mistake became apparent, in part when eachbattalion called up the same artillery battalion torequest that artillery fire be directed at the other.

A similar exchange became one of the worstcases of fratricide in the Vietnam War. Oneartillery unit aimed its guns correctly but used thewrong powder charge so the rounds went too farand landed on another U.S. artillery position. Thesecond position responded with deadly accuratecounterbattery fire. This duel went on for over 20minutes and resulted in 90 casualties, all fromfriendly fire.14

Command and control procedures can preventfratricide when identification is difficult. Fromthe time that aircraft were first used for groundsupport in World War I, airmen knew thatidentification of ground units would be difficult.General William ‘‘Billy’ Mitchell said, “Ourpilots had to fly right down and almost shakehands with the infantry on the ground to find outwhere they were. ’15 To avoid fratricide, both ofinfantry and pilots, World War I military com-

B-17s of Eighth Army Air Force bombing Berlin in1944. One plane strays beneath another and has its tailclipped by falling bombs. Even perfect identification offriends will not eliminate all fratricide.

14 sc~ader, op. cit., footnote 5) P. 21.

15 Ric~rd Hallion ~frjjre From f~e Sk),,. The ~i$tov, ~}f Bar(/efie/dAir Ar~ack, ]9]]-1945 (Was~~on, DC: Smithsonian Institution pI13SS,

1989), p. 39.

---- . . . - ..-. — - -- -- - - - — --- - - - -


manders divided up combat areas into ‘‘no-fwe”zones and “free-fire" zones.16

At the outbreak of World War II, the Germanshad the best developed system for air-groundcoordination. They began with a World War Isystem of colored panels to mark infantry posi-tions. This worked well in the attacks on Polanduntil Polish defenses broke and the German armybegan a war of maneuver. The German 10thPanzer Division then reported ‘constant” attacksby friendly airplanes.17 The same story was .

repeated on the Western front, the Germansintroduced a system of safety lines (Sicher-heitslinie) to avoid attacks on their own troops,18

which worked well at first but, again, once a warof maneuver began deep in French territory,fratricidal attacks increased sharply.

Clearly, in both cases above the change was notin the ability to identify; German pilots did notsuddenly forget how to identify tanks that theycould have identified the day before. Rather,identification had been difficult all along andoperational command and control procedures,developed to serve in lieu of identification, brokedown when the character of the fighting changed.

Recognizing the importance of rapid maneu-ver, and the strain it placed on any operationalmeasures to avoid fratricide, the Germans workeddiligently to develop better ground-to-air sig-naling and training programs to increase pilots’identification skills.

The single most famous case of fratricideillustrates the extreme difficulty of coordinatinga complex attack by hundreds of elements, evenalong a stationary front. The carpet bombing in

World War II classroom instruction in aircraftidentification using resin models.

preparation for the Normandy breakout near St.Lo was filled with problems, with those precipi-tated by mechanical failure of a bomb rack citedabove being just the beginning.

German defenses around the Allied armies inNormandy were tenacious but spread thin. Alliedcommanders decided that a single, concentratedblow would breach the defenses, allowing Alliedarmor to pour through the gap. The plan topuncture the defenses was called Operation COBRA.The first phase of the breakout was to be a carpetbombing of German positions. The attack was tostart with 380 medium bombers hitting specifictargets, followed by over 1,500 heavy bombers,B-17s and B-24s, dropping over 3,300 tons ofbombs, with over 500 fighter-bombers and dive-bombers attacking anything that survived. Forthree hours, the sky would be filled with air-planes. 19

16 M~~~ ~Wer, ~d,, The united StateSAi~S~~iC~ in world w~~], vo~, II (washingto~ Dc: Office of Air Force fistory, 1978), p. 373.

17 Williamson Murray, “The Luftwaffe Experience, 193941, ‘‘ in B.F. Cooling, cd., Case Studies in the History of Close Air Support(Washington DC: Office of Air Force History, 1990), pp. 85-86.

16 Ge.do. ~. ~eekows$ Ia, ~age~eft Z. Km Nr 2, ‘‘Zusamrnenarbeit d der Luftwaffe, ’ National ArChiWS and Records

Administration Washington (NARA) T3 14/615/372-393.19 ~ese nub~s come from Cmven ad Cate, op. cit., foo~ote 9, p. 232. Most of the narrative is taken from MilrhXl“ Blumensoq Breakout

and Pursuit (WashingtorL DC: Department of the Army, OffIce of the Chief of Military History, 1961), pp. 228-239. Part of the problem witha historical review of fratricide can be seen by a comparison of these two “official” histories, one by the Air Force, the other by the Army.The Army was on the receiving end and their history relates much of the controversy between the air and ground commands, while the Air Forcehistory states somewhat matter-of-factly that “Wchnically viewed, the bombing was good. ” (p. 233)

Chapter 2—Historical

The attack was delayed a week by weather,Then the July 24, 1944 attack was partiallyunderway when it, too, was called off because oflow visibility. (But not before the inauspiciousshort bombing described above.) Finally, themain attack took place on July 25.

General Omar Bradley thought that the benefitof the bombing would be greatest if the infantryattack could follow immediately after. He wantedground forces as close as 800 yards even thoughthe air commanders warned that 3,000 yards wasthe closest safe distance. They compromised at1,450 yards.

Bradley and other ground commanders hadinsisted that the bombers approach parallel to thefront, so that any ‘‘short’ bombing would resultin bombing the wrong Germans, not in bombingthe wrong side. The air commanders argued thatthis made their machines too vulnerable toantiaircraft fire. Bradley seems to have believedthat he had agreement when he left the lastplanning meeting. Years later he insisted that“the Air Force brass simply lied” about thedirection of the attack.20

With so many aircraft, mistakes were inevita-ble, Visibility was poor. Heavies were to bombfrom 15,000 feet but a layer of clouds forcedmany down to 12,000 feet, which in turn forcedgroups to reassemble in crowded skies andbombardiers to recalculate bombing solutions inflight. Allied positions were marked with smoke,which was hard to see in the haze and essentiallyuseless once the bombing started, since the bombexplosions raised mountains of dust that mixedwith the smoke.

Human error was the cause of most of the shortbombings. Mistakes were disastrous when com-mitted by the lead plane of a group becausecommand and control procedures called for thelead plane to sight the target and all other planesto release when the leader did. One lead plane had

. .

Review of the Causes of Fratricide 13

a broken bomb sight and released visually.Another bombardier thought he was on target butwas orienting on the wrong landmarks.

Succeeding flights of bombers would almostnever be able to see their targets because of thedust raised by first salvos. Therefore, their at-tempts to bomb targets were really the bombingof dust clouds, under which they hoped the bombswould find targets on their own.

Unfortunately, in this case, wind blew the dusttoward U.S. positions and every wave of bombersstruck a little closer. The war correspondent ErniePyle wrote later:

As we watched there crept into our conscious-ness a realization that windows of explodingbombs were easing back towards us, flight byflight, instead of gradually easing forward, as theplan called for. Then we were horrified by thesuspicion that these machines, high in the sky andcompletely detached from us, were aiming theirbombs at the smokeline on the ground, and agentle breeze was drifting the smokeline backover us! An indescribable kind of panic comesover you at such times. We stood tensed in muscle

Soldiers of the U.S. 30th Infantry Division dig out afterbeing bombed near St. Lo by the U.S. Eighth Army AirForce.

20 l-l~lio~ op. cit., fOOtIIOte 15, p. *11.


and frozen in intellect, watching each flightapproach and pass over us, feeling trapped andcompletely helpless.21

A company commander wrote,The dive bombers came in beautifully and

dropped their bombs right where they belonged.Then the first group of heavies dropped them inthe draw several hundred yards in front of us. . .The next wave came in closer, the next still closer.The dust cloud was drifting back toward us. Thenthey came right on top of us. We put on all theorange smoke we had but I don’t think it did anygood, they could not have seen it through thedust. ..22The results of the misplaced bombing were

deadly. Added to the casualties of the abortiveattack on the 24th, the short bombings on July 25caused official casualties of 490 wounded and111 dead.23 In addition, the 30th Infantry Divisionalone reported over 160 casualties due to “com-bat fatigue, ” that is, soldiers simply stunned bythe experience but not necessarily showing anybodily damage.

Among the dead was Lieutenant General LeslieMcNair, Commanding General of the ArmyGround Forces, pro tern commander of the 1stU.S. Army Group, and a strong supporter ofair-ground combined operations. He had come tothe forward area on the 25th specifically to helpmorale after the short bombings of the 24th.24 Hewas killed instantly and could be identified onlyby his West Point ring.25

The bombings at St. Lo caused resentmentbetween air and ground commanders. The com-

mander of the 30th Infantry Division said, ‘Theresno excuse, simply no excuse at all. I wish I couldshow some of those air boys, decorated witheverything a man can be decorated with, some ofour casualty clearing stations." 26 General DwightD. Eisenhower reportedly swore never to useheavy bombers in combat support again, but theirusefulness was too apparent and the ban did notlast.

On the positive side, Operation COBRA alsomotivated important U.S. improvements in com-mand and control of bomber groups and inprocedures for marking of friendly lines on theground. During Operation QUEEN, the Alliedattempt to breach the Roer River, a carpetbombing preparation like that for OperationCOBRA was to open the way for the infantry.This time giant fluorescent cloth panels markedthe positions of friendly troops and tetheredballoons flew parallel to the front line. U.S. troopsalso marked their positions by using their 90millimeter antiaircraft cannon to fire red smokeshells straight up, and the bombing went well.27

Repeating the earlier German experience, how-ever, the Allies found that these command andcontrol procedures-depending on careful marking—broke down as soon as a war of maneuver began.

In operations near Cherbourg, a single Britishdivision, the 51st Highland, on a single day,August 18, 1944, reported 40 separate attacks byfriendly aircraft (more than occurred during theentire Persian Gulf War).28

On August 7, 1944 during OperationTOTALIZE—a franticly paced and fluid attempt to cut off a

21 Cited @ Mm H~~gs, over/oral: D.Day, June 6, 1944 (New York NY: Simon and Schuster, 1984), p. 254.

22 Blume~oq op. cit., footnote 19, P. 237.

23 Ibid., p. 236.

24 Rob@ L, HeWrit~ ~~~~ ~orSe ~~t~e ~eSfern ~~ont: ~~e ~Iory Of f~e j~t~ ]~~a~fry Djvjsion ~uhingt(l~ DC: hlfillltry Jod PIESS,

1946), p. 37.25 Russ~~ Weigley, ~~Sen~ower’S f.ieu(enanfS @100mingtou IN: Indiana University press, 1981), p. 170.

26 Hmtigs, op. cit., foomote 21, p. 255.

27 Kent Ro~~ Gree~lcld, ( ‘Ax-DIy Ground Forces md Air-Ground Battle Team, ” Study No. 35 (Historiczd Section: Army Ground Forces,1948), p. 89.

28 HM~g, Op. cit., fo’o~ote 21? p. 303-

—— ...—.— — — ——. -. —

Chapter 2–Historical Review of the Causes of Fratricide [ 15

huge German force fleeing through Falaise--U.S.heavy bombers bombed short, causing 300 casu-alties among British, Canadian, and Polishground forces.

A week later, British bombers attacked U.S.Army forces. The primary culprit was a failure ofinter-allied coordination. U.S. Army units usedyellow smoke to mark their positions while theRoyal Air Force used yellow smoke to mark itstargets. A historian records one witness saying," . . . the more the troops burnt yellow flares toshow their position the more the errant aircraftbombed them.”29

Sometimes failures of communication haveforced gunners to fire on forces knowing full wellthat they are friendly. By the very strictestdefinition--that is, willful but mistaken attackson friendly forces—this would not be friendlyfire. The German general Guderian recounts how,during the blitzkrieg into France, Luftwaffeairplanes attacked his units. The ground unitsknew that the airplanes were German but wereforced to return fire in simple self-defense. Onepilot bailed out when flak hit his plane and Guderianhimself was waiting for him on the ground.30

General Omar Bradley recalls that a flight ofAmerican A-36s attacked his armored column inSicily. The tankers properly identified the aircraftas friendly and lit off yellow smoke flares, themarkers for ‘‘friendly’ armor, but the attackscontinued so the tanks returned fire and downedone of the planes, When the pilot parachuted toearth, the tank commander said, ‘‘Why you sillysonuvabitch, didn’t you see our yellow recogni-tion signals?” To which the pilot replied, “Oh. . .is that what it was? ’31

On August 15, during the breakout fromNormandy, one American fighter pilot had thebad luck to mistakenly strafe the headquarters of

the XIX Tactical Air Command near Laval.Antiaircraft gunners knew full well that the planewas American but again for self-defense wereforced to return fire. Flak brought him down.32

I Fratricide Due to Fire DisciplineFailures

At the lowest level, ‘‘command and control’devolves into something as straightforward as“fire discipline.’ Indeed, where command andcontrol concerns the actions of units, fire disciplineconcerns the actions of the individual shooter.

The following case illustrates a string ofmistakes, fire discipline being just one: In thefighting in France, a group of eight tanks set outin low visibility in late afternoon of July 9, 1944.They were under strong pressure from superiorsto make a symbolic advance by the end of the day.At a critical road junction, the group commanderturned right instead of left, bringing them uponCompany C of the 823 Tank Destroyer Battalion,later to hold a U.S. record for most Germanvehicles killed.

Based on the tactical situation, the companyexpected no U.S. tanks to be on that road and thetanks were approaching from the direction ofGerman lines; thus the company reasonablyassumed that they were German. The first tanktook a direct hit from a 76 millimeter antitankcannon and was destroyed. The others continuedto advance and opened fire. At 400 yards, thedefenders recognized the tanks as American. Onevery brave sergeant stood up and waved wildly toget the tanks to stop firing but they kept on. Thedefenders stopped firing, took cover, and hopedfor the best. One of the passing tanks shot at anuncamouflaged halftrack at a range of 15 yards,wounding a driver. 33 This incident shows that

29 John Terraine, A Time for Courage: The RAF in the European War, 1939-1945 (New York NY: Macmillan, 1985), p. 661.

30 Hek Guderian, Panzer Leader (New York, NY: Duttoq 1952), p. 113.

31 Hallioq op. cit., footnote 15, p. 178.

32 Craven and Cate, op. cit., footnote 9, VO1. HI, p. 255.

33 Sckader, op. cit., footnote 5, p. 82.

. . . . - -- - .- . - . - -- - - -- . - - -- - - —


!*u);”nz kzr< . &. -.. J-Cr . . co ‘z “’w’’”* ‘“ ““”” . . ,4 ., ,

*OU %, %,;*‘+, “

Aircraft Spotters’ Guides distributed to troops in thefield during World War II. The most basicidentification system is comprised of the human eyeand the human brain, but these require training to beeffective.

fratricide occurs because of compounded errors;in this case, starting with poor navigation, poorcommunication, and faulty identification, andending with lack of fire discipline, since inabilityto identify can hardly justify firing at a halftrackat 15 yards.

Lack of fire discipline was the major contribu-tor to the worst antiaircraft fratricide in WorldWar II. On July 11, 1943, the beachhead on Sicilywas to be reinforced with a drop of 2,000paratroops. The troops left Tunisia in 144 C-47transports. The fleets and ground forces had beenalerted to the drop. The flight went well until,

when crossing over the coast at a very visible butvulnerable altitude of 1,000 feet, the transportswere fired on by a lone machine-gunner. Thebeachhead forces had been attacked just hoursearlier by German bombers and, once the firingstarted, discipline collapsed; everyone on theground was ‘throwing everything they had at us’as one airborne company commander later put it.A destroyer even fired on transports that hadditched at sea. The results were a disaster. Of the144 transports, 23 were shot down and 37 dam-aged. One hundred forty one paratroops and aircrewmen were killed. Many of the transports thatsurvived did so by scattering; thus, of the originalforce of 2,000, only 500 or so could be effectivelyorganized on the ground in the drop zone.34

Just two days later, the British suffered asimilar incident in their zone of Sicily. Nineteenhundred paratroops were to capture the Primosolebridge, but more than half the transports were hitby either ship- or ground-based antiaircraft fire.Three hundred men did reach and capture thebridge. 35

I Fratricide Due to Navigation FailuresClosely related to command and control are the

problems of navigation. During World War IIPacific fighting and in Vietnam, many artilleryfratricides resulted from forward observers cor-rectly calling in fire relative to their putativepositions, but not knowing their own locationsprecisely. 36 This seems characteristic of junglefighting, when forces could go long periodswithout knowing just where they were. TheMarine Corp still refers to “The Battle of theTenaru River’ (on Guam), which actually tookplace on the Ilu River, but because the maps wereso poor and the vegetation so thick, the men on theground did not know that at the time.37

34 (XUlo r)’)ZSte, Bitter Vic(ov: The B~tr/e for Siciliy, Ju/y-A~g~Sr J$)43 mew York NY: William cobs, 1988), pp. 307-308.

35 A&fiGml~d ~dHowmd smy~,si~i/y ~~thesurrenderof]taly (was~gtou Dc: Us. Army, Office of the Chief of Military History,

1965), p. 218.

36 Sckader, op. cit., footnote 5, p. 23.

37 ~fim stew- Gw&l~ana[: World War II’S Fier~e~t Naval campaign (@redo% l%gl~d: Willi~ Kimber, 1985), p. 50.

Chapter 2–Historical Review of the Causes of Fratricide I 17

Navigation, not identification, clearly was theproblem when Navy dive-bombers were called into attack Japanese defenders on the tiny island ofTanambogo. The planes bombed the wrongisland, adjacent Gavutu, killing three Marines andwounding nine others.

Navigational blunder was responsible for mis-taken aerial bombing of civilians in World War II.One of the earliest incidents was a German error.On May 10, 1940, 20 Heinkel 111s set off tobomb Dijon. One Luftwaffe lieutenant got sepa-rated from the group due to bad weather. When acity suddenly appeared below, he took it to beDijon and bombed it but it turned out to be theGerman town of Freiburg. (Instantly, the Ministryof Propaganda announced that Allied aircraft hadinitiated a deliberate policy of ‘terror bombing, ’with the innocent citizens of Freiburg as the frostvictims. )38 In general, however, these sorts ofgross navigational error rarely caused fratricide inWorld War II.

H Fratricide Due to IdentificationFailures

Finally, many fratricide are due to straightfor-ward misidentification. The first aircraft used incombat in World War I did not even displaynational insignia. When German ground firebrought down a Zeppelin on August 23, 1913, theGermans painted Iron Crosses on all their aircraft.The British adopted markings too, but quicklylearned that gunners confused the Iron Cross andthe Union Jack insignia so they switched to red,white, and blue roundels similar to those used bythe French at the time.

Oftentimes an observer sees what he is lookingfor, not what he is looking at. The “scientificmethod’ calls for first forming a hypothesis andthen searching for evidence that it might bewrong. Psychological tests show that peoplerecognize things in the opposite way, by forming

L

o

A World War II classroom for training sailors in theidentification of naval ships and airplanes.

a hypothesis and then searching for additionalevidence that it is correct.

On May 15, 1941, a formation of FaireySwordfish took off from the British carrier ArkRoyal as part of the epic search for the Germanbattleship, Bismarck, They soon spotted a largewarship and launched torpedoes against it. But itwas the cruiser HMS Sheffield, a ship that did notlook anything at all like the Bismarck. A historianwrote, “Expecting to see the Bismarck, Bismarckis what they saw. ’ Fortunately, skillful evasionby the Sheffield ensured that the torpedoesmissed, One Swordfish pilot radioed, ‘‘Sorry for

the kipper. ’ ’39

Seeing what one expects to see accounts for aparticularly dangerous opportunity for fratricide:patrols returning to friendly lines. Since scoutsand patrols are coming from the direction of theenemy, getting past friendly, but nervous, guardsand lookouts can be tricky. Thomas ‘‘Stonewall’Jackson went ahead of his own troops to recon-noiter Union lines during the Battle of Chancel-lorsville. Just as Jackson was re turning to confed-erate lines, forward units of General JosephHooker’s Union infantry reached the North Caro-lina troops near Jackson. Some shots were firedand the hastily dug-in Confederates were greatly

38 Wolfgmg Dierich, Kampfgeschwa~er “Edelweiss”: The History of a German Bomber Unit (Londou England: 1~ Men, 1975), p. 26

39 Ludovic Kennedy, F’ursuir: The Chase and Sinking of the Battleship Bismarck (New York NY: PinnaCle, 1975), pp. ls~lss.

. . . . . . . . . . . . . . . . .


I

● A 1 \ ‘ ‘ ‘“l +’10

-ka-+

; \ z

—wr’w--* \01’

World War 11 troops were given free packs of AirplaneSpotter Playing Cards. Maintaining goodidentification skills requires constant practice,

excited by anticipation of the oncoming assault.Hearing the firing, Jackson hurried back towardhis own lines but under the nervous conditions aforward Confederate picket—seeing riders ap-proach from the direction of known enemyforces—shot and mortally wounded the general.40

Some admirers of Jackson argue that his deathchanged the course of the war, which, if true,would make it the fratricide of greatest conse-quence uncovered during this research.

Misidentification due to similarities betweenweapons are more understandable. Many friendlyfire losses during the Battle of Britain wereattributed to the similarity of the SupermarineSpitfire Mark I and the Messerschmitt Me-109Efighters. Also, the Bristol Blenheim twin-enginefighter resembled the German Junkers Ju88

medium bomber. The latter similarity lead to thedestruction of three Blenheims. One section ofCanadian Hurricanes, thinking that the planesbelow them were Junkers, attacked but pulledaway at the last second when the leader realizedhis mistake. The next section still went in for theattack, in part because they mistook the yellowand red Very recognition flares for tracer fire frommachine-guns. One Blenheim blew up in the airand the other two crash-landed.41

During World War II, considerable effort andattention went to improving identification ofaircraft. As one response, the British developedelectronic IFF devices. A touring “air circus”was also organized so ground forces couldpractice identifying real aircraft overhead, notresin models in a classroom.

TACTICAL CONSTRAINTS DUE TO FEAROF FRATRICIDE

At times in the past, fratricide has beenaccepted as a costly necessity of combat. Forexample, World War I tactics made some fratri-cide almost inevitable. Trench defenses could becaptured if the defending machine guns weresuppressed by artillery fire while the attackersapproached the trench works; thus, the attackerswanted the artillery to pound defensive positionsup to the last second. In this situation, attackerswere willing to have friendly artillery fall veryclose because they believed that the losses due tofratricide would probably be less than those fromenemy machine-gunners .42 A World War II bat-talion commander said, “We must teach oursoldiers to remember that when they follow theartillery barrages and air strikes closely, they

40 G,F.R. Hendersoq stonewall Jackson and the American Civil War (New Yorlq NY: Da GPO, 1988), P. 678.

41 Ricmd Collier, Eagle Day (New York NY: AVO~ 1%9), p. 140.

42 For ample, ~m ~ V+ Hogg, Ba~age: the Gum in Acfi~n (New York ~: B~lentine BOOkS, 1970), p, 21, “me Frenck With their

greater elan and still-unconquered spirit of attack at all costs, were known to observe that unless the infantry suffered 10 pereent casualties fromtheir own artillery, they weren’t following the barrage close enough!” or”. . the ‘creeper’ [rolling barrage] covered the ground progressivelyin front of and behind the objectives. All the infantry had to do was to stay close to it even if the occasional short round sprayed them withshrapnel.” from Shelford Bidwell and Dominick Graham, Fire Power: British Army Weapons and Theories of War, 1909-1945 (Imndo~England: Allen and Unwiu 1982), p, 111.

——.—

Chapter 2–Historical Review of the Causes of Fratricide 19

eventually suffer fewer casualties even though anoccasional short may fall on them. ’ ’43

Yet no commander can afford to be indifferentto fratricide and avoiding it can limit tacticaloptions. For example, the need to insure safety ofinfantry operating with artillery severely ham-pered the flexibility of the massive British battleplans for the great World War I attack at theSomme.

Fratricide has a greater cost than the directcombat loss of the forces hit. Fear of fratricide canso inhibit a commander’s actions that combatefficiency is much reduced. The Center for ArmyLessons Learned (CALL) reports several conse-quences of fratricide incidents that reduce combateffectiveness.44 These are listed in Table 2-1.

In addition, friendly fire does not need to kill tohave a suppressive effect. In some instances, forexample in World War II battles on the islands ofBiak and Luzon, groups of infantry as large asbattalions spent whole afternoons pinned downby friendly fire of various sorts, seriously disrupt-ing coordination of attacks.

Nighttime World War 11 naval battles are filledwith cases of tactical confusion in general andfratricide in particular. During the Battle of CapeEsperance, control broke down from the start withthe charge of the destroyer Duncan, describedabove. The situation became so confused soquickly that the American group commander,Admiral Scott, ordered ‘‘Cease firing, our ships!Scott further ordered that ships flash their recog-nition light (colored lights up either side of thebridge) .45

The Americans’ problems with sorting out thesituation benefited the Japanese enormously.Firing was halted for four minutes. This may not

Table 2-l—Detrimental Effects of Fratricide

● Hesitation to conduct limited visibility operations. Loss of confidence in unit’s leadership● Increase of leader self-doubt● Hesitation to use supporting combat systems● Oversupervision of units● Loss of initiative● Loss of aggressiveness during fire and maneuver● Disrupted operations● Needless loss of combat power● General degradation of cohesion and morale

SOURCE: Center for Army Lessons Learned.

seem like much until considering that the heavycruiser Salt Lake City had fired 80 eight inchrounds in the first two minutes of the battle andthe light cruisers were averaging an incredible150 rounds a minute each—so four minutes wasa long time. Furthermore, the signal lights re-vealed the ships’ locations to the Japanese.

Finally, the lights identified the U.S. ships forthe Japanese. Ironically, their commander, Admi-ral Goto, had thought that he was under mistakenattack from another Japanese force and had beenhesitant to return fire, but the distinctive recogni-tion lights showed the force to be American.Within an eight minute period shortly aftermidnight, Scott ordered recognition lights to beflashed three more times. The group was nevercompletely under control and both sides withdrewwithout a clear decision.

During the Battle of Savo Island in the earlymorning hours of August 9, 1942, the crew of theheavy cruiser Vincennes believed that she wascoming under friendly fire so she hoisted a hugeAmerican flag up the mast. The Japanese assumedthat this must mean that she was the flagship andtherefore concentrated their fire on her. Twominutes later, Japanese cruisers took the U.S.destroyer, Ralph Talbot, under fire. Her skipper,

43 Sctiader, Op, Cit,, foo~ote 5, p. 17. Also from footnote 53 of Cbptcr one of Sctider: ‘‘An experienced infantry officer who send asa battalion S-3 in Vietnam related to the author that it was his common practice (and that of others) to accept up to 5 percent friendly casualtiesfrom friendly artillery in the assault before lifting or shifting f~es. The rationale, of course, is that it is preferable to suffer 5 percent easuahiesfrom one’s ovm fire plus 5 percent from the enemy than to permit the encmy, through lack of adequate suppression, to inflict 15 percentcasualties on the attacking force. ’

44 Center for Army ~ssom~ ~amed, Fratricide Risk Assessment for Company Leadership, CALL H~dbook No. 92-3 @ofi ~avenwo~

KS: Center for Army hssons Learned, March 1992), p. 4.

J5 Udcss o~cWise cit~ the following naval accounts arc from Kilpatrick, op. cit., footnote 10.

. . -— . - - . - . . - -


Lt. Commander Callaham, also believed that hewas being fired on by friends so he turned on hisrecognition lights. This action was so unexpectedthat the Japanese were momentarily flummoxed,worrying that perhaps they were firing on anotherJapanese ship. The Japanese cruisers were forcedto use search lights to illuminate the Talbot. Shetook several hits but escaped into a squall.

Nearby, on August 21, again in the earlymorning darkness, the U.S. destroyer Blue and theJapanese destroyer Kawakaze detected one an-other at almost the same time. The American shipwas equipped with radar, which the Japanese haddeployed on only a very few ships-and thoseunits were primitive. For night fighting, theJapanese relied instead on specially trained look-outs equipped with oversized, night-use binocu-lars. During clear weather, the Japanese visualmethod was as good as U.S. radar. In thisparticular case it was better. For whereas the Bluedetected a blob on a radar screen, the Kawakazedetected and identified target. Thus, as the Bluewas creeping forward to get a better view, theKawakaze was launching 21-inch diameter "LongLance” torpedoes that ripped the stern off theAmerican destroyer. She was later scuttled.46

During the closing moments of the Battle ofEmpress Augusta Bay, the cruiser Montpelierreceived direct orders to fire at a target atspecified coordinates. In the pre-dawn darkness,the Montpelier’s commander was uncertain of thetarget and intentionally directed that the firstsalvo should miss. He then listened in on the TBS(Talk-Between-Ships) radio for two minutes and,failing to hear any complaints, commenced firingto hit. This time, the incoming rounds quickly gotthe attention of the target, which turned out to bethe American destroyer Spence. Radio calls for acease fire were heeded before any damage wasdone but, clearly, the lack of identification

required a tactical solution that would have givensome advantage to an enemy, either a chance toevade further fire or return fire.

Caution induced by fear of fratricide can beexploited by an enemy. During World War II,artillery fire began the preparation for an attack bythe 3rd Infantry Division against the town ofOsheim on January 23, 1945. The leading battal-ion reported that shells were landing on theirposition and the barrage was halted. More rangewas added and the barrage resumed but roundsstill fell on the lead battalion. Finally, theAmericans discovered that the fire was comingfrom nearby German tanks that held their fireuntil the barrage started, specifically to fool theAmericans into believing they were receivingfriendly fire and so trick them into calling a haltto the barrage. (Incidently, the 3rd Division lateradopted rules that called for finishing plannedbarrages regardless of reports from forward units,which may have contributed to later fratricides.)47

The Japanese also used the technique ofsynchronizing the artillery fire with their enemyartillery, although perhaps for different reasons.On Guam and elsewhere in the Pacific theater,Japanese artillery and mortar crews would waituntil U.S. artillery was firing before firing theirown guns, thus increasing the difficulty of locat-ing them by their sound. In addition, of course,U.S. troops noticed that when U.S. artillery fired,they often received incoming rounds, causingthem to believe it was friendly fire.

THE PREVALENCE OF FRATRICIDEAny conclusions about the general prevalence

of fratricide developed from a collection ofanecdotes must be regarded with healthy skepti-cism. Several sources use a figure of two percentof casualties that have been due to fratricide in

46 stew~, op. cit., foo~ote 37, p. 50.

47 Sctiader, op. cit., footnote 5, p. 9.

——..——.


20th century wars. In fact, the two percent figureseems to have become almost a rule of thumb.48

One of the great difficulties is knowing whofired which shot. Only rarely is reliable evidenceavailable, but it is sobering to discover that whenevidence is there, it often reveals fratricide thatparticipants at the time were unaware of. Twocases from the same campaign provide interestingexamples. The reader has doubtless noticed thatmany examples used here come from the Solo-mons naval campaign. This is not surprising whenone considers that many of the major battles theretook place at night, resulting in poor coordinationand frequent misidentification. But the sameconditions that made mistakes likely made itunlikely that they would be detected.

In one case, however, U.S. friendly fire left aclear fingerprint. U.S. warships carried a limitedsupply of ‘‘dye-loaded” shells, with each shipcarrying a different color. The added dye allowedtwo ships shooting at the same target to distin-guish the splashes of rounds hitting the water andthus independently adjust their fire. When thelight cruiser Atlanta’s crew examin ed battledamage after nighttime engagements off the coastof Guadalcanal, they discovered nineteen hitsfrom eight inch shells loaded with green dye, thecolor of the heavy U.S. cruiser San Fransisco.

The depleted-uranium rounds used by U.S.tanks during the Persian Gulf War left a similartelltale and again, fratricide rates turned out to behigher than previously suspected. That combat isdiscussed further in the last section of thischapter.

Perhaps the best estimate of overall rates offratricide come not from traditional military

histories and action reports but from the medicalrecords. The U.S. military has long kept recordsof the causes of casualties. Unfortunately, thenormal reports lack enough detail to determineconclusively that a casualty was caused byfriendly free. Moreover, without a detailed ac-counting of all casualties in a given time or place,we cannot know the numerator and denominatorneeded to calculate the fraction of casualtiescaused by friendly fire.

In a few cases, starting in World War II,however, detailed casualty surveys have beencarried out that allow a reliable estimate of thefrequency of friendly fire losses in those cases.The frost was conducted by Dr. James Hopkins,who maintained detailed records of cause ofwound for every casualty in his batallion. Heserved for part of the war on New GeorgiaIsland-near Guadalcanal--and part in Burma.He examined the wounded and conducted inter-views after actions. Hopkins was able to deter-mine that 16 percent of those killed and 19 percentof those wounded were the victims of friendly fireby his broad definition, which includes accidentsin the use of weapons.

49 By TRADOC’s current,

narrower definition, as applied by Dr. DavidSa’adah of the Department of the Army, SurgeonGeneral’s Office, the figures would be 13 percentand 14 percent.50

Two other comprehensive surveys examinedall of the casualties from two divisions inBougainvillea, in the South Pacific, in 1944.Almost a hundred of the killed were morecarefully examined by autopsy to determine causeof death.51 These surveys reveal that 24 percent of

48 See CJCtiader, ~p ~lt,, foo~ote 5, ~, 105; Trevor Duprey, At~ition: Forecasting Battle cas~[h’es a&Equip~nt~sses in Modern War(Fairfax, VA: Hero Books, 1990), p. 3; and Dwight Dickson and Ehin Hundley, “Avoiding Not So Friendly Fire,” Military Review, July 1992,p. 57.

49 James Hopkins, ‘‘Casualty Survey-New Georgia and Burma Campaigns, ‘‘ in James C. Beyer, cd., Wound Ballistics (Washington DC:Office of the Surgeon General, Department of the Army, 1962), pp. 237-280 from the series, Medical Department, United StatexArmyin WorldWar 11, Colonel John Lada, Editor in Chief.

50 see colonel David M. SaJa&@ Office of & Sugeon Gen~~, Headquarters, Dqartment of the Army, “Friendly Fire: wi~ we &t It

Right This Time?’ p. 7.

51 As~ey W. Oughterson et. al., ‘‘Study of Wound Ballistics-Bougainville Campaign, “ in Beyer, ed. op. cit., footnote 49, pp. 281-436.


the deaths were due to fratricide, using thenarrower current TRADOC definition.52

In Vietnam, only the United States and its allieshad certain types of weapons, for example,air-delivered ordnance of any sort-especiallynapalm, certain types of artillery, and so on. Thus,by examining the wounds of casualties anddetermining g the type of weapon that caused them,one can estimate the fraction that are caused byfriendly weapons. Using this approach, someunpublished reports cited in the press estimatethat perhaps 15 to 20 percent of the casualties inVietnam were fratricides.53

The U.S. Army also conducted careful casualtysurveys during the Vietnam War that werecompiled in the Wound Data and MunitionsEffectiveness, or “WDMET” study. The datawere collected between 1967 and 1969 fromelements of one cavalry and three infantry divi-sions. An absolute figure for fratricide is notavailable from the WDMET survey. However, thedata include the type of weapon causing theinjury, and in four cases the type is very specificand was almost certainly in the hands of U.S.-orat least allied-troops: the M16 rifle, the M79grenade launcher, artillery (excluding mortars),and Claymore mines. These four weapon typesalone accounted for 11 percent of all U.S.casualties, including 10 percent of the fatalities.54

The summary of the data compiled by ColonelSa’adah is shown in table 2-2.

These casualty surveys cover only limitedcases for which data are available, but again it isworthwhile to note that in every case where dataare available, the fratricide rate is significantlyhigher than the two percent that frequently

appears in print as the nominal fratriciderate. 55

Despite the hit-and-miss of using historicalanecdotes, the types of fratricide do show somepatterns. As might be guessed, indirect fireweapons or long-range weapons (in past wars,artillery and bombers) have been more likely to beresponsible for friendly fire. Also, the damagedone by these weapons is disproportionately greatbecause mistakes involving single-shot weapons,like tank guns, kill one friend at a time, whileartillery barrages and bomber attacks can devas-tate whole units. The Persian Gulf war did nothave any artillery fratricide. This may be goodluck or reflect an important change brought aboutby better communication and navigation.

Perhaps surprisingly, there seems to be nostrong correlation between type of action andlikelihood of fratricide; it is just as likely duringoffense or defense. Fratricide between neighbor-ing units appear to become more likely the greatertheir separation in the chains of command.Whenever units operate near one another andhave poor communication, poor navigation, or arepoorly controlled, fratricides can occur.

Fratricide of almost any type are more likelyduring periods of limited visibility when identifi-cation is harder. However, although better identi-fication is frequently presented as the solution tofriendly free, Schrader classified only about aquarter of the cases in his review as due primarilyto misidentification. The majority of fratricidewere more properly explained by failures ofcommand and control or fire discipline. See figure2-1.

52 sa’~~, Op. cit., footnote 50, p. 8.53 David M~rnS~~ ~~FWS Not he CMy hat in he GuM WU,’ Nationul Journa/, Feb. 9, 1991, p. 335, cittig tie work of Colonel David

HackWorth.

54 Sa’adti, op. cit., footnote 50, p. 10.

55 D~g tie Coww of ~~ ~~e~~ment, tie au~or ~ ~d seve~ oppo~ties to coll~t perso~ WCOUIItS from military perSOMel that

served in Vietnam. This is admittedly a totally umcientilic sampling but the clear concensus is that a 2 percent fratricide rate is a seriousunderestimate. See also app. A in Colonel Sa’adah’s paper, ‘‘The 2 Percent Nonsense. ’

Table 2-2—Friendly Fire Data in Combat Casualty Surveys—World War II Through Operation Desert Storm

No. of cases KIA + DOW WIA by Prevalence: Prevalence:Survey location/name No. of cases KIA + by friendly f ire friendly fire survey TRADOCforces in survey Line in survey DOW WIA No. % No. % definition definition

New Georgia and Burma/HopkinsJungle perimeter defenseJuly 18-Aug. 5, 1943,Spearhead across Burma,Feb. 15-June 8, 1944

Bougainvillea Beachhead perimeterdefenseFeb. 15-Apr. 21, 1944

Bougainvillea autopsyMar. 22-Apr. 21, 1944, 25% of allKIA + DOW within Bougainvillea survey

Vietnam WDMETComponents of 5 Army divisionsJune ‘67-June ’69, preferably inoffensive engagements

Vietnam WDMET autopsyJuly ‘67-Nov. ’68, 500 consecutiveautopsies within VN WDMET

All U.S. Forces, Jan. 17-Dee. 15, 1991Operation Desert Storm

l alb

2a2b

3a3b

4a4b

5a5b

6a6b

370 102 268 16 16 50353 99 254 13 13 36

1,788 395 1,393 63 16 1561,778 392 1,386 60 15 149

99 99 0 30 30 091 91 0 22 24 0

5,993 1,279 4,714 NC NC NC5,993 NC KIA+DOW=WIA=667

500 500 0 NC NC o500 500 0 51 10 0

613 146 467 35 24 72613 146 467 35 NC 72

19 17.97014

11 12.3%11

0 Not computedo

NC “almostcertainly morethan 10%”

o Not reportedo

15NC Not reported

14%

12%

24%

11%

10940

17%

Within each survey, Line a displays the data as presented in the original study. Line b standardizes this same data to the current TRADOC definition of “fratricide.”

KEY: DOW = died of wounds; KIA = killed in action; WIA = wounded in action; NC= not calculated.

—. -.—. .


Figure 2-1 —Causes of Fratricide: Direct FireFratricide in World War II, Korea, and Vietnam

Inexperience19“/0

(’- Unknown factorsTarget 10%

misidentification260/o

\/Coordination

45%

“,0 = Incidents by category58 total Incidents

SOURCE: U.S. Army

THE NATIONAL TRAINING CENTERData collected from the National Training

Center (NTC) at Fort Irwin, California are impor-tant enough to warrant special notice. The Armymaintains training centers where visiting unitscan engage in mock combat with “OPFORS’ or“opposing forces. ” The NTC is of particularinterest because it has been equipped with asophisticated laser direct-fire engagement systemcalled the Multiple Integrated Laser EngagementSystem, or MILES. The NTC is also equippedwith location and engagement recording systems.

\ Guns on tanks, infantry fighting vehicles, evenpersonal rifles, are equipped with lasers. Whenthe gun ‘fires’ a blank round, the laser also fries.Each of the major weapons and each of thepersonnel have detectors that sense when a laser“hit” occurs.

The laser pulses are coded in such a way thatI the detectors know the type of weapon that freed.

Thus, if a tank detects that it has been shot by a

rifle, nothing happens; but if a rifleman’s sensordetects that he has been shot by a tank, his sensorregisters a‘ ‘kill.’ A kill is signified by activatinga flashing light. Each shot and each hit activatesa small radio pulse indicating type of shooter andtarget, which is picked up by antennas spacedaround the training center. These data are re-corded along with their time, and data areanalyzed by computer after each exercise toexplain to the participants various mistakes thatwere made.

The Army emphasizes that its training centersare for training, not experimenting or data col-lecting, so the system has not been set up withfratricide data collection in mind. It neverthelessprovides invaluable insight into the causes offratricide. Some questions arise about the rele-vance of the data since these are not real battles soperhaps the participants do not behave the waythey would in actual combat. Nevertheless, thedata are enormously rich compared to informa-tion about real battles, so if one believes thesimulation valid, the data are proportionatelyuseful. One can then try to examine hits caused byfriendly fire and hope to learn something of thecircumstances.

Much of the following is taken from a briefRAND study that evaluated data from 83 battalion-sized battles.56 See figure 2-2. Considering thecauses of the fratricides observed, the studystates:

Of the 18 cases of fratricide, one-half couldhave been prevented had the shooting vehiclebeen aware of the location of a sister organiza-tional unit, for the destroyed vehicle was locatedin a friendly formation with no enemy nearby.Another third of the cases could have beenprevented if the shooter had knowledge of thelocation of individual isolated friendly vehicles,a more difficult requirement. One-sixth of thecases involved the killing of a friendly vehiclewhile close to opposing force (OPFOR) elements.

I56 M~~GoldSfi@APplying the National Training ce~terExpen”ence-Incidence ofGround-to-GroundFratricide, R4ND Note N-2438-A

(Santa Monica, CA: The Arroyo Center, R4ND Corporation February 1986).


In this class, only an Identification Friend or Foe(IFF) device could provide the information neces-sary to positively avoid fratricide.57

The data indicate that at least 1 percent of the“blue” vehicles killed were killed by friendlydirect fire. This figure is much less than in thePersian Gulf War. Two possible biases mayexplain the difference. The NTC data may under-estimate some fratricide. For example, if thelasers cannot penetrate through dust, then no killis recorded even though a tank round would havescored a hit in actual battle. This effect works toreduce hits from enemy attacks as well asfratricidal attacks, but since fratricide is morelikely to occur in dustier conditions, it might beunder-recorded to some degree. In addition, theOPFOR train all year long on the same groundand are excellent troops (according to their ownevaluation, not just good but the best). Thus, totalblue “casualties” are usually high in the simu-lated combat, unlike the Persian Gulf experience,and the resultant ratio of friendly fire casualties tototal blue casualties unusually low.

Artillery cannot be simulated with MILES butother means are available, The data availableindicate that 3.6 percent of artillery fire missionsresulted in some fratricide. This result appearsworse when one considers that only about a thirdof the artillery missions hit anything at all, friendor enemy. Thus, of those artillery fires that hitanything, about one-tenth hit friendly forces.Experience at the NTC also suggests that fratri-cide resulting from artillery-delivered mines andunexploded submunitions may be almost asserious a concern as other artillery fratricides,although the MILES data do not now allow aquantification of this effect.

Some fratricides were clearly caused by misiden-tification but more were due to disorientation.(This result depends on the terrain; preliminaryunpublished data from a similar test facility in

Figure 2-2—Types of Information Needed to AvoidFratricide: Direct Fire Fratricide at the

National Training Center

IFF

Locationblue vehl(

330/0

\

/

- - = ’ ’ ’ ’ ’ ~catlonofblue units

50%

“,0 = Number of incidents by cause77 total incidents

SOURCE: U.S. Army.

Hohenfels, Germany suggests that disorientationis not as important there as misidentification.However, the German test range is much smallerand some nearby hills apparently provide easyorientation landmarks.) Several fratricides at theNTC occurred when no enemy forces were nearbyand, moreover, the commanders knew that enemywere unlikely to be near.58 These cases could becured by better fire discipline. Mistakes due totrue misidentification were most common inmelees and poor visibility.

The data collected at the NTC is now beingexploited for fratricide ‘‘lessons-learned.’ Mostof this work is now coordinated through CALL,or the Center for Army Lessons Learned, part ofthe U.S. Army Combined Arms Command at Fort

57 Ibid., p. vi.

58 Ibid., p, 13,


Leavenworth, Kansas.59 Since the Persian GulfWar, observers at the NTC have had to fill outfratricide incident reports whenever MILES de-tects friendly free. Preliminary unpublished re-sults show that observers on the ground attributefratricide most often to identification failure, butthat this accounts for just over a fifth of the cases,with another fifth due to failures of command andcontrol, another fifth due to planning failure, anda combination of communication and navigationproblems and simple mistakes making up thebalance. Incorrect assessment of the tacticalenvironment was cited as the most commoncontributing cause.

MILES is being expanded at other trainingcenters and will no doubt provide increasinglyvaluable information about how fratricides occurand can be avoided.

FRATRICIDE DURING THEPERSIAN GULF WAR

The current surge in interest in fratricide is duelargely to the experience of the recent PersianGulf War. Wars are complex and no two areidentical, so “lessons” from the war should notbe considered universal truths. Although manyconditions in the Persian Gulf were special, someargue that this war was a first example of the“high tech” wars of the future. It may also berepresentative of a type of war that will be morecommon for the United States in the future: onein which massive, overwhelming force is appliedquickly and decisively. A primary appeal of thesetypes of actions is that U.S. casualties arepotentially very low considering the scale of themilitary operation. This report has pointed outalready that one reason the fraction of casualtiesdue to fratricide was high is that total U.S.

casualties were so very low. Another way oflooking at this is that if low casualties arecharacteristic of an important class of futureconflict, then the relative importance of fratricidewill be much greater.

There were 615 U.S. battle casualties in Opera-tion Desert Storm, 148 of which were fatal. Of the148 fatalities, 35-or 24 percent—were causedby friendly fire. Of the 467 nonfatal battlecasualties, 72--or 15 percent—were caused byfriendly free. These percentages seemed high atthe time when compared to those assumed fromwars past but the review of past rates of fratricidesuggest that there has been a substantial underap-preciation of the rate of fratricide in past wars.

Of the 35 soldiers killed by friendly fire, sevenwere on the ground while the 28 remaining werein vehicles. This distribution reflects the highlymobile, mechanized nature of the combat—thatis, most U.S. forces were in vehicles so one wouldexpect more casualties there-but it is alsohopeful for those seeking a technical solution,since mounting combat identification equipmenton vehicles is much easier than mounting it onindividual infantrymen.

A case-by-case description of the known fratri-cide incidents is listed in table 2-3.60 F e windividual cases stand out as being unique to themodern equipment used in the Persian Gulf War.In one instance, a radar-seeking missile lost trackof the Iraqi radar for which it was intendedand-while attempting to reestablish a targettrack-locked onto a nearby U.S. radar. This typeof technology-dependent mistake is, however, theexception; many of the descriptions of friendlyfire—with a change of weapon designation—could have taken place in the deserts of North

59 At he tie of ~~ ~~g, c~’s mse~ch iss~ ~ tit form ~d ~published but their results seem so far to confkn the ~ work.

Major Rick Bogdeu personal communication, February 1993.@ ~s ~omation is ~en from “W@ Robes Friendly Fire Incidents, ” a news release from the Offke of the Assistant Secretary of

Defense (Public Affairs), dated Aug. 13, 1991. Note that some of the incidents are not ‘‘fratricide’ by the narrowest definition now used bythe Training and Doctrine Cornmand. For example, casualties due to faulty missiles or artillery rounds would be considered accidents becausethey did not result from a deliberate act of ftig, believing one was firing at the enemy.


Table 2-3—Persian Gulf Friendly Fire Incidents—1 991

Ground-to-Ground

January 29-Four Marines werekilled when their Light ArmoredVehicle (LAV) was struck by aTOW missile which was firedfrom another LAV west of Kafji,Saudia Arabia.

February 14-Three soldierswere wounded in a small armsexchange during urban clear-ing operations in the town ofArky Amah Al Jadid, SaudiArabia.

February 24-One Marine waskilled when the convoy he wasin received fire from a tank.

February 26-Three soldierswere killed and three woundedwhen their armored personnelcarrier (APC) was hit by ma-chine gun fire from a tank.

February 26-One soldier waskilled when his vehicle was hitby a premature burst of anartillery round.

February 26--Five soldiers werewounded when their BradleyFighting Vehicle (BFV) was in-correctly identified and hit by aTOW missile.

February 26--Two Ml Al Abramstanks were hit by fire fromanother M 1 Al tank. No casual-ties occurred.

February 26-Two soldiers werekilled and six wounded whentheir BFV, which was operatingin reduced visibility, receivedfire from a M1A1 Abrams tank.

February 2&Two BFVs, whileoperating at night in reducedvisibility, were fired upon by aMl Al tank. No casualties oc-curred.

February 27—Six soldierswere killed and 25 woundedwhen five M1A1 tanks andfive BFV’s engaging enemyforces were incorrectly iden-tified at night in reducedvisibility and engaged byother M1A1 tanks.

February 27—Two soldierswere killed and nine werewounded when three BFVswere fired upon by a M1A1tank because of incorrectidentification.

February 27—Three dam-aged M1A1 tanks were delib-erately destroyed by otherMl Al tanks to assure theycould not be used by theenemy.

February 27-One soldierwas killed and one woundedwhen 2 BFV’s were incor-rectly identified at night inthe rain and fired upon by aMl Al tank.

February 27-One soldierwas killed and two werewounded when two BFV’swere hit by fire from a M1A1tank while operating in rainand smoke at night duringan attack on a bunker com-plex.

February 27—Two soldierswere killed and two woundedwhen their BFV was firedupon by a Ml Al tank whileoperating at night in re-duced visibility.

February 27----One soldierwas killed and one woundedby machine gun fire whenthey were incorrectly identi-fied as Iraqi forces.

Air-to-Ground

January 23-A USAF A-10Thunderbolt fired on a Ma-rine observation post withno casualties.

January 24--One Marineand one sailor were woundedwhen a USAF A-1 O strafeda USMC Hummvee and afive-ton truck about 60 mileswest of Kafji, Saudia Ara-bia.

January 29-Seven Marineswere killed and two woundedwhen a USAF A-1 O fired aMaverick missile whichmalfunctioned in flight andhit a LAV,

February 2-One Marine waskilled and two were woundedduring an air attack by aUSMC A-6E using 500-pound bombs after their ve-hicles were incorrectly iden-tified as Iraqi.

February 2—Two soldierswere wounded when a HARMmissile fired by a USAFF-4G Wild Weasel did notproperly acquire its intendedtarget and locked onto thesoldiers’ radar.

February 4-A HARM mis-sile is suspected to havelanded close to the USSNicholas (FFG-47) result-ing in no casualties andonly superficial damage tothe ship.

February 17—Two soldierswere killed when a BFV wasstruck by a Hellfire missilefired from an AH-64 Apachehelicopter. Six soldiers werewounded and aground sur-veillance vehicle was alsodamaged in the incident.

February 23-One Marinewas killed and one woundedwhen a HARM missile froman undetermined sourcestruck a radar unit.

February 24-A HARM mis-sile is suspected to havelanded close to the USSJarrett (FFG-33) with no cas-ualties or damage to theship.

Ship-to-Ship

February 25--USS Jarrett(FFG-33) fired at a chaffrocket launched by USSMissouri (BB-63) resultingin superficial damage to USSMissouri. No casualties oc-curred.

March 27—USS Avenger(MCM-1) received small armsfire while in the vicinity ofRas Al Qalayah. No casual-ties occurred and the shipmoved out of firing range.

Ground-to-Air

February 15-A USN A-6Epilot reported he was firedupon by a surface-to-air mis-sile, resulting in no casual-ties.

SOURCE: Assistant Secretary of Defense (Public Affairs).


Africa in 1942 rather than the deserts of Iraq a halfcentury later.

In one barely avoided fratricide, a group oftanks was waiting for a second unit to catch up.Radio communication confirmed that all of thesecond unit’s forces were behind the frost unit.Two Iraqi T-55 tanks crossed in front of the firstunit, which quickly destroyed the enemy tanks.Just minutes later, two more armored vehicleswere detected, moving in the same direction asthe original T-55s. From consideration of thetactical situation, they obviously seemed part ofthe same Iraqi group. An alert tank gunnernoticed, however, that the vehicles showed on thethermal imager the characteristic ‘hot wheels’ ofU.S. infantry fighting vehicles and called out tohold fire. In fact, these were the scouts from theother units reported behind-but showing upahead of—the first unit and reported headingnorth but actually going west, which unfortu-nately was the same direction as the nearbyenemy force.61

Another case did not turn out so well. Twounits were traveling at night in parallel but not inconstant visual contact because of a gentle risebetween them. The units passed on either side ofan Iraqi infantry force armed with rocket-propelled grenades. The Iraqis fired at U.S.infantry fighting vehicles to one side. The explo-sions were seen by U.S. tanks in the other unit,which mistook the explosions for gun flashesfrom Iraqi tanks. The U.S. tanks returned fire andhit some of the U.S. infantry fighting vehicles.62

There were no fratricides of airplanes. Airsuperiority was so complete and accomplished so

quickly that very restrictive rules of engagementwere possible, which might have hampered theeffectiveness of the air arm but avoided anyfratricide.

SUMMARYA historian might wince at drawing lessons

from a collection of anecdotes, but some generalpoints come through. First, fratricide result mostoften from a complex chain of errors. The storiesoften read: identification was wrong, yes, butmisidentification would have been unimportant ifnavigation had been reliable, navigation errorscould have been overcome if communication hadbeen adequate, and so on. Also, these anecdotesmake clear that while misidentification oftenleads to fratricide, failures of command, commu-nication, coordination, and fire discipline are atleast as important. Although an accurate estimateof the overall frequency of fratricide is impossibleto determine, the two percent rule of thumbpresented by Schrader and others is almostcertainly too low. In every case in which gooddata are available, the actual rate of fratricideturns out to be much higher than two percent andhigher than most would guess. Finally, the typesof fratricide change much less quickly thanmilitary technology. This suggests that technol-ogy is only part of the solution; reducing fratricidewill always depend on the trainin g and Skills o fthe combatant in the field.

61 Center for ~y LMSOILS Learned, “Fratricide: Rduchg Self-Inflicted ~ss~,” Newsietrer (April 1992) No. 92-4, p. 1.

62 ~s ~~e is ~-bly sw to one ~cm at the Natio~ ‘Tr~g center, de~ri~ by GoMsmith. h tkit case the fkheS Were

simulated incoming artillery, but friendly units on either side took them to be gun flashes from enemy forces and returned fire.

—. —

I

AvoidingFratricide:

A

GeneralConsiderations 3

merica’s recent combat in the Persian Gulf Warbrought new attention to an old problem: fratricide.Twenty-four percent of all U.S. combat fatalities in thewar were caused by friendly fire.1 This figure seemed

much higher than in previous wars and caused a sudden focus onavoiding fratricide in future wars,

The U.S. military and the American public are becomingincreasingly sensitive to the human costs of military involve-ment, especially for contests of less than national survival. Forthis reason, the United States has invested much in energy andequipment to keep casualties low. As casualties from hostileaction decrease, the relative importance of fratricide increasesand fratricide should receive more attention.

The previous chapter found that only in a very few cases arereliable estimates of fratricide available but in each of thosecases, the fratricide rate was much higher than the nominal twopercent rate that frequently appears in the military literature,Fratricide has been, and probably will continue to be, asignificant source of combat casualties.

Moreover, the political and psychological cost of losses due tofratricide will always be greater than those due to losses inflictedby an enemy. In military operations involving allies, fratricidebetween countries can cause international friction at a time whenstrong cooperation is of utmost importance.2

1 This value derives from 35 friendly fire fatalities out of a total of 148 combat fatalities.

II

“Military Probes Friendly Fire Incidents, ” Office of Assistant Secretary of Defense(Public Afairs), Aug. 13, 1991.

2 During Operation Desert Storrq 9 British soldiers were killed when Maverickmissiles from A-10 ground-attack aircraft destroyed their armored personnel carriers,causing considerable political controversy in the United Kingdom. See Glenn Frankel,‘‘In Britain, Fallout from Friendly Fire, ” Washington Post, May 18, 1992, p. Dl,

29


Fear of fratricide has been a constraint oncombat tactics and maneuvers. The previouschapter described how lack of effective identifi-cation during both World Wars limited the use ofair power for close-in support of ground troops. Inat least one case, danger of fratricide is theprimary constraint on tactics: the joint use of airdefense fighters and air defense missiles in thesame area. Conversely, development of a good,reliable antifratricide system could open up newtactical options. Better identification could allowmore rapid attacks on enemy strong points, moreaggressive defense and covering fire by dug-insecond-line defenders as first-line defenders with-draw, closer and more agile air-to-ground orartillery support, and so on.

Fratricide becomes increasingly important notjust because of its relative increase due to thesmaller numbers of total casualties, but becauseof the way that the United States wants to keepthose numbers low. The U.S. military believesthat the best way to win quickly and decisivelywith least loss is to apply overwhelming fire-power against the enemy. However, if the onlypeople on the battlefield shooting are Americans,then it follows that the only way for Americans toget killed is from fratricide. Indeed, some simpletheories suggest that fratricide may blunt theadvantage of overwhelming advantage in number(or “force ratio”) so that the U.S. approach todecisive combat may require solving the fratri-cide problem to be viable.3

TRENDS IN THE FREQUENCYOF FRATRICIDE

Some incidents in the Persian Gulf show trendsthat may exacerbate the problem of fratricide asweapons’ development continues to advance.First, the tempo of battle has increased, often-times allowing combatants only seconds to makelife and death decisions. Second, engagement

ranges have increased. Mistakes of identificationwere difficult at the close ranges needed in the ageof sword, but many modern weapons’ range farexceed the range at which the human eye, or eveninstruments, can distinguish friends from foes.Also, the destructiveness of weapons has in-creased. In the past, fratricidal attacks couldsometimes be stopped if the mistake was realizedquickly, but now the first shot is often fatal,making an initial mistake irreversible. Finally, apotential problem often overlooked during theCold War is that future enemies may haveweapons similar or identical to those of theUnited States or its allies.

Other technical developments make avoidingfratricide easier. A British investigation early inWorld War II showed that among strategicbomber crews reporting that they had attackedtheir assigned targets, only one-fifth had actuallydropped their bombs within five miles of them.4

Absolute rates of fratricide may have peaked inWorld War II because the destructiveness ofordnance increased faster than the ability todeliver it precisely. Clearly, improvements innavigation, communication, and weapons-delivery accuracy improve the control of fire,making avoidance of fratricide easier (at least inprinciple). This hypothesis is supported by expe-rience in Operation Desert Storm where, contraryto past wars, there were no artillery fratricides.

Since estimates of past rates of fratricide havebeen unrealistically low, any telltale that allowsunambiguously attributing a casualty to fratricidecauses a jump in the number of visible incidents.This accounts for part of the picture coming outof the Persian Gulf experience. For example, onlyU.S. tanks were armed with depleted uranium(DU) antitank shells in the Persian Gulf. Theshells leave a small but distinctive and easilydetectable trace of uranium on any target they hit.Thus, after Operation Desert Storm, a quick testcould reveal clearly any fratricide caused by U.S.

3 hwd A. Wojcik, “The Manchester Equations in Defense Policy Analysis,” July 3, 1984, unpublished.4 John lhai.ne, A Time for Courage: The RAF in the European War, 1939-1945 @Jew York NY: Macmillaq 1985), p. 292

Chapter 3–Avoiding Fratricide: General Considerations 31

tanks. In World War II, the same poor communi-cation and navigation that might cause fratricidewould allow it to occur in the midst of theconfusion of battle without anyone, shooter orvictim, ever realizing it. Today’s improved com-munication could make fratricide less likelywhile also making it more likely to be discoveredwhen it occurs.

By any absolute measure, fratricide was notworse in the Persian Gulf War-or in Panama andGranada-than in previous wars. The fraction ofdeaths due to fratricide was apparently high, butthis was due in part to the very low total Americanfatalities from all causes and, in part, to pastunderestimates of fratricide rates.

MEASURES OF THE SEVERITYOF THE PROBLEM

Part of the challenge of evaluating any antifra-tricide effort is deciding on an appropriate meas-ure of how bad the problem is. On the one hand,the most common—and most public—measure isthe fraction of all U.S. casualties caused by U.S.weapons; that is, a comparison of the number ofU.S. casualties caused by U.S. forces to the totalnumber of U.S. casualties.5

On the other hand, some argue that U.S.casualties caused by U.S. forces are more appro-priately compared to the number of casualtiesinflicted on the enemy by U.S. forces, In the caseof the Persian Gulf War then, the dozens ofmistaken fratricidal attacks by U.S. forces should

properly be compared to the tens of thousands ofappropriate attacks on enemy targets.6

Comparing friendly fire losses to losses in-flicted on the enemy is probably more appropriatein cases of wars against comparable enemieswhere the outcome is uncertain, such as the ColdWar contest between NATO and the WarsawPact. In this case, relative rates of attrition willdetermine, in part, which side wins. In contests ofless than national survival, such as the PersianGulf War, final victory is less uncertain-if theNation is willing to pay the price. The question iswhat that price will be in lives lost. In these cases,casualties should be as low as possible andmilitary planners should address the causes ofcasualties in their order of importance. Thus, inthese cases, comparing fratricide to total friendlycasualties is the more appropriate measure.

Neither of these simple measures is adequateby itself. Avoiding fratricide is never the soleobjective of a military force; it must be balancedwith other military goals and efforts to hold downoverall human costs. Combat is inherently dan-gerous and casualties are inevitable, and some ofthose casualties inevitably will be due to fratri-cide. Moreover, some measures to reduce fratri-cide could be so stringent that they would reducemilitary effectiveness and, in the end, increase thecasualties inflicted by enemy forces.7 (See figure3-l).

Yet a death due to fratricide will never have thesame psychological effect as just another casu-

5 When possible, this report uses fatalities as a measure of comparison rather than total casualties because of ambiguities of defining‘‘wounded.

G For example, if the 35 Americans killed arc compared to the 20,000 or more enemy killed in Desert Storm, then the fratricide rate is wellunder one percent. See Center for Army Lessons Learned (CALL), “Fratricide: Reducing Self-Inflicted Losses, ” CALL Newsletter No. 92-4,April 1992, p. 5.

7 This is precisely what some alleged happened during the opening phases of the air war against Iraq, that over-stringent IFF requirementsallowed to go free an Iraqi fighter that later downed a U.S. jet. See Mark Crispin Miller, “Death of a Fighter Pilot, ” New York Times, Sept.15, 1992, p. 27 and New York Times, ‘‘Officer Says Iraqi Jet Downed Navy Plane During Gulf War, Sept. 15, 1992, p. 5. The Departmentof Defense is uncertain but believes that the airplane was downed by a surface-to-air missile. See the rebuttal letter from Pete Williams, AssistantSecretary of Defense for Public Affairs, New York Times, Sept. 26, 1992, p. 20.


Figure 3-l—Casualties and AntifratricideMeasures

~ \otallosses..I \ Increasing losses

\ from enemy fire > /

\ Decllnmg losses/

8 ‘,< from fratricide /ga \ /

3 \ /(uu ‘ \

/3 . /

F/

‘A

s p.——’”-. \- .— - -

INo antifratricide Too stringent anti-measure fratricide measures

>More stringent anti-fratricide measures


alty. The destruction of morale, esprit, andmilitary cohesion from a fratricide is far greaterthan that from a similar loss inflicted by anenemy. If fratricide’s importance is measuredsimply by its effect on the outcomes of battles orwars, it seems of little significance: the historicalreview found no cases in which a fratricidal errorclearly reversed the outcome of a battle. Butincidents of fratricide can cause soldiers tobecome too cautious, too timid, and too conserva-tive. These psychological effects may be intangi-ble but every experience of combat shows how

very real they are. In the end, military effective-ness is reduced. Thus, some military analystsbelieve that the secondary, hidden effects offratricide on the psychology of the survivingtroops may be greater than the direct effects oflosses of forces.

THE ATTACK SEQUENCEThe review of fratricide incidents shows that

very few result exclusively from mechanicalmalfunction; in almost all cases, fratricide resultsfrom deliberate-but mistaken-human decisionsand actions that cause casualties among friendlyforces.

The final decision to attack a target is the laststep in a multistep process. If, at any stage of thisprocess, the shooter could get information show-ing that his weapon is directed at friendly forces,then fratricide might be reduced.8

Attacking a target begins with the detection ofsomething. That ‘‘something’ might be an intui-tion, a slight movement among the trees, a blip ona radar scope or other sensor, or even incomingfree.

The next step is classification. The process ofclassification can itself contain several steps. Atthe moment of first detection, the observer istypically uncertain whether the small blob seenthrough the thermal sight is an enemy tank or alarge warm rock. Is the object a rock or a vehicle?If a vehicle, is it wheeled or tracked? If tracked,is it a tank or armored personnel carrier? If a tank,is it foreign- or U.S.-built? And if U.S. forces arefighting alongside allies, can the foreign-builttank be identified as an allied or enemy foreign-built tank?9

8 This discussion is based largely on the Defense Advanced Research Projects Agency briefing by Lt. Col. Joseph H. Beno, “BattlefieldIFF” (undated).

g Readers familiar with the intelligence and photointerpretation process should note that this sequence of identilcation is different fromthat used by photointerpreters of recomaisaince images. They assume that all the vehicles parked in the interior of an enemy country are enemyvehicles. First, they want to know whether the spot on the photograph is a vehicle, if so is it a tracked vehicle, if so is it an armored persomelcarrier or a @ and finally, what type of tank and what are its capabilities. IFP can be in some ways a much harder problem. Combatidentification is less concerned with the problem of whether a vehicle is a truck or a tank but in the potentially very hard problem of whetherthe truck is U.S. or enemy.

Chapter 3–Avoiding Fratricide: General Considerations I 33

Once a target is identified as enemy, a decisionto attack must be made. This decision is notautomatic. For example, an air-defense missileunit may clearly identify an aircraft as hostile, butthe aircraft may also be beyond the range of theunits missiles. An infantryman may see a enemytank but be armed with a shoulder-launchedrocket that is effective against armored personnelcarriers but not against tanks; a tank gunner maysee several enemy targets and decide to ignoredistant ones because those closer are more threat-ening.

Finally, after a decision to attack has beenmade, the attack must be carried out. Weaponsmust be aimed properly and ordnance deliveredwhere it is intended and not elsewhere, Withmodern weapons, ordnance almost always goeswhither the weapon is aimed; in the overwhelm-ing majority of cases, fratricide occurs when theweapon is aimed at the wrong place.

Separating the attack sequence—which mayonly take a few seconds—into these individualsteps may seem a more complex description thanneeded. But most fratricide are errors that couldhave been avoided if proper information had beenavailable at any one of these steps. Thus, whenfratricide is the result of a chain of errors, it couldbe avoided if the chain is broken at any link.Breaking a link in the chain requires that theshooter be given correct information about afalsely identified target.

THE COMBATANT’S SOURCES OFIDENTIFICATION INFORMATION

Three overlapping types of information affecteach step of the decision to fire or not to fire ona potential target. The first and highest level ofinformation is an overall, general knowledge ofthe tactical environment during the battle: wherefriendly units are-or at least supposed to be—and where enemy units are thought to be, plus theplan of action for the battle. The Army calls thisknowledge ‘‘situational awareness," an awkward

but useful term; other Services include it under“battle management.”

High-level, general knowledge is not adequatealone; people involved in the battle must alsohave specific information about whether anyparticular weapon or vehicle is friendly or enemy.This information is usually called the Identifica-tion of Friend and Foe, or IFF. The military callsthe synthesis of these two types of information“Combat Identification, ” or CID.

Connecting these first two sets of informationis another type of a priori information broughtalong to the battle: doctrine and rules of engage-ment. These rules tell those engaged in the battlehow to treat information from other sources. Inparticular, rules of engagement contain assump-tions about how to make decisions with imperfectknowledge; specifically, is an ambiguous targetassumed friendly until proven hostile, or assumedhostile until proven friendly? Different forcesunder different circumstances will use differentassumptions.

Because destroying a target is a multistepprocess and fratricide can be avoided by properlyintervening at any step, more than one approachcan be used to prevent fratricide. This must bekept in mind when comparing claims about theefficacy of various approaches. Those proposingsolutions that improve knowledge of the tacticalenvironment might claim that, say, 75 percent offratricide could be avoided by improving situa-tional awareness. While those proposing IFFsolutions might also claim that 75 percent offratricide could be avoided by improving IFF.Clearly, both systems will not eliminate 150percent of a problem, but both claims might betrue because either approach could reduce fratri-cide. One study using computer simulation ofland combat illustrates this point in an interestingway: fratricide was eliminated entirely if forceswere assumed to have either perfect tacticalknowledge-provided by hypothetical perfectcommunications and navigation equipment--orif they were assumed to have perfect IFF—


provided by hypothetical perfect sensors or IFFtransponders. 10

1 Knowledge of the Tactical EnvironmentA sense of the tactical situation on the battle-

field is so important to avoiding fratricide that itis sometimes taken for granted. A tank com-mander in a rear assembly area is surrounded byother potentially lethal tanks but does not eventhink of firing on them because of the firmknowledge that they are all friendly. This is a caseof almost subconscious tactical awareness.

As the likelihood of encountering hostile forcesincreases, tactical information provides clues tomake the search for targets more productive andhelps in the classification of targets as friendly orenemy. Rarely does a ground force expect anattack from any direction; typically, their knowl-edge of the battlefield suggests to them the likelydirection from which an enemy might approach.The tactical environment thus helps to classifypotential targets as most probably friendly orenemy. For example, a unit approaching from therear is first assumed friendly but a unit approach-ing from enemy-held territory is frost assumedhostile. Or, if a commander knows that his and allother friendly units have orders to advance in acertain direction, then a unit seen moving at rightangles to that axis will be assumed to be mostlikely hostile.11

Before these types of presumptions can beusefully reliable, however, each unit must haveconfidence that it, and its neighboring units, areunlikely to be heading in a wrong direction andthat wherever each unit is and whichever direc-tion it is headed, nearby units can inform eachother. This requires, in turn, reliable navigationand communication procedures and equipment.

Communication may be through some central-ized clearinghouse. Tactical air forces typicallyuse this approach. It allows the optimal allocationof resources across the entire theater of opera-tions. Since airplanes travel so far and so fast,centralized control is almost required to respondto enemy attacks and to avoid unintended encoun-ters among friendly forces. The disadvantage of acentralized system is that large quantities ofinformation must be transmitted up and down thechain from the controllers to the units in the field.

Communication may also be through localnetworks that link nearby forces. Ground units aremore likely to use this approach. It has theadvantage of minimizing the required flow ofinformation up and down the chain, Problems canoccur, however, since ground units report upthrough a chain to some central commander butalso need to communicate across the chain ofcommand to units that might be remote in thecommand network but happen to be geographi-cally close. Armies have long recognized thisproblem and have developed special proceduresto ameliorate it. The simplest approach is to makemajor command divisions correspond to realgeographic barriers-for example, rivers or moun-tain ridges—thus minimizing the need for com-munication across command lines. Failing that,command lines must be clear and special atten-tion must be given to liaison between adjacentforward units that report up through differentchains of command.

Navigation, like communication, can be globalor local. The pilot of a long-range aircraft clearlyneeds to know where it is in absolute terms, thatis, its latitude and longitude. World War IIexperience shows that when artillery fire wasmisplaced, the cause often was a forward observer

10 See I$fIT/Lincoln Laboratories briefing by A.B. Gschwendtner, “DARPA Combat Identitlcation Program at IWT/Lincoln Laboratory”(Aug. 28, 1992). In the extreme, perfect tactical awareness, that is, precise and complete information about the location of all nearby friendlyforces, would provide as one benefit the equivalent of perfect identification of friends and, by the process of el iminatiou foes. Thus, perfecttactical awareness is equivalent to perfect IF’F.

11 Rec~l tie case described inch. 2 from Operation Desert Storm in which lost U.S. armored personnel carriers were assumed tO k @i

because they were cutting across the general direction of advance.

— .— ———— -. —

Chapter 3—Avoiding Fratricide: General Considerations 35

correctly calling in fire relative to his position buterroneously reporting his absolute position. Inmany if not most cases, however, position relativeto nearby friendly forces is good enough or evenbetter, Perhaps only the leader of a tank platoonneeds to know his absolute position, while the restof the tank commanders are primarily interestedin local terrain features and the tanks’ positionsrelative to each other.

1 Rules of EngagementThe criteria for deciding on the nature of a

target and the response to it are called “rules ofengagement. Military commanders, theorists,historians, and tacticians have long recognizedthe chaotic confusion of combat. No battlefieldparticipant has perfect knowledge, and criticaldecisions must be made, sometimes quickly, withincomplete and occasionally flatly contradictoryinformation. Of necessity, every combatant entersa battle with some a priori knowledge in the formof a set of decision-making rules to use underthese difficult conditions. This a priori knowledgeis learned through training, exercise, and indoctri-nation. At the tactical level, military doctrineprovides rules or criteria that are used to decidewhen to attack, when to defend, and so on.Fratricide is affected more by some of the rulesused by each individual combatant to determinewhether an unidentified target is hostile orfriendly, Targets are rarely classified by appear-ance alone but also by location, behavior, andrecent experience. Thus, the combatants’ tacticalenvironment affects strongly their decisions aboutthe danger posed by ambiguous targets.

Two kinds of identification errors are possible:friendly forces or neutrals can be mistakenlyidentified as hostile, and hostile forces can bemistakenly identified as friendly or neutral. Rulesof engagement will also depend on a comparisonof the consequences of each of these mistakes. Forexample, interceptor aircraft pilots typically usedifferent rules of engagement in different tacticalsituations because of the differences in the

consequence of each type of mistake. The differ-ences are clearest when comparing point defenseof an extremely valuable target and defense of alarge area.

For example, Navy fighters fly from an aircraftcarrier, which is a single asset of enormous value.The loss of the carrier implies the potential loss ofall the aircraft and crew on the carrier. Carriers areheavily defended, but even a single attackergetting through can do substantial damage, per-haps disabling the carrier for an extended period.Moreover, the carrier is a‘ ‘point” target, makingthe geometry of an attack unusually clear, Thus,the Navy’s rule of engagement during hostilitiesis that an aircraft approaching the carrier must beable to prove that it is friendly, or at leastnonthreatening, or else it is assumed to be hostile.

Tactical interceptors trying to defend a largearea face a very different situation and respondwith different rules of engagement. For example,the classic defense of NATO would have in-volved a large and confusing array of fighters overthe battlefield. Air defenses of NATO weremultilayered. The first ‘ ‘defense” was attackagainst enemy air fields. Enemy intruders wouldthen have to pass through a screening force ofdefending interceptor fighters. These aircraftmight be backed up by a band of defensescomposed of surface-to-air missiles (SAMs). Anyintruders that survived could be attacked bydefending fighters specifically left prowling inrear areas to catch these ‘‘leakers. ’ Finally,particularly valuable targets would be defendedby their own local air defenses of short-rangemissiles or guns, Thus, the battle would include aconfused and dense air traffic, with fighterscoming from many places and heading towardmany targets. Most of the defended combatassets—for example, Army assembly areas—would be more dispersed than an aircraft carrier,so that a successful attack on any one asset wouldbe far less critical than a similar attack against acarrier.

The consequences of firing on a friend are assevere in the case of area defenses as in the case


of defense of a valuable point target-a friendlyaircraft gets shot down-but the consequences ofpassing up an opportunity to shoot at an enemyare much less severe—if the first layer of defensesdoes not get an intruder then the next will, and soon, until he approaches the target when he finallymust make his intentions clear. Therefore, the ruleof engagement for unknown aircraft during areadefense is different from that used by the Navywhen defending a carrier: area defenders assumethat an unknown aircraft might be friendly unlesssome positive evidence is available to show thatit is hostile.

Ground combatants have a slightly differentperspective on the consequences of the two typesof misidentifications. They tend to think of theproblem as a comparison of the consequences ofeither shooting or not shooting. The situationfacing a tank commander can provide a specificexample. Historical evidence from World War II,training exercises, and testing show a substantialadvantage to the tank that shoots frost in tank-on-tank engagements. To hesitate is to risk destruc-tion. And the loss is not a vague, difficult-to-quantify overall loss of combat capability, but aloss that spurs the very compelling motivations ofself-preservation: the tank that fails to aim andfire first will itself be receiving the fire that mightcause the crew’s death. Thus, ground combatants’rules of engagement frequently are closer to thoseappropriate for defense of a valuable point target;indeed, the shooter is his own ‘‘high-value”target. A potentially threatening unknown targetis generally assumed hostile unless there is someevidence that it is friendly.

The danger that a shooter perceives also affectshis incentives to wait or shoot; the greater theperceived threat, the greater the urgency to shootfrost and the more likely a shooter is to assume anambiguous target is hostile. This dynamic isrepresented in figure 3-2. The vertical axis

i represents the shooter’s estimate that the target heI sees is an enemy. His estimate can range fromI zero at the bottom of the axis, that is, absolute

confidence that the target is a friend, to 100

100”/0

z’a)cal(n

5pga-c5%E.-6w

0“/0

DON’T FIRE

Low Perception of danger High>

SOURCE: Office of the Assistant Secretary of Defense (C31).

percent at the top of the axis, that is, absoluteconfidence that the target is an enemy. (Note thatthese estimates of classification confidence referto the shooter’s perception; these perceptions, inturn, may or may not be correct.) The horizontalaxis represents the shooter’s motivation to firebased on his perception of danger and urgency. Itcan range from a minimum motivation, repre-sented on the left-hand side, to maximum motiva-tion, on the right-hand side.

The motivation to fire will change due toseveral factors, but the most important variable isthe shooter’s estimate of his own danger. Thegraph lays out the different regions that result indecisions to fire or not fire. If the shooter feelsreasonably secure, then he is willing to wait tomake absolutely certain that the target he hasdetected is, in fact, an enemy. Thus, the “free”area of the graph occupies only the upper comer

-.—.— —. —


of the left-hand side, where confidence is high.Toward the right-hand side of the graph, theshooter’s perception of danger increases and hebecomes increasingly “trigger happy” until fi-nally, at the extreme right-hand side, where hefeels in immediate grave danger, the shooter mayneed high confidence that a target is not an enemyto keep from shooting. That is, ambiguous targetswill be assumed hostile until proven friendly.

The effect of a shooter’s sense of danger on hisidentification accuracy has, in turn, an indirect butinteresting effect on the relationship of targetrange and identification accuracy. That increasedweapon range makes IFF more difficult is acommonplace. The development of long-rangeweapons, especially guided weapons not requir-ing any form of forward spotting, has increasedthe range of engagement beyond that at whichtargets can be identified, thereby inviting mis-taken attacks on friendly targets. Some militaryanalysts have suggested that reliable IFF range bemade a design constraint on weapon range.12

Some data, for example, from tank combattraining exercises at the National Training Center,show a clear relationship between range andlikelihood of fratricide. Accidents are less likelyat shorter engagement ranges, with an importantexception: at very short ranges when the rate offratricide engagements is again very high.13 Thisphenomenon occurs because, all else being equal,identification becomes easier at shorter ranges—reducing fratricide-but all else is not equal.Engagements occur at very short ranges onlyunder the desperate and chaotic conditions of aclose-in melee. The effect of greater ease ofidentification is overwhelmed by the greater

Figure 3-3—Fratricide and Engagement Range

I

/

L-------- -–-–--T-- ---- ----——–—- ---10 500 1,000 1,500 2,000 2,500 3,000

Range (meters)

SOURCE: U.S. Army Training and Doctrine Command.

stress and confusion characteristic of close com-bat. See figure 3-3.

Similarly, air defense training shows that thehighest frequency of identification errors occursamong those manning the shortest range weap-ons. A Vulcan gun crew cannot engage anapproaching jet aircraft until it is virtually on topof them, thus forcing an ‘‘us-or-them’ approachto target engagement decisions. Increased weaponrange certainly is not the solution to fratricide, butneither is its effect entirely negative: it mightmake quick identification more difficult, but italso can allow more time for decision andevaluation.

The rules of engagement include other tacticalinformation. Some is explicit. For example, if afighter plane determines from, say, detection ofan enemy radar frequency that an approaching

Iz some ~Wents ~ tie debate about the usefihess of increasing weapon range have gone so f= ~ to suggest ‘it ‘i-t@ti ‘iSstie ‘we

ought to be tied to accurate target identification range such that targets at greater ranges are impossible to engage. For example, horn acongressional report on IFF: “Correcting this imbalance between missile and ID capabilities now requires an accelerated, closely coordinated,interservice and NATO-wide effort to concentrate resources on achieving the needed identification capability. A concomitant slow-down intactical missile acquisition could support that effort. ’ U.S. House of Representatives, Committee on Government Operations, “Identificationof Friend or Foe in Air Warfare-A Capability Imng Neglected and Urgently Needed’ (Washington DC: U.S. Government Printing Office,Nov. 1, 1985), p. 3. However, an artificially imposed constraint on range persists even when identifkation presents no problem due to, forexample, the clarity of the tactical situation.

13 See “CALL Fraticide Study” (undated), Center for Army hSSOILS Learned (C~), pp. 8-9.


aircraft is hostile, other aircraft flying in the sameformation are also assumed hostile. (This isinformally called “guilt by association.”) Thisassessment is usually accurate, but scenarios inwhich a friendly fighter is in hot pursuit of anenemy are also easy to imagine. Some other“rules” are implicit and seem to result fromhuman nature, or can be understood in terms ofthe effects of an increased perception of threat—

whether real or not-discussed above. Analysisof training exercises shows clearly that a gunner’srecent experience strongly influences his judg-ments about friend and foe. If, for example, an airdefense unit has just been attacked by “hostile”aircraft in a training exercise, then the nextairplane that flies over is much more likely to bejudged hostile than it would have been if theprevious overflight had been by friendly aircraft.Gunners are not, of course, taught to make theseprejudicial assumptions, and this is not an explicit‘‘rule,’ but few will be surprised by the observa-tion.

IDENTIFICATION: FRIEND OR FOE‘‘The first requirement in warfare is the ability

to distinguish friend from foe.”14 This statementfrom a World War II field manual makes clear theimportance-but not the difficulty--of IFF. Eventhe most straightforward technique, looking at apotential target with human eyes, is neithersimple nor reliable; combatants need training toidentify forces quickly. Even with training, mis-takes that appear egregious in the calm ofpeaceful retrospection are all too common in theconfusion of combat. The great increase inweapon range-made possible primarily by de-velopments in weapon guidance systems sinceWorld War II-has in many cases far outstrippedthe ability of human observers relying just ontheir eyes. IFF today requires additional informa-tion from longer range sensors.

Different weapon systems have different re-quirements for IFF systems. The IFF needed forshort-range and long-range weapons, for exam-ple, may have very different reactions times,abilities to track multiple targets, likelihood ofrevealing position, and so on, as well as theobviously different requirements for range.

~ IFF and the Rules of EngagementDifferences in the rules of engagement will

shift emphasis among technical approaches toIFF. There is no technical solution that is optimalfor providing positive identification of friend andfoe, where “positive identification” means iden-tification based on positive presence of someevidence. For example, response to a reliable IFFinterrogation system is positive evidence that thetarget is a friend, but lack of response is notpositive evidence that it is an enemy.

In defense of critical point targets, aircraftcarriers being the premier example, defendersmust assume that ambiguous targets are poten-tially hostile until proven friendly. Thus, theemphasis is on systems that can prove a targetfriendly, such as cooperative question-and-answer systems. The rules of engagement for areadefense will give relatively greater emphasis topositive identification of foes, hence, noncooper-ative foe identifiers will be relatively moreimportant with question-and-answer systems pro-viding frost-order sorting of targets into friendsand unknowns.

Ground forces, when under extreme pressure tofire quickly, also find themselves in a situationwhere ambiguous targets must be assumed hostileuntil proven friendly. Thus, ground forces willconcentrate on developing question-and-answerfriend identifiers. (In addition, active IFF byground forces is in a primitive state compared tothat of air forces and question-and-answer sys-tems, being easier to implement than noncoopera-tive systems, are a good place to start.)

14 Recognition Picton’ai Manual, War Department Field Wutd 30-30 (June 1943), P. 1.

Chapter 3-Avoiding Fratricide: General Considerations I 39

Figure 3-4--Sources of Information for Identification

/\Direct ,,/” “\ Indirect

“\,,,//‘ \

7I

Passive,, , Active}

OBSERVER‘ \/ ‘.

/, ~., J9 *

passive ‘ , Active ,/’ ‘\,,/ ‘

% Passwe ,’ “1 Active/“ / \// /,, ‘,

I I● )) TARGET

/’ \\

Cooperative ,’”‘.\ Noncooperative

/ \/’ ‘\ ‘\

SOURCE: Office of the Assistant Secretary of Defense (C31).

I Sources of IFF InformationSuccessful IFF requires that the shooter have

information--carried by some form of energy—from the putative target. The sources of thisenergy and how it is collected can provide ageneral, overall classification of the varioustechnical approaches to IFF. Figure 3-4 shows thepossible combinations of energy and informationsources. The first branch is between direct andindirect collection. A ‘‘direct’ system is one inwhich the shooter collects the information abouta target, while in an ‘indirect’ system some otherobserver collects the information and passes it onto the shooter. An artillery forward observer is agood example of indirect IFF. The forwardobserver, or spotter, uses whatever technique hehas to determine whether a target is friend or foeand then radios that information back to theartillery, which accepts it as accurate without anyability to confirm the information. The rest of thedirect and indirect branches are identical to eachother, so only the direct branch is shown in thefigure.

The observers may be passive or active.Passive observers do not transmit any energy

themselves but only collect energy normallytransmitted or reflected from the target. An activeobserver transmits energy at the target to some-how affect the target in a way that can beobserved.

The target as well may be either passive oractive. A passive target only reflects energy fromits environment. This energy may be natural, likesunlight, which a passive observer can detect or,in the case of an active observer, the energy mightbe artificially produced specifically to be directedto and reflected back from the target, as is thecase, for example, with a radar beam. An activetarget transmits its own energy, for example,radio signals or sound.

Thus four possible observer/target energy trans-mission routes are available: passive/passive,passive/active, active/passive, and active/activeshown in figure 3-5. An example of the first caseis the simplest IFF system imaginable: a passiveobserver identifying a passive target by sensingreflected sunlight. Or, a passive observer mightalso sense energy that is actively transmitted bythe target. This could be radio transmissions in aforeign language, or radio or radar transmissions

.-


Box 3-A-The Electromagnetic Spectrum

Chapter 3 discusses in general terms-and the next two chapters discuss in much more specific detail--someoft he identification and communication techniques important for avoiding friendly fire. Many of these techniquesdepend on detecting some form of electromagnetic radiation. The figure 3-A represents the electromagneticspectrum. Visible light is the most familiar part of the spectrum but radiation at higher and lower wavelengths isexactly the same physical phenomenon, just not detectable by the human eye.

Radiation of different wavelengths interacts with matter in different ways, which has two importantconsequences: first, the wavelength determines how easily the radiation passes through the atmosphere,including obscurants, such as fog, smoke, rain, or dust; second, the wavelength affects how the radiation isgenerated and detected. In general, radiation is little disturbed by particles smaller than the radiation’s wavelength,thus longer wavelengths tend to pass more easily through airborne particulate like smoke or fog. Infraredradiation--with a longer wavelength than visible radiation--can be used to see through clouds of obscurants thatappear impenetrable to the human eye and even longer wavelength radar waves can pass through rain that wouldstop infrared radiation.

at a characteristic frequency not used by friendlyforces, or the sound of a tank that is different fromthat of friendly tanks.

The observer could be active and transmitenergy to the target. Most commonly this is in theform of radio signals of some sort. Even if thetarget is passive, the shooter can still bounce radarsignals off it—this is standard radar. Thus anactive shooter can detect the radar return of apassive target and get some information about the

Figure 3-5—Examples of Approaches to IFF

TARGET

Passive Active

Visual ID of Beacons:weapons type, e.g., Budd-lights,painted insignia DA RPA-lights

Question-Radar and-answer

systems

L

SOURCE: DoD and OTA.

target that might allow it to be classified as friendor foe.

Alternately, the target might actively transmitenergy in response to the shooter’s transmission.In such a system, the shooter sends a specialsignal to the target. If the target wishes to beidentified, and is equipped with the propertransponders, then it sends back a signal signify-ing that it is a friend. This is the principle of‘‘cooperative’ IFF systems, such as the MARKXII. These are often called ‘‘question-and-answer” systems.

Finally, each branch of the tree shown in thefigure can be cooperative or noncooperative.Painting special insignia on a vehicle for others tosee is an example of a passive/passive cooperativesystem. Normal radar is an active/passive nonco-operative system, but if the target adds specialradar reflectors to enhance radar echoes, it be-comes an active/passive cooperative system. Justbecause a target gives an active response to anactive observer does not imply necessarily thatthe target is cooperative; it might be tricked intoresponding. Military electronics experts are al-ways trying to find ways to cause an enemy totransmit energy revealing his position, intentions,or capability.

Chapter 3—Avoiding Fratricide: Genera Considerations 141

Longer wavelength radiation may be able easily to penetrate the atmosphere but another problem arises aslonger wavelength radiation is used: the ultimate resolution of any imagining device is limited by the ratio of thewavelength and the size of the imaging optics. Radars using radiation with waves several centimeters or meterslong may easily penetrate long distances, but to achieve any resolution requires proportionally large antennas.Thus, long wavelength radiation may be useful for communication and detection but is not much use for derivingan image.

Several parts of the spectrum are of importance to the problem of avoiding fratricide. Communication usuallyuses UHF and VHF radio bands. The radar bands shown are used to detect objects and perhaps identify them,The infrared bands are important because ground forces use them for seeing at night and under conditions oflimited visibility. With increasing wavelength, infrared blends into the millimeter wave bands that will be used bythe proposed question-and-answer identification system for ground combat vehicles.

The Electromagnetic Spectrum

Frequency (Hz) 1072 1 o?~ 1 ().8 1 o’~ 1014 1 o’? 1 o’~ 1 Oa 1 OF 10’ lo~ 1

Wavelength (m) 10 ‘3 10 “ 109 10’ 105

0.4 0.9 1.1 8 10 12

I

Visible Iight ‘

103 0.1 10 103 1 o’ 1 o~

Nominal frequency ruu~cpprpy-CDbm

range (GHz) Li.3~r’dO-J03

Wave lengths (microns)

I I I {Sensitive range of IR lmaglng/targeting

night vision goggles viewer


I Advantages and Disadvantagesof Each Approach

Each approach to IFF has comparativestrengths and weaknesses. The military prefers, atleast as an ideal, systems that allow the shooter toremain completely passive because these poseleast risk of revealing information about theshooter. But passive systems must exploit subtledifference in emissions, that is, the “signature,”of a vehicle. Keep in mind that the objective is notto detect a tank but to distinguish one type of tankfrom another against a very complex battlefieldbackground, and in spite of possible efforts by theenemy to hide or alter their weapons’ appearance.Thus, purely passive systems may require sensitivehence expensive—sensors. Most IFF expertsinterviewed by OTA felt that no single passivesensor would be adequate but that the fusedinformation from several may have future appli-cation.

Active IFF systems—the most obvious exam-ples are radars and question-and-answer systems—can have longer range but they, of course, mightbe detected, providing the enemy informationabout friendly forces. This risk can be minimizedby transmitting the least power necessary, trans-mitting intermittently, using and looking for1special transmission patterns (or “waveforms’known only to other friendly forces, and so on.

In general, cooperative systems are cheaper perplatform and have longer range; it is always easierto get information out of a target if it wants thatinformation to get out. Cooperative systems havethe cost disadvantage, however, of becominguseful only when virtually all of the platformswithin a theater of action are equipped, makingpartial deployments unattractive.

Clearly, cooperative systems are really justfriend identifiers, they do not positively identifyenemy. If no reply is received, the shooter mightassume that the target is enemy but perhaps it isa neutral or a friend without an operating trans-ponder. The final classification in these cases willdepend on the rules of engagement.

Noncooperative systems can, under some cir-cumstances, positively identify enemy. For exam-ple, classification could be based on type ofweapon. This approach was easy and reliableduring the long NATO-Warsaw Pact confronta-tion; NATO forces had NATO weapons andWarsaw Pact forces had Warsaw Pact weapons.Thus, the type of weapon identified it as friend orfoe. This approach can, with luck, still work butin general the world is much more complex today.For example, in the Persian Gulf War, Syrianallies were armed with the same Soviet tanks asthe Iraqi enemy while the Iraqis also had the sameFrench-made aircraft as the French allies. Thisprofusion of weapon types does not make nonco-operative IFF impossible, only more subtle anddifficult. A weapon is much more than the shapeof an outside shell; for example, the same type ofaircraft in two different air forces may usedifferent engines, have different weapons mounted,and have different radio or radar frequencies.Noncooperative IFF may even try to exploit thehabitual tactical behavior of the enemy.

Under some circumstances noncooperative sys-tems may have a cost advantage even when costsper platforms are higher since the anti-fratricidebenefit is roughly proportioned to the extent ofdeployment in contrast to cooperative systems,which have little benefit until deployment isvirtually complete within a combat group.

Differences in rules of engagement push to-ward different technical solutions to IFF. As wasdiscussed before, area defense rules of engage-ment really require an enemy identifier since,without some positive evidence that a target isenemy, it will not be engaged. Thus, in theextreme, with no way to identify enemy, theywould not shoot at anything. A cooperativequestion-and-answer IFF system identifiesfriends, but that provides little additional help tothe interceptor pilot: he starts off assuming thatthe unknown target is potentially a friend. Thepilot needs some system that can identify anenemy as an enemy; thus, the need for noncooper-ative IFF system.


In contrast, a Navy pilot returning to a carrierhas to convince the carrier’s defenders that he isa friend. Anything that identifies enemies doesnot give much additional useful informationbecause the defenders start off assuming thatunknowns are enemy. Thus, the Navy needs afail-safe friend identifier and concentrates itsattention on cooperative question-and-answer IFFsystems.

I Information SecurityA potential danger with any question-and-

answer system is that the enemy might exploit it.Of course, the occasional transmission of a querymight reveal the shooter’s position, but two evengraver failures are possible. First, if an enemycould receive the query and then respond with theproper answer to make himself appear to be afriend, he could then penetrate defenses easily andcause great damage. Second, if an enemy couldproduce the proper query, he could fool friendlyforces into responding, revealing both their posi-tions and their identity.

Securely encrypting the query and the answer15 An enemy t r y ing toeliminates both dangers.

exploit a question and answer IFF system couldtransmit queries to try to have U.S. weaponsidentify themselves, but if an encrypted message,which only U.S. or allied forces knew, were usedas the query then friendly forces would simply notrespond to any queries that were not properlyencrypted. But what if the enemy could not createa valid query of his own but at least couldrecognize that it was a query? The enemy couldrespond with an answering signal to make himselfappear to be a friend. This is called ‘‘spoofing,

Again, if replies were encrypted and only friendlyforces knew the encryption technique, then theenemy might try to reply with some signal but itclearly would be counterfeit.

To be genuinely secure, an encrypted messagemust be sufficiently complex to foil enemy effortsto figure out a way to create a valid query andreply. Additionally, even without understandingthe encryption method, an enemy could simplyrecord valid queries and then retransmit them.Therefore, the encrypted form of the query mustbe changed frequently. The combination of thesetwo requirements creates a burden of generating—and distributing to forces in the field-encryptionkeys and other materials. Transport and dissemi-nation of these materials is further complicatedbecause they must be protected during distribu-tion; if the encryption keys fell into enemy hands,then the IFF system would be compromised untilnew keys could be created and distributed.16

Some front-line forces in Europe, such as Armyair defense units, claim that handling of theclassified IFF encryption keys is so onerous thatproper training with IFF systems is stifled.17

New technical developments allow easier han-dling and dissemination of secure informationlike cipher keys. Systems are under development,called electronic key management systems, thatwould allow the electronic dissemination ofpartial keys to regional distribution centers, andthither to a “Local Management Device” con-taining a “Key Processer. ’ At the local level a“Data Transfer Device” is used to carry the keyto the individual weapon. The partial keys alonewould not allow any enemy to decrypt messagesor IFF queries and answers. Instead they would

15 me words t ‘code” ad “cipher” ~d, sifil~ly “CnCOde’ ad ‘ ‘encrypt” are often used interchangeably h the nOnteC~Ca.1 literature.Experts, however, make a very clear distinction. Code is any set of symbols that represents something else and mayor may not be secret. Thus,“20500” is the ZIP code for the White House, but anyone that calls can get the code. If a message is encrypted, then anyone who knows thecypher can read the code but no one else. Sec Nationat Security Agency briefiig, ‘‘Cryptography and Security for Combat ID Systems, ” May6, 1992, p. 5.

IS See Nation~ Security Agency briefing, ‘‘The Electronic Key Management System (EKMS)” (undated), pp. 34.

17 Monti Callero, “Combat Identification and Fratricide Analysis, ” a briefing presented at the Modeling and Simulation Work.rhop onCombaf Idennjicafion sponsored by the Defense Advanced Research Projects Agency (D#@PA), Nov. 4, 1992. In fact the keys we onlyclassified ‘‘cotildential; part of the problem may be more the users’ perception and lack of motivation,

.. —- ——


provide an input to a computer that wouldgenerate the actual keys. Thus, the partial keyscould be transmitted over the open airwaves forall to hear. The local computers could load theactual keys directly into the IFF question-and-answer transmitters so there would be no need tohave the keys written on easy-to-steal paper andno person-or potential spy—would even need toknow what the key was.

Advocates of noncooperative IFF systems areready to point out the burden of handling encryp-tion keys, but noncooperative systems have theirown data-handling challenges. Noncooperativesystems, especially ones that rely on a passivetarget, must have available stores of informationon each weapon that might be encountered.Identifying aircraft from their radar return, forexample, would require detailed informationabout the radar returns from dozens of aircraft,each with several different configurations orweapon loads, from all possible perspectives.This information might take the form of databasesin which the characteristics of enemy aircraftwould either be looked up or calculated quicklyfrom models, but whatever approach is used thedata required could be substantial.

Moreover, these data need to be protected justas much as encryption keys. Not every detail of aweapon’s signature will be used for identificationand loss of the data would help an enemy discoverwhich particular attributes were being used.Updating data or adjusting for compromisedinformation during combat could pose severe datadistribution problems. Thus, using a noncoopera-tive approach to IFF will alter but not remove therequirements for data security and handling.

Compiling the databases required for noncoop-erative IFF raises delicate diplomatic and intelli-gence questions. Would allies be willing toprovide sufficient data on their weapons? (Wouldthe United States be willing to provide compara-ble data to allies?) If not, would surreptitiousattempts to collect data cause international fric-tion? Would the very existence of data collectedsurreptitiously need to be kept secret from allies?

How would information on enemy weapons becollected? If weapons data are available, couldthey be shared with allies during joint operationswithout revealing intelligence methods?

Whether using cooperative or noncooperativeIFF systems, operations with allies pose specialproblems. In the new global strategic and politicalenvironment, the United States will almost al-ways be engaged militarily with allies at its side,for political if not military reasons. Moreover,unlike the relatively stable allied relations inNATO, future alliances are more likely to be adhoc, with even former-and perhaps future--unfriendly nations acting as temporary alliedpartners. The Persian Gulf War, which includedthe participation of Syrian forces, provides aperfect example.

Clearly, it is in the interest of the United Statesto help allies not to commit fratricidal attacks onthemselves or on U.S. forces. But the need forboth protection of identification techniques andallied cooperation creates dilemmas.

For example, encryption keys needed for question-and-answer IFF systems would have to be sharedwith allies. But today’s allies might be hostiletomorrow. At the very least, the United Statesmust be able to make any system secure again bychanging the encryption keys. Thus, a temporaryally would not have irrevocable ability to triggerIFF systems on U.S. weapons.

Encryption keys that fell into the hands of othernations can be changed quickly-and thus ren-dered useless—but the same is not true forhardware and technology. Cooperative question-and-answer systems require that allies have compati-ble equipment. Presumably the United Stateswould have to provide the equipment or at leastexplain the encryption systems in enough detailthat other nations could produce their owncomparable systems. Yet the United States hasinvested significant resources to develop reliableand secure encryption systems because expertopinion holds that the resulting capabilities offerprofound tactical and strategic advantages. Eventhough keys are changed, if the equipment is left


behind, other nations can generate their own keysand use the equipment for their own IFF with allof the tactical advantages that accrue.18

The United States must decide whether thebenefits of having a unique technical advantage inIFF outweigh the benefits of allied interoperabil-ity. Without exchange of IFF technology, geo-graphic areas of responsibility would have to beclearly divided among allies, with each responsi-ble for fratricide avoidance within its area andspecial procedures used to reduce accidents at theboundaries between areas. This problem becomesparticularly important in ad hoc coalitions; theUnited States may be willing to share technologywith NATO allies that it is unwilling to share withtemporary allies. The ideal solution might be atechnically less sophisticated, or ‘stripped down,’IFF system that could be shared with some alliesand was compatible with, but not have all theadvantages of, the complete system.

MAKING CHOICES ABOUT HOW TOAVOID FRATRICIDE

B Criteria for Judging AntifratricideTechnology

Any successful antifratricide system must meetseveral criteria. First, and probably of greatestimportance to the combatants in the field, is thatusing the system must not significantly increasethe users’ danger. For example, transmitting anIFF query in a question-and-answer system mustnot give the enemy a chance to intercept the radiosignal, thereby giving away the shooter’s loca-tion. A second, and related criterion, is thedifficulty of using the system. If it is too complexor time-consuming, then it will not be used at all.The whole process of using the IFF system mustnot take so long that the enemy can shoot firstwhile friendly forces are still fiddling with knobs.

A third criterion is how widely applicable thesystem is. Can it tell the shooter something about

everything on the battlefield from infantrymen tojet fighters? Or is the system limited to thoseplatforms that can carry a large and expensivetransponder? A fourth criterion, related to thethird, is the system’s reliability. This is more thansimply a question of how long a piece ofequipment will work without maintenance, but ofhow well it will work under a complex—andultimately unpredictable--array of battlefield con-ditions that might include active measures by theenemy to undermine or exploit the system. Forexample, how reliably can a noncooperative IFFsystem identify an enemy if the enemy knowswhat characteristics are being observed and usedfor classification and then tries to alter those verycharacteristics? Required reliability also dependson how an IFF system is used. Human eyes andcommon sense avoid most potential fratricide; ifanother system is used in addition to currentprocedures as one last check, then 90 percentreliability is adequate, but if used instead ofcurrent procedures, 90 percent reliability wouldbe disastrous.

The fifth criterion is cost, both total costs andcosts compared to alternative requirements. Inabsolute terms, total cost is important becauseresources are finite and combat can never be madesafe so decisions must be made about when asystem is “good enough. ” Consideration ofcomparative costs is potentially less judgmentaland more analytical. If the objective is to savelives while maintaining combat effectiveness anda better IFF system is just one way of doing that,then within a finite budget every dollar spent onIFF is a dollar not spent on other things that mightalso save lives, Perhaps casualties-includingthose from fratricide--could be lowered moreeffectively by using the same resources to buyreactive armor for tanks, or for more training, andso on. The relative importance of a fratricidecasualty and a hostile casualty-which willdetermine the allocation of marginal resources—

18 Natio~ sec~~ Agency bfiefkg, “Smey of Target Identiflcatioq ” Sept. 9, 1992


is a difficult policy question that will not just behard to decide but hard to discuss dispassionately.

1 Comparison of Different Approaches ToAvoiding Fratricide

Awareness of the Tactical Environment v. IFFThe main two broad approaches to preventing

fratricide are increasing tactical awareness andimproving target identification. Either approachoffers benefits, each has areas of particularimportance, and both will need to be pursued inparallel.

The differences are clearest for ground combat.In general, as ground units approach the enemy,knowledge of their tactical environment is mostimportant to avoiding fratricide. This means thatunits have the navigational information needed toknow where they are, know where they aresupposed to go, and not stray across avenues ofadvance. Nearby units need to communicate toavoid confusion along their boundaries.

As the battle is entered, one-on-one identifica-tion becomes increasingly important. In theextreme of a melee, the tactical situation couldbecome so confused and fluid that no personcould keep up with the changes even if thetechnology were available to report them. In this

I situation, IFF would be of greater importance,(See figure 3-6)

When comparing costs of the two approaches,the multiple benefits of each must be weighed inthe balance. A reliable point-to-point IFF systemwould be valuable for IFF but since identificationrange can set the limit of effective engagementrange, it might also allow engagements at longerrange. Any system that improved the knowledgeof the tactical environment would reduce fratri-cide but also allow better control of maneuvergroups, better coordination of fire and combatunits, faster attacks, more efficient movementacross country, and so on, thereby increasing theoverall combat capability of the force. Thus,when comparing costs of reducing fratricide, onlythe appropriate portion of the cost of each system

Figure 3-6-Changing Relative Importance ofTactical Information and Identification

IFFA

~>.—=5

1 Tactical informationv

\

Deliberate Meldeattack

SOURCE: Combat ID System Program Office.

should be allocated to the antifratricide require-ment.

Training and Simulation

The solution to the problem of fratricide is notjust black boxes; training and combat skill are atleast as important. Of course, the skills needed toavoid fratricide-alertness, unit coordination, anddiscipline-are exactly those needed for any unitto be effective in combat. However, the Servicesrealize that the danger of fratricide needs specialattention. The Army recently has begun to carryout separate after-action analysis of simulated“fratricides” at the National Training Centers.This includes a standardized format for an evalua-tion, called the Fratricide Incident Report, whichallows comparison of cases and analysis of mostcommon causes. Avoiding potential fratricidedangers are now an explicit part of battle plan-ning. The Army promulgates this new emphasisthrough the Training and Doctrine Command(TRADOC).

Realistic training can reduce casualties incombat—including fratricide-but the inherentdanger of training creates a connundrum. Onemilitary analyst has suggested that some of theproblems encountered in the Persian Gulf oc-


curred because the military did not train enoughand not realistically enough.19 Intense and realis-tic training results in better combat performance,no doubt, but it costs lives itself. For example,1991 was the Air Force’s safest year on record,with only ten noncombat deaths due to flyingaccidents, In one particularly unfortunate 24-hourperiod late in 1992, the Air Force lost four aircraftand 17 crew in three separate accidents.20 Train-ing for war will never be both perfectly effectiveand perfectly safe. Those that suggest morerealistic live-fire training must weigh the possiblereduction in losses in some future war against thehigher annual peacetime losses incurred from thetraining.

Simulators are an important component ofmodem training. These need to be improved tobetter represent real combat identification prob-lems. For example, most ground combat fratri-cide in the Persian Gulf War occurred at night,but tank training simulators do a much better jobof reproducing daytime than nighttime conditions—that is, the world seen though infrared imagingdevices.

On a large scale, simulation becomes combatmodeling and, until recently, almost all computer-ized combat models simply ignored the possibil-ity of fratricide. In fact, most computer modelsproceed as though both sides have perfect tacticalinformation and fratricide never occurs. Indeed,no provision is allowed in most computer pro-grams to investigate the effects or causes offratricide. This past neglect of the problem is nowbeing addressed.

Doctrine

Doctrine affects the causes and cures of fratri-

cide. Avoiding fratricide has never been the sole,or even primary, determinant in doctrinal devel-

opment and rightly so. As fratricide is becomingrelatively more important as a source of casual-ties, however, it is receiving increasing emphasis,especially within the Army. Yet fundamentaltenets are not being examined as closely as theymight. For example, the Army emphasizes a veryaggressive approach to ground combat, arguingthat speed and ‘‘momentum” are key to successand holding overall casualties low. Yet manyfratricide in the Persian Gulf occurred at night atranges where the enemy—because of inadequatesensors—was unable to shoot back. At least onecivilian analyst has suggested that-with theshooters in less immediate danger-slower, moredeliberate, attacks could have been carried out toreduce the risk of fratricide. The Army argues thatdoctrine and training are not something that onecan turn on and off like a switch. ‘‘Today, troops,we will use Doctrine B’ will not work. Perhapsfurther development of the “weapons tight/weapons free’ approach would suffice.

Clearly, the Services must coordinate theirapproaches to avoiding fratricide. The interactionof the ground forces with the air, is one clearexample. Close air support has never been the AirForce’s primary interest and identification ofground targets has suffered. This is the time, withnew interest in developing antifratricide technol-ogy, for careful coordination.

With the end of the Cold War, some militaryanalysts are discussing quite radical suggestionsfor the reassignment of roles and missions of theServices. Avoiding fratricide will not be theprimary determinant in these decisions, but itshould be important in several cases, for example,whether to assign close air support to the Army orwhether to assign ground-based air defense mis-siles to the Air Force.

19 1‘T~aini~~ ~eed~ to ~~ge, too, ~though o~ soldiers we of tie highest quality, they Stfl do not main redkticdly enOUgh fOI WaT. Thele

is too much emphasis on sa~ery. [emphasis added] As a result, units do not train for integrated combat in a live-fire environment, where artillery,armed helicopters, close air-support jets, armored vehicles and soldiers replicate the violent and confusing conditions found on the battlefield.Col. David H. Hackwortht “Lessons of a Lucky War,” Newsweek, vol. 117, No. 10, March 11, 1991, p. 49.

ZO Ayiafion Week and Space Technology, “USAF Loses Four Aircraft, Including B-lB, in 24 Hr., ” Dec. 7, 1992, pp. 24-25.

. . - .-

48

—— —

Who Goes There: Friend or Foe?

CONCLUSIONFratricide is becoming more visible and may

frequently be a relatively more serious source ofcasualties in modem U.S. combat because majormilitary operations are now sometimes possiblewith remarkably few losses, as the Persian GulfWar demonstrated. Cures for the fratricide prob-lem deserve serious, continuing attention, butfratricide is not a cause for panic and will not losethe next war.

Avoiding fratricide takes more than just identi-fying targets properly. And target identification is

a complex problem that will not be solved by a“black box.” Much of the information needed toavoid fratricide is exactly the information neededto be a coherent combat force.

Future wars will include joint operations amongeach of the Services and among the United Statesand its allies. Today, the military R&D commu-nity is pursuing several antifratricide develop-ments. Existing efforts to coordinate with sisterServices and allies should be vigorously main-tained.

AvoidingFratricide of

Air andSea Targets 4

As described in the previous chapter, reducing fratriciderequires more than improving identification. This chap-ter discusses ways to improve both tactical knowledgeand identification to avoid air fratricide. There are

several technical approaches to better identification; each ofthese is described briefly below with a discussion of itsadvantages and disadvantages. The chapter ends with a discus-sion of the interaction of military identification systems withcivilian air traffic control and a brief discussion of avoidingfratricide of ships.

BATTLE MANAGEMENT‘‘Battle management’ includes collecting information about

where combat resources are needed, setting priorities, andallocating resources to needs. The tactical knowledge or‘‘situational awareness’ provided by battle management is sointegral a part of air combat that its importance to avoidingfratricide is easily overlooked, Yet the foremost antifratricidemeasure is properly coordinating friendly forces.

Any efforts that improve coordination also improve combateffectiveness—and that typically is their primary justification—but these same efforts can help reduce fratricide. The Navy andthe Air Force discovered during joint operations in the PersianGulf War that air tasking orders (ATOs) were difficult totransmit between the Services’ strike planners. Air Force andNavy radios were not always compatible and the ATOs were sovoluminous that transmission was time , consuming. These andother uncovered communication problems are now beingcorrected. The resulting improvement in attack efficiency willbe obvious, but better communication and coordination alsowill make fratricide less likely.

~-—-- ------ ..-–.. ~

II

1 --1

49


At the tactical level, long-range surveillancecan sometimes track an enemy airplane from themoment of takeoff. If a fighter can be seen takingoff from an enemy airfield, few would argue withthe assumption that it is an enemy airplane. Thiscapability is partially in hand today with theAWACS. The limitations of the system are rangeand, more importantly, tracking. Once the enemyairplane gets close to a friendly airplane the radarmay no longer see them as separate targets or theradar may loose track of the airplane when it fliesbehind mountains. Then, when a distinct radarecho is again detected, the tracking radar cannottell whether the aircraft came from an enemyairfield.

AWACS can hand down information, butgreater benefits accrue with a two-way communi-cation between some central coordinating pointand forward shooters equipped with IFF capabili-ties. For example, to avoid fratricide, surface-to-air missiles are specifically allocated defensiveareas in which friendly aircraft are not to fly. Testsare currently underway to evaluate the feasibilityof transmitting target identification down toindividual missile batteries so that missiles andfighters can operate in the same area. But much ofthe information handed down from the centercould come from analysis of data collected by theindividual missile batteries in the field.

The ultimate goal for any identification andcommand system would be an array of individualshooters equipped with point-to-point identifica-tion of friend and foe (IFF) capabilities, collectinginformation and sharing it through some networkso identifications are based on a compositepicture built up from all available information.With the possible exception of the missile-fightercoordination, none of the communications im-provements currently proposed or underdevelop-ment are being justified solely-or even pri-marily—as antifratricide measures, but their con-tribution to avoiding fratricide could be substan-tial. If allies can tap into the information-sharingnetwork, they can still get the benefit of U.S. IFFinformation without acquiring the technology.

NONCOOPERATIVE IFFChapter 3 discussed how cooperative systems

really only identified friends; noncooperativetechniques are able to identify foes as well.Noncooperative identification of aircraft variesfrom the very simple-visual recognition—todetection, analysis, and classification based onextremely subtle differences among target air-craft.

The end of the Cold War will change theequations governing noncooperative IFF. In thosefuture Third World conflicts in which the UnitedStates has overwhelming air superiority, positiveidentification of enemies-and hence noncooper-ative IFF--will be important. After all, anyunidentified aircraft picked at random is likely tobe friendly under those conditions, so failure torespond to an IFF query will probably not bejustification to fire. At the same time, the techni-cal challenge now will be in many ways muchgreater than during the clearer confrontationbetween NATO and the Warsaw Pact forces,since both Western and Soviet equipment are nowwidely proliferated. Therefore, allies and enemiesmight very well be using the same equipment, asthey indeed did during the Persian Gulf War. TheServices agree that no single measurement will beadequate to identify enemies, rather that a com-posite picture formed from many sources ofinformation will be needed to be definitive. Someare discussed below.

1 Radio-Emission InterceptPerhaps the simplest noncooperative tech-

nique—short of visual identification-is passiveinterception of radio and radar transmissions.Each radio and radar system transmits at charac-teristic frequencies, with characteristic signalmodulation, and-at least for radars--character-istic pulse shapes and repetition rates. Someaircraft will transmit radio-frequency energy rou-tinely, while others will at least occasionallytransmit. It is also theoretically possible to induceenemy aircraft to transmit signals, perhaps by

— —— — . .

Chapter 4–Avoiding Fratricide of Air and Sea Targets 51

sending false communications requiring answersor by appearing to threaten the aircraft in a waythat forces the enemy pilot to turn on defensiveradars or otherwise communicate with his com-mand and control network.

9 RadarCareful analysis of radar returns reveals much

more about a target than just its bearing and range.Soon after the development of radar, operatorsnoticed that the propellers of aircraft modulate thefrequency of the radar return in a characteristicway. The modern equivalent is called jet enginemodulation (JEM). The air intakes of jet enginesreflect radar signals very efficiently. Some of theradar waves entering the inlets are reflected off ofthe rapidly rotating compressor or fan blades. Themotion of the blades causes a slight Doppler shiftin the frequency of the reflected waves. Thesesubtle frequency shifts are readily detectable bysophisticated radars and are, moreover, character-istic of particular jet engines.

The principal limitation of identification by jetengine modulation is clear: the technique identi-fies engines, not aircraft. There are a limitednumber of military jet engines available world-wide and very different airplanes can be poweredby the same type of engine, At the same time,individual aircraft in a particular fleet might havedifferent engines. For example, some U.S. F-16shave been fitted with the General Electric F-1 10engine while others have been fitted with modi-fied Pratt and Whitney F-100 engine originallyused in the F-15.

Jet engine modulation should at the very leastdistinguish fighter aircraft from transport aircraft,These two types of aircraft use very differenttypes of engines: fighter engines typically havelow bypass ratio engines and therefore havesmall, high-speed fans, while transports have highbypass ratio engines with much larger, slowerfans. This method will not, however, necessarily

distinguish military from civilian transports, Forexample, the commercial Boeing 757 and themilitary McDonnell Douglas C-17 both use thePratt and Whitney F-1 17 turbofan engine.1

In addition, the technique is highly dependenton a proper geometry between the radar and thetarget. This dependence can restrict the tech-nique’s application in a dynamic air engagement.

More detailed information about the airplanestructure itself will be available from high-resolution radars (HRR) under development.Radio or radar waves are just a form of electro-magnetic radiation, like light, and travel at thesame speed. Light travels about 300 meters in amicrosecond. A typical radar sends out pulses, orbursts of radio waves, that are on the order of amicrosecond in duration. This means that theradar pulses are many meters long. Resolvingfeatures much smaller than the radar pulse lengthis difficult; thus conventional radars are good atdetecting objects but not much use for providingdetails of objects as small as airplanes. Targetsappear just as blobs on the radar screen. If,however, a radar had a very compact pulse,perhaps the individual reflective surfaces of anaircraft could be resolved, which would allowidentification. Such HRRs are currently in re-search and development. See figure 4-1.

The challenges facing high-resolution radardevelopment are substantial. First, of course, isdesigning and building a radar that can emitpulses with duration of only several nanoseconds(billionths of a second). Proponents of high-resolution radar are confident that the technologyis available or can be developed. In addition,however, are the operational challenges. Forexample, each target will have a different echopattern depending on the perspective of theviewing radar. Side views will look nothing likehead-on views and data catalogs must be devel-oped of all potentially hostile aircraft seen fromall possible aspects.

1 Mark Larnbert, cd., Jane’s All the WorZd’s Aircraft (CoulsdoU UK: Jane’s Information Group, 1990), p. 748. The civilian designationfor the engine is “PW2040.’

.— - -- - - - - - - ..- .- . — . .


I

Figure 4-l—High Resolution Radar

Identification with High-Resolution Radar

A ->Transmitted radar pulse

<

Reflected radar pulse

n

AL

Transmitted radar pulze

<

Reflected radar pulse

\

Foe

Friend

~~Reflected pulse

\

m “IDENTIFICATION

I /

-=

/

Catalog of radar signatures

Office of Technology Assessment, 1993.

Moreover, different aircraft can look similar thinner profile and is therefore more difficult tofrom particular directions. For example, from a detect. From certain attitudes, if the range is nottrade journal review of the Soviet Su-27 fighter: known, the forward portion of the Su-27 also“From a head-on or trailing position, the Su-27 resembles that of the USAF/General Dynamicsresembles the Navy/Grumman F-14, but has a F-16 because of its prominent bubble canopy and

—.——


U.S.-built F-18 from the Australian Air Force fliesbelow a similar-looking Su-27 from the former SovietUnion. Superficial resemblances of weapons can makequick identification difficult and unreliable.

its forward aerodynamic strakes, which blend intothe wing leading edges.”2

The algorithms used to discriminate amongaircraft will not look at every detail but willextract and concentrate on certain defining char-acteristics. If an enemy knew which characteris-tics were used for discrimination, then it could tryto suppress or alter them. Thus, the algorithmswill need to be strictly secret and cannot be sharedwith all allies.

~ Surface-to-Air Missiles andNoncooperative IFF

Difficulties of working out aircraft identifica-tion has forced surface-to-air missiles (SAMs)and air interceptors into different zones of respon-sibility. Typically, SAMs defend strips of air-space from which all friendly aircraft are ex-cluded; thus, anything entering the zone could beconsidered hostile and attacked. These areas arecalled Missile Engagement Zones or MEZs.Undefended corridors through the strips allowfriendly aircraft to pass from one side of the stripto the other. Since enemy aircraft can trackfriendly aircraft and soon discover the locations

of the corridors, the corridors must be movedfrequently. Areas outside the MEZs are left to theinterceptors. SAMs would not routinely engageany aircraft within these Fighter EngagementZones, or FEZs.

The introduction of the Patriot missile changedthe utility of this allocation of responsibilitybetween interceptors and SAMs. The range of thePatriot is so great that, at least in the Europeantheater of operations, there would be little areathat was not accessible to both interceptors andPatriot. The Army, which operates the Patriotbatteries, could severely and artificially restrictthe engagement range of Patriot, but that obvi-ously eliminates much of its capability and thejustification for the cost of the system. To resolvethis conflict, the Army and the Air Force starteda program called Joint Air Defense Operations(JADO) to test a concept known as Joint Engage-ment Zones, or JEZs. The test of the feasibility of

An infantryman prepares to launch a hand-heldsurface-to-air missile against an attacking aircraft.The combined-arms battlefield is complex anddynamic, with many different types of weapons able toengage enemy targets, thereby greatly complicatingand broadening friend and foe identificationrequirements.

2 Donald E. Fink, “Sukhoi, Australian Pilots Fly in Joint Maneuvers, ’ Aviation Week and Space Technology, Mar. 5, 1990, pp. 64-67.

. . . . . . . . . . . . .


Box 4-A—History of Aircraft Identification

From the time aircraft were first used in combat in World War 1, fratricide has been a problem. Early fratricidesmotivated the application of national insignia on the wings and fuselages of aircraft and were at least part of thereason that some pilots-like the famed “Red Baron’ ’resorted to garish color schemes.

Between the World Wars, the military gave some slight attention to the problem of identifying friendly aircraftbeyond visual range, or in aloud or fog. As early as 1928, the British speculated on the possible use of sirens,whistles, or “singing” wires to create a signal that could be heard even if the aircraft could not be seen. Bombercommand also considered schemes to use special light signals to identify returning aircraft.1

The problem of fratricidal attacks on aircraft and surface ships took its modern and familiar form with thedevelopment of radar. Early radar inventors foresaw immediately that radar-a device that allowed detection ofaircraft and ships at long range, at night, and through clouds--created the difficult problem of identifying thedetected object?

Early radar developers’ first IFF attempts were to alter the radar returns of friendly craft in some characteristicway. Radar pulses are nothing more than radio waves that, when reflected from an object, create detectable echopulses. The apparent size, or radar “cross-section,” of an object determines the intensity of the radio wavesreflected back. The cross-section in turn, depends roughly on the physical size, shape, and orientation of the objectbut also on the electrical characteristics of the object. For example, a conducting rod equal in length to half theradio wavelength will resonate with the radio waves, which causes a particularly strong reflection. If the resonantreflection could be turned on and off, then the radar return would vary in a way that could be used for identificationpurposes.

. ,.~ .“*

:+ ““’’””’”w m~a as .c.~~*

og

-,-+s

c 1’m ,“n i’z

. .(

<a

,, ,.& w

< .

i

. . . . ,-- -u..- * . “ ! * - - . . ~ ~-,—..- &“”m&y* =:--~:.., ..” . *,y* - - - , =*-& * - - - :.-— . .

I tim.m-“-*-** u-a

$+

T e ‘+ - T::IF4

?~

..-

A German aircraft spotter’s field guide used in World War I. From the very first uses of aircraft incombat, identification has been a challenge.

In 1937, the British mounted a few test aircraft with an antenna wire running the length of the fuselage. Aswitch at the center of the antenna was turned on and off in a regular pattern by a cam. Toggling the switcheffectively changed the length of the antenna and hence the degree of resonance and the radar cross-section.Thus a radar operator would seethe “size’ ’of the target changing in a pattern known to be characteristic of friendlyaircraft. Tests on individual aircraft were very successful but the crude cams would not have worked for groups

1 ~ans. Swords, A T~~n/~/jjjstoryoft~e &@~~j~~s ofF/a&r, Tedrliml Report MEE1 (Dublin, Ireland:

Department of Microelectronics and Electrical Engineering, Trinity College; 1983), p. 99.2 [ndeed, early radar developers mined the term “[F~’ for “interrogation, Friend or Foe” or, as Some eW[Y

operators referred to it, Wziefriendorfoe?” See Robert Morris Page, 7?)e Odginofl?acfar(Garden City, NY: AnchorBooks, 1962), p. 166.

— — —

Chapter 4–Avoiding Fratricide of Air and Sea Targets I 55

of aircraft: without perfect synchronization, one airplane’s switch might open just as another’s might close creatingan indecipherable jumble and the system was never adopted.

In 1939, the U.S. Navy mounted atop a destroyer a set of half-wavelength rods on a pole. A motor rotatedthe pole and the rods along with it. The rotation changed the orientation of the rods, hence their degree ofresonance with a distant radar and thus the strength of the radar echoes. The radar echo from the destroyeroscillated in an obvious way that identified it as a friend. This technique, while simple, had the same limitationsas the aircraft system and because of its simplicity was easy for an enemy to copy.

The limitations of passive cooperative techniques led radar researches to active radar reply devices, nowcalled “transponders.” The first transponders operated at the radar’s frequency; whenever the transponderdetected a radar pulse it would transmit its own pulse at the same frequency. The radar would detect this pulse,interpret it as a powerful radar return, and the target would show up brightly on the radar screen.

The first transponders were the Mark I and Mark II developed in Britain and similar devices developed aroundthe same time by the U.S. Naval Research Laboratory (NRL). These devices scanned all radar frequencies in useby friendly forces and retransmitted a pulse at the appropriate frequency whenever a radar was detected. In theearly days of radar, this technique was possible because only two or three radars frequencies were common, butas more radar frequencies became available, this approach became untenable simply because the transpondercould not handle the range of frequencies.

By 1941, the proliferation of available radar frequencies required that IFF devices go to a single frequency,independent of the radar’s frequency. Thus, the radar could operate on whatever frequency was most appropriateand an additional signal, part of the so-called “secondary” radar, would query the target’s identity. The Mark Illwas the first such device, sending and receiving signals in the 157-187 MHz (that is, megahertz or millions of cyclesper second) frequency band.3 The Mark Ill became the standard IFF device used by the American, British, andCanadian air forces during World War Il.

The Mark IV, developed at the U.S. Naval Research Laboratory (NRL), was the first IFF system to usedifferent frequencies for the query and the response--470 MHz and 493.5 MHz-but it never came intowidespread use.4 In 1942, the NRL began development of the Mark V, also called the UNB or “United NationsBeacon,” which was to operate near 1 GHz (that is, gigahertz, or billion cycles per second). This program was notcompleted until after the war but is important because t he f requencies used-1 .03 GHz for queries and 1.09 GHzfor replies-are still used today on both civilian and military transponders.

The next set of refinements appeared in the Mark “X,” which had a dozen query and response channelsavailable. 5 Mark X originally allowed aircraft to identify themselves as friendly but did not allow different responsesfrom different friendly aircraft. A capability, known as SIF,6 allowed different responses from different transponders.This capability, plus an encrypted query and response mode added to the Mark X became the current Mark XII.(The Mark XII used for civilian purposes without the encryption capability is still frequently referred to, especiallyin Europe, as the Mark X-SIF.)

The Mark XII is today used by U.S. aircraft and ships but is not widely used among U.S. allies.

3 The u-s. equivalent of the Mark Ill was ~lled the SCR 595. See Swords, op. cit., footnote 1, p, 102.

Transponders operating on this principle, now in near universal use on aircraft for civilian air traffic control as wellas military IFF, are usually called ‘(semndary surveillance radars.” See Michael C. Stevens, SeconclWy Surwei//anceRadar (Norwood, MA: Artech House, 1988), p. 7.

4 Much of this and subsequent wartime development of IFF devices was a joint effort of U.S. and British

scientists, called the Combined Research Group and headquartered at the Naval Research Laboratory.5 The 11)(11 was a pla~+older until a decision could ~ made about whether a new Mark number WaS justif ied

but it became, perhaps inevitably, the Roman numeral 10. Thus, there are no Marks Vl, Vll, VIII, or IX in thechronology.

6 U.S. sourmsstate fhaffhe acronym stands for “Selective Identification Feature, ” but some British sour~sstate that the “F’ stands for “Facility.”


this approach is called JADO/JEZ. Tests are nowbeing carried out at or planned for Nellis andEglin Air Force bases. Since the establishment ofthe test program, the Navy has become involved,to enable it to test operationally some noncooper-ative surface-to-air identification systems, suchas the Shipboard Advanced Radar Target Identifi-cation System, or SARTIS. In February 1994,coordination of Army ground-based SAMs, AirForce interceptors, and Navy off-shore SAMswill be tested along the coast of Florida at EglinAir Force Base.3

Patriot missiles belong to the Army but duringjoint operations their rules of engagement aretypically set by the Air Force. The Air Forcewants positive hostile identification before allow-ing the missiles to fire. Lacking any intrinsiccapability to do so, Patriot batteries must dependon higher echelons to provide the information.During large-scale battles involving many poten-tial targets, the transfer of information from highechelons down to individual fire control centerscan saturate current data networks, causing delaysin identification data transfer.

Any development of noncooperative IFF mustconsider SAMs from the beginning if they are tocontribute to their full potential. For example,Patriot missiles, while gaining fame for theirinterception of Scud missiles in the Persian GulfWar, were on permanent weapons-hold statusagainst aircraft targets. There were, therefore, noground-to-air fratricides despite numerous viola-tions of Army air defense areas,4 but also, ofcourse, Patriot played essentially no role in airdefense operations.5

COOPERATIVE QUESTION-AND-ANSWERIFF SYSTEMS

Cooperative question-and-answer systems havebeen central to combat aircraft identification

since World War II; indeed, some discussions of“IFF” really only treat this one approach. Adescription of the Mark XII and now-canceledMark XV provides specific information on thesetwo systems but also offer a convenient frame-work for presenting general design considerationsfor all question and answer IFF systems.

Although the early “Mark” numbers referredto specific types of hardware, the Marks XII andXV—and presumably future models—really referto protocols. That is, “Mark XII” refers to anagreed format, or protocol, for sending andreceiving information. This includes the relevantfrequencies, the length of the radio pulses, thetiming between them, the meaning of differentpulses, and so forth. The Mark number does notspecify the hardware configuration required, andMark XII equipment has gone from vacuum tubesto transistors while using the same protocols. Ofcourse, some hardware must embody the proto-cols, but any one of several interrogators andtransponders can handle these formats and are,therefore, ‘‘Mark XII” devices. For example, theUPX-23 and UPX-27 refer to two specific ship-board Mark XII interrogators while the APX-72is an example of a Mark XII airborne transponder.Thus, the following discussion can combinespecific questions related to protocols with gen-eral descriptions of hardware.

The current Mark XII sends out a query in the‘‘L” radar band, at a frequency of 1.03 GHz. Thequery is a pair of radio pulses. The time betweenthe two pulses can be varied and the transponderwill interpret the query differently depending onthe separation time between the pulses. Theimmediate predecessor of the Mark XII, the MarkX, used three different pulse separations, eachreferred to as a ‘‘mode. ’ A pulse separation of 3microseconds is “Mode l,” 5 microseconds is

q Capt. Rocco Erse~ N6X1, brief~ titled: “Introduction: JADO/JEZ, ” (undated).4 Briefing entitled, ‘‘Combat Identification: TRADOC Briefing,” Lt. Colonel Mike Bro~ U.S. Army Training and Doctrine Command

(March 12, 1992).

5 U.S. Congress, General Accounting Office, Aircrajl Identification System(U), GAO/C-NSIAD-92-13, p. 13.

Chapter 4–Avoiding Fratricide of Air and Sea Targets! 57

“Mode 2,” and 8 microseconds is “Mode 3.’These modes are still in use today.

The reply signal from the Mark X contained atleast a pair of 1.09 GHz “framing" pulses 20.3microseconds apart. These pulses indicate whenthe reply message starts and stops. Between theframing pulses of the response from the originalMark X lay six time slots 2.9 microseconds wide,each of which may or may not contain a radiopulse.6 A pulse in a particular time slot representsa “1“ and lack of a pulse represents a “O,’ thusallowing transmission of binary numerical data.The improvements implemented for the SelectiveIdentification Feature, or SIF (see box 4-A), andMark XII included an increase to twelve slotsbetween the framing pulses to allow for 4,096possible replies. With the available number ofpossible replies, airborne transponders can give adistinct reply that identifies not just whether theaircraft is a friend but which aircraft it is-exactlyas in civilian air traffic control today.

A simple transponder can be exploited easilyby the enemy. If the enemy could get thetransponder to respond to a query, friendlyaircraft would reveal their positions and identi-ties. To avoid this weakness, a program todevelop an encrypted query mode was started in1954. Mark XII IFF devices have this encryptedquestion-and-answer mode, called Mode 4. TheMode 4 query starts with four time synchroniza-tion pulses followed by up to 32 pulses thatcontain encrypted information telling the receiv-ing transponder that the query is a valid, friendlyquery. Invalid queries are simply ignored by thetransponder. The response to a Mode 4 query is astring of three pulses. The reply can start after anyof 16 possible time delays; thus by changing the

delay the reply can convey limited information.See box 4-B for the specifications of the interro-gations and replies.

The IFF interrogators and transponders cannotwork by themselves. If the transponder sent out aquery in all directions and got a response back, theIFF interrogator would only know that there issomewhere out there at least one friendly aircraft.To query particular aircraft, interrogator antennasare mounted along with conventional radar anten-nas to point in a particular direction. When atarget is detected by the radar, the IFF interrogatorcan send out a query in the same direction as theradar beam asking the aircraft to identify itself.The information from the reply can then bedisplayed on the radar screen directly.

The transponder can be simpler than theinterrogator. When it receives a query, the answerneed not be directed to the questioning radar;instead a simple omni-directional antenna isadequate and the interrogating radar can deter-mine which aircraft is responding by correlatingthe response with the radar beam’s direction.

Because of the difference in complexity, inter-rogators are generally more expensive than trans-ponders. This affects the extent of their use; forexample, some aircraft are fitted only withtransponders. Ships and control aircraft like theE-2 are routinely fitted with the more expensiveinterrogators, but many interceptors are not.

The different modes can be used for military airtraffic control. For example, an aircraft carrier’sradar will show all the aircraft in its vicinity but thismay be over a hundred aircraft in some crowded aircorridors, such as over the Mediterranean.7 Theradar operator can use IFF modes to highlightfriendly military aircraft.

b Oral history at the Navat Research Laboratory holds that the time slots were intended to be an even 3.0 microseconds wide but the fustdelivexy of delay lines for the prototype proved faulty, testing at only 2.9 microseconds. Rather than wait for new delay 1ines, researchersproceeded with what they had available. Thus, it came to pass that framing pulses of 20.3 (that is, 7x2.9) microseconds will be in use well intothe twenty-first century. This tale illustrates how protocol standards should be developed carefully because they tend to stay with us for a longtime.

7 Briefing entitled ‘‘Automatic Identification (AutoID), ” Applied Physics Laboratory, The Johns Hopkins University, (undated).

. . . - .- - . -- - . - - - . . . . - - - . . - . - -- — -_ — —- -—_ -_


Box 4-B-Formats and Protocols for the Mark XII

Mark XII interrogations and replies are pulses of radio waves at 1.03 and 1.09 GHz. Information is conveyedby changes in the number and timing of the pulses. The interrogation for military Modes 1,2, and 3 and civilianModes A, B, C, and D are a pair of pulses separated by a time delay, the length of which is specified by the particularmode. This format is illustrated in figure A.

Mark XII interrogation format

J L_

F 7 Suppressor-~ ~ pu lse

IL

A< 0.8 p sec

< Delay

Mode 3/A reply format

>< 0.45 p sec

ti 1< 20.3 p sec

Mode 3/A reply pulse 7654

F1 ~ Al A2 C4 A4 B2 B4 D4 F2

JLJ-JL.-LIL---L-l ULll -

Mode 4 formats

Query Reply

JLLILTime synchronization

pulse

-...-UJLL JLILUp to 32 data

pulses

SOURCE: NATO STANAG 4193 (Part 1) (Edition 2), 1990.

———— —


Box 4-B–Formats and Protocols for the Mark XII

No antenna is perfect, sending signals Table A—IFF Modesonly in the direction wanted and no other. —

Military civilEnergy leakage through “side lobes” could Delay (p see) mode mode Use

cause transponders to f ire even when t he radar 3 1 Military Function IDis not pointed directly at them. To avoid this 5 2 Military Function IDproblem, some interrogators send out a sup- 8 3 = A Aircraft identification

pressor pulse in every direction except the 17 B Not used internationally

direction the radar is pointed. A transponder 21 c Altitude25 D Not used internationally

can compare the size of the two signals and ifthe suppressor pulse is larger than the framing SOURCE: NATO

pulses, the query is ignored.Each interrogation mode has a different time separating the pulses, except that military Mode 3 is equivalent

to civilian Mode A. The various Modes are shown in table 4-1. The pulse separation in Mode 1 is so short that notail interrogators and transponders can handle the insertion of a suppressor signal in Mode 1.

The reply format consists of a pair of framing pulses 20.3 microseconds apart with up to 12 signal pulsesbetween them, although not all modes use all the availabie signal pulses for information. The format is shown infigure B. Numerical values are transmitted in the replies in the form of four-digit “octal” or base eight numbers ofthe form ABCD. Each of the these digits is the sum of three pulse values. in the figure, for example, three of thepulses are Iabeled Al, A2, and A4. The first digit in the four-digit number is A which equals Al +A2+A4 where Alhas a value of one, A2 a value of two, and A4 a value of four if a pulse is present in the appropriate time slot, andzero otherwise. Thus the decimal number 4,012, which is 7,654 in octal notation, would be represented byA=7=A4+A2+A1, B=6=B4+B2, C=5=C4+C1, and D=4=D4. The resulting pulse pattern is shown in figure C.

Mode 4 pulses, the encrypted mode, have a different format. The interrogation pulse starts with four timesynchronization pulses. These are followed by up to 32 data pulses. The arrangement of these pulses validates

that the query is indeed from a friendly interrogator and transponders should send a reply. The reply is a set ofthree pulse delayed by various amounts. These formats are shown in figure D.

1 Passive Cooperative IFF Measures appropriate noncooperative identification tech-

Cooperation by friendly targets could enhancethe discriminating ability of passive friendlyobservers using the noncooperative IFF tech-niques described above. Few of these ideas arebeyond the conceptual stage. They include addingradar highlights that would stand out on high-resolution radar, inducing vibrations that wouldshow upon Doppler radars,8 and even modulatingor doping exhausts to make them stand out toinfrared sensors. Application of any of thesetechniques will require first the deployment of the

nique.

1 Limitations of Mark X11, Improvements inMark XV, and Next Generation IFF

This section briefly discusses some of thecurrent and potential short-comings of the currentMark XII IFF system. These problems were to becorrected with the development of the Mark XV.The Mark XV development was canceled inDecember 1990 and the still-to-be-defined suc-cessor question-and-answer system is now re-ferred to as the Next Generation IFF, or NGIFF.

g Briefing entitled, ‘ ‘Achieving Covert Communications and Ground-Combat Identification Using Modulated Scatterers, ’ E.K. Miller andD.M. Mctzger, Mechanical and Electronic Engineering, Lns Alamos National Laboratory (Mar. 11, 1992).

I

II

. -. . -. . - -— - — - - - - - -- - -- - - - — - - - - - - - - - — - -


A typical Mark XII airborne transponder. IFF formatsand protocols can be handled by any of several typesof interrogators and transponders.

As discussed in chapter 3, the three ways inwhich an enemy can defeat the purpose of an IFFsystem are exploitation, spoofing, and denial. Anenemy exploits an IFF system by getting informa-tion from it. For example, if an enemy couldrecord queries from a Mark XII interrogator andthen rebroadcast them, then he could trigger theMark XII transponders and have friendly aircraftidentify themselves and reveal their positions.Even if recording valid queries were impossible,an enemy could guess at queries, hoping to hitupon a valid combination. With thousands ofpossible queries this may seem daunting, but infact modem electronic devices should allowtransmission of scores of guessed queries persecond.

The Mark XV would have reduced or elimi-nated the possibility of enemy exploitation bychanging the query codes rapidly. Valid Mark XIIquery codes are changed regularly but the processis cumbersome, involving the use of printed keysand the mechanical insertion of key values intothe transponder. In the late 1960s, the Navy flighttested a system called TACIT (Time Authenti-cated Cryptographic Interrogator/Transponder),

which allowed the rapid, automatic changing ofquery codes. Indeed, codes could be changed sorapidly that an enemy could not record andretransmit a code before it became obsolete

The Mark XV was to have a capability likeTACIT. This means that all transponders, at leastwithin a theater of operation and ideally world-wide, would switch codes in an agreed pattern,say, every second on the second. Clearly therequirements for time synchronization acrossliterally thousands of platforms around the worldis a major technical challenge. But modemdevelopments in electronics makes the task possi-ble if not easy; modem electronic clocks haveaccuracies of a fraction of a second per year. Theencryption computer contained with each interro-gator would be supplied with a “seed” key thatwould allow generation of a query code, thenalgorithms embedded in the computer wouldgenerate the next code from that seed and so onuntil a new seed were supplied.

An enemy’s ability to appear friendly to an IFFsystem is called spoofing. Just as an enemy couldtry to exploit the Mark XII by recording orguessing at proper queries, an enemy could alsotry to spoof the Mark XII either by recording orguessing proper replies. The Mark XV wouldhave incorporated several improvements to in-hibit spoofing. These included a greater possiblenumber of responses creating a greater barrier tostraightforward guessing. The Mark XV wouldhave used “spread spectrum” pulses, that is, asharp pulse converted by the transmitter electron-ics into a broad frequency pulse that has lowerpeak energy at any given frequency, makinginterception more difficult. The receiver hassimilar electronics that operate in reverse, com-pressing the spread pulse into a sharp informationpulse. A receiver that knows the proper spreadingand compressing function gets substantial re-ceiver gain, but an enemy that does not know thedetails of the function will have a difficult timeresolving the signal from background noise.9

v Don J. Torrieri, Principles of Secure Communication Systems, Second Edition (Bestow MA: Artech House, 1992), pp. 95-99.

Chapter 4-Avoiding Fratricide of Air and Sea Targets I 61

Also, the strength of the reply signal would havebeen adjusted for the distance of the interrogator,with the reply no stronger than it needed to be thusmaking more difficult the interception and re-transmission of a valid reply,

An enemy might be able to deny the use of anIFF system, For example, jamming of the radiosignals is one straightforward approach. Almostany radio can be j ammed if an enemy is willing toinvest adequate resources and can get jammers inthe right places. Military radio and radar systemsare designed to make j amming more difficult, but

Jamming can never be made impossible and theamount of effort that is appropriate to invest ine l e c t r o n i c antijamming capability depends sensi-tively on 1) the presumed combat environment, 2)a judgment about the value of working in a

Jamming environment, and 3) a comparison withother approaches to solving the jamming problem(e.g., blowing up the enemy jammers).

Military users of Mark XII consider it easily

Jammed. Improvements contemplated for theMark XV included significant antijamming capa-bility. For example, the query and response pulseswould be much longer than those of the Mark XII.Longer pulses mean the potential for greater totalenergy in the pulse and a longer signal integrationtime, which makes j amming more difficult, Eachbit of information in a pulse would be spread overthe whole pulse length, which makes reading thepulses in the presence of j amming more reliable.The spread spectrum electronics that contributeso much to communications security also makejamming more difficult. Finally, the structure ofthe pulses would allow for detection of transmis-sion errors so that if jamming occurred, theoperators would know that information receivedwas faulty and should be rebroadcast.

Jamming is one potential threat, but an indirectform of denial is perhaps the most important: ifthe operators do not have confidence in thesystem, they turn it off. Several experienced pilots

reported to OTA that during the Vietnam War,they turned off their Mark XII IFF systems assoon as they entered enemy air space. Other saythat more recently, on missions flown nearpotentially hostile forces around the Mediterra-nean Sea, Mark XII were turned off. The reasonsare the same in both cases: a fear that enemy-oreven friendly--queries would trigger the trans-ponders and thus reveal the aircraft’s presence.Against the more sophisticated Soviet threat,pilots expected to turn off IFF transponders longbefore they crossed enemy lines. If the operatorbelieves that an IFF system increases his danger,then it will not be used to its full potential, if it isused at all. Thus, the ON/OFF switch becomesanother means of denial. Mark XV improvementswould have mitigated the problem of inadvertentfriendly triggering of the transponder but this isclearly more than just a technology problem anda solution requires careful coordination amongthe eventual users.

Despite its promised improvements, Mark XVdevelopment was canceled in 1990, both becauseof increasing technical complexity and the grow-ing estimated cost of deploying the system. Over40,000 Mark XII sets have been produced10 andapproximately 25,000 are still in use. With a hugenumber of platforms needing IFF devices—and aneed for at least 17,000 Mark XVs was forecast—acquisition costs multiply rapidly. Indeed, thecost of the Mark XV would have precludedoutfitting every vehicle with a device, whichraises the distressing question of the value of anIFF system that is so expensive that it is notcarried by all platforms.

The cancellation of the Mark XV programleaves the future of cooperative aircraft IFFuncertain. The original motivation for the pro-gram, obsolescence of the Mark XII, still stands,so some replacement capability is probably needed.Most studies by the military Services assume that

1° Navy Briefing, “Defense Acquisition Board Milestone ‘O’ Review: Cooperative Friendly Aircraft Identification Presentation for C31Systems Committee,” July 22, 1992.


a cooperative IFF system will provide some partof that needed capability.

Perhaps the most obvious Mark XII replace-ment is a system with performance that is acompromise between that of the Mark XII and ofthe Mark XV. By backing off on the Mark XVperformance goals, costs could be reduced for anyr e p l a c e m e n t . Antijamming capability was one ofthe major determinants of costs. Justifying anyp a r t i c u l a r antijamming requirement is difficultfor any piece of electronic equipment. No one canever be entirely sure what resources a futureenemy might devote to j amming, how its jammersmight operate, be deployed, and so on. Thus,making very conservative assumptions is enticingbut has substantial cost consequences for elec-tronic systems. Options for future IFF systemsrange from using the current Mark XII protocolsand radio wave forms-that is, making no an-tij amming improvement to the Mark XII-tousing the spread spectrum waveforms proposedfor the Mark XV. In between are compromisesolutions that include using a modified andreduced spread spectrum waveform or usingdirectional antennas to reduce jamming interfer-ence. The Germans have made proposals formoving to a different frequency band altogether,in the S-band, which covers the range from 2-4GHz. One advantage of S-band is that theairwaves are less crowded at those frequencies.As a general rule, however, the higher thefrequency of radio transmission, the shorter therange. Thus, an S-band system has shorter rangethan an L-band system. The shorter range ofS-band actually has one advantage: it is harder tojam with a few distant but powerful jammers.Short range is less of a handicap in the crowdedEuropean theater in which the Germans operatebut it makes S-band unattractive for the world-wide operations needed by the United States.

Maintaining communications security was asecondary contributor to the cost of the Mark XV.

The Mark XV was to have anew cryptocomputer,the KI-15, that would have provided very goodsecurity through electronic handling of cipherkeys and rapid, automatic changing of keys thatwould have saved operational costs. While thecryptographic security component of the MarkXII is quite old, the whole system need not bejunked. For example, according to the NationalSecurity Agency, new cryptocomputers could beadded to the current Mark XIIs or to similarfollow-on models. These would provide bettersecurity without the full cost of the Mark XV’scapability.

The important point is that the Mark XVimprovements were not all-or-nothing; compro-mises in performance are possible. DoD couldbuild a system better than today’s Mark XII butcheaper than the Mark XV. The performancegoals for the Mark XV were established duringthe Cold War. Now that war with Russia seems farless likely and the more likely opponents are farless challenging, backing off on the performancerequirements of the Mark XV should be givenserious consideration. The exact emphasis is, ofcourse, subject to judgment, but technical expertsinterviewed by OTA felt that the end of the Sovietthreat has probably reduced the need for antijam-ming improvements more than the need forcommunication security improvements.

COORDINATION WITH CIVILIAN AIRTRAFFIC CONTROL

B Civilian Airborne TranspondersA mid-air collision on December 16, 1960

between a United DC-8 approaching Idlewild(now Kennedy) airport and a TWA Super Con-stellation approaching LaGuardia killed the 128aircraft occupants and eight others on the groundand highlighted the limitations of the air trafficcontrol system even when aircraft were underpositive control.11

11 Ric~d J. Ken~ safe, separated, and Soaring: A History of Civil Aviation Policy, 1961-1972 (Washington DC: U.S. Department ofTransportation, Federal Aviation Administration; 1980), pp. 6-7.


The New York City mid-air collision acceler-ated adoption of several technical improvementsfor civil air traffic control, for example, secondaryradar transponders on civilian aircraft. By 1953,Mark X frequencies and protocols had beenreleased for international civilian use.12 Thecivilian airliners were given use of Mode 3, whichin civilian use is called Mode A. Civilian airlinersuse the 4,096 possible combinations of replysignal to identify the aircrafts’ flight numbers,with some numbers reserved for special mes-sages; for example, any aircraft that responds toa Mode A query with a “flight number’ of octalcode 7700 is saying that it has an emergency.

Additional civilian modes called, not surpris-ingly, Modes B, C, and D have been addedsubsequently, but only Mode C, with a 21-microsecond delay between query pulses, iscommonly used. Ground-based radars can deter-mine ground coordinates directly but are not ableto determine aircraft altitude. Therefore, mostaircraft now have transponders connected to theaircraft altimeter allowing the transponder toreport automatically the aircraft’s altitude inresponse to a Mode C query, The use by civiliansof military IFF frequencies and protocols hasimportant implications anywhere the two types ofaircraft expect to share the same airspace.

Mode S

Secondary surveillance radar has become souseful and important to civilian air traffic controlthat it has become a victim of its own success. Insome particularly heavy air corridors-for exam-ple, Europe and the east and west coasts of theUnited States—aircraft can receive identificationrequests from dozens of radars. This may includeair corridor radars and overlapping radar coveragefrom closely spaced airports. Moreover, not allnations’ air traffic control procedures are identi-

cal or even particularly well-coordinated, soneighboring countries may simultaneously querythe same aircraft for information important totheir air traffic control.13

The multitude of interrogators can result inhundreds of queries each second. This volume cansaturate the airborne transponders, making themunavailable when essential radar queries comethrough. Moreover, all of the aircraft are sendingout identification calls more or less continuously;interrogators are receiving the replies from prop-erly designated aircraft and a multitude of strayreplies that are picked up as interference.

The ongoing, but gradual, adoption of a newmode-Mode S, where ‘S’ stands for ‘select"—is the internationally agreed technical solution tothese problems. In Mode S, interrogating radarswill not routinely send out a general ‘ ‘who’s outthere” query with every sweep. Mode S requiresthat each aircraft have a unique, permanentidentifying number. A Mode S interrogator willoccasionally send out an ‘‘all-call’ message toestablish the identities of all of the aircraft withinits coverage. The radar and interrogator can thensend out specific messages to specific aircraft,requesting altitude, for example. As an airplaneleaves the coverage area of one radar, theinformation about that specific airplane can besent by land lines to the next radar, which will beexpecting it.14

Proponents of the Mode S system claim that itwill reduce transponder information loads by atleast two-thirds. It will also allow denser airtraffic near airports. One current constraint ontraffic density is the resolution of air traffic radars,not just safe separation distances. Two closelyspaced aircraft will respond to the same Mode Aor Mode C query and their answers will overlapwhen received by the interrogating radar, which

12 Mi~lael c, Stevem, Se~o&~ Sumeillance R~ar @oWood, MA: fi~h House, 1988), p. 9.

13 Lief IClette, “Europe’s Crowded Skies: Managing Civil-Military Airspace, ” International Defense Review, July 1992, pp. 659-662.

14 Radio ~~~~~ Cotission for Aeromutics, “ Minimum Operational Performance Standards for Air Trafilc Control Radar Beacon

System/Mode Select (ATCRBS/Mode S) Airborne Equipment, ” Document Number RTCA/DO-181 (March 1983).

— . . . . . . - -- - - - — . . . . . . . . . -. .- - . . . - — . . .- - .


I

will not be able to make sense of either of them.15

With Mode S, a radar could interrogate oneairplane and then, a millisecond later, send aseparate interrogation to the second airplane. Thistime lag is so short that it will not affect air safetybut will allow electronic separation of the replies.Thus, while Mode S is not needed by the U.S.military, it is part of a much-needed world-widecivilian standardization and modernization effort,of particular importance in Europe.

Mode S will allow messages longer than thecurrent Mode A and C 12-bit messages. Thegreater number of message bits plus the uniqueidentification numbers for aircraft allow limitedtwo-way digital and text data transfer betweenaircraft and the ground that would be impossiblewith today’s free-for-all interrogation system. Forexample, just as Mode C now queries altitude,Mode S could, in principle, query aircraft speed,bearing, rate of decent, destination, even fuelload. These data would be used by air trafficcontrollers to better manage aircraft in densetraffic.

Information can also be sent up to the aircraft.For example, changes of air traffic radio fre-quency could be sent automatically when aircraftmove from one control zone to another. Weatheradvisories or changes in flight plan instructioncould be sent up, then written or otherwise storedso they would be available to the pilot at a time ofher choosing.

The basic Mode S is called “Level l;” “Level2’ includes the ability to receive short commandsof 56 bits; “Level 3“ includes the ability toreceive longer command sets of 112 bits andtransmit short ones; “Level 4“ includes theability to receive and transmit long commands.Since the system is not yet operational, all of the

ultimate uses for the data transmission channelsof Mode S are not clear but enough is known nowto see that they will be useful to civilian pilots, butless so to military pilots.

Mode S will be adopted in phases in the UnitedStates. Large commercial passenger carriers musthave Mode S transponders operating by the end of1993. Commuter carriers and general aviationwill acquire Mode S capability more gradually.Perhaps within 10 years, all aircraft, even generalaviation, will need Mode S if they are to operatein dense air-traffic areas of the United States.

The military must operate in a world where theoverwhelming majority of airspace is undercivilian control (at least in peace time). Thismeans that the military must adopt at least someof the Mode S capability into any future IFFsystem. The Services have two major problemswith Mode S, neither of which is insurmountablebut each of which must be handled. First is cost.

The cost to the military of Mode S will be morethan just the cost of the equipment. Another blackbox can always get squeezed into a transport, butevery cubic inch on a fighter plane is alreadyoccupied. That means redesigning existing equip-ment, placing antennas, dealing with waste heatfrom more electronics, and so on. Mode S was tobe an integral part of the Mark XV. With itscancellation, DoD must find some other way tooutfit its aircraft with Mode S or come to someaccommodation with civilian air traffic control.16

The benefits of Mode S are greatest in thedensest air traffic. Service resistance to the cost ofMode S is easy to understand because densetraffic areas are just those in which militaryaircraft are least likely to fly, at least in the UnitedStates. But since commercial aircraft can hardlyavoid these areas, they will need Mode S. Then,

15 “f’hi~ is due t. he ffite speed of tie radio waves sent out by he &ansponders. Since radio @avels about S(XI meters per microsecond, a

20-microsecond reply is 6,000 meters long and any two aircraft whose distances fkom the interrogator differ by less than that will send messagesthat will overlap, or be ‘garbled,’ at the receiver. Note that the aimraft need not be dangerously close; since air tratllc control radars normallydo not discriminate aircraft altitude, two airplanes with substantial vertical separation can still have very close ground tracks.

lb ~~er fmm Barry Mbefl Harris, Deputy AWstrator, Federal Aviation Admum“stratiorL U.S. Department of Transportationi to RichardG. Howe, Acting ChairmmL Policy Board on Federal Aviation, Office of the Secretary of Defense, Apr. 20, 1990.


when Mode S is widespread, it doubtless will beused-even if not essential-in the less heavilytraveled airspace where military aircraft do fly.

The FAA does not foresee Mode S messagecapability as a requirement for operating ingeneral U.S. airspace, but requirements will beimposed on aircraft that wish to operate near theNation’s busiest airports. In principle, militaryflights could simply avoid those areas. Currentlythe DoD foresees that fighters eventually will beoutfitted with Mode S/Level 2 and cargo trans-ports will be outfitted with Mode S/Level 4.17

The second problem is more subtle: the infor-mation that Mode S might provide to potentialenemies. This is not a wartime problem—pilotswould just turn civilian modes off in war theatersand the military would take over air trafficcontrol. It is, rather, a problem of long-termpeacetime intelligence information loss. CurrentMark X-SIF identifies an airplane by its flightnumber for that day. Mode S will identify eachairplane uniquely by its tail number. The Serviceswill not tolerate this since it would allow apotential enemy to build up over time a valuabledatabase. For example, long-term compilation ofaircraft tracks might reveal how often particularaircraft shuttle between deployment areas anddepot maintenance sites. One solution is to allotto the Services a block of numbers that they couldmix around at random. Civilian and foreign airtraffic controllers would then know that theaircraft is a U.S. military aircraft but not whichone,

The military may object to even this muchinformation being available, but presumably theUnited States is also interested in clearly identify-ing civilian airliners as such. Any system thatloudly proclaims all civilian aircraft inevitablyidentifies military aircraft as well, at the very leastby default, since any airplane not proclaimingloudly will be assumed military. Thus, this

weakness may bean inevitable price that has to bepaid for the protection of civil aircraft fromaccidental attack.

fl Other Areas Requiring Civil-MilitaryCoordination

Today the military and the FAA jointly operateair traffic control (ATC) systems covering theUnited States and the air approaches to it, with themilitary providing about a fifth of the ATC assets.The Nation is in the process of converting to aunified system, called the ARSA-4, to be operatedsolely by the FAA. The unified system should bemore capable and cheaper. The Air Force willreceive data from the FAA radars and interroga-tors, which will be used for the identification ofaircraft approaching the United States .18 Some ofthe current Air Force air traffic control computerscannot keep up with the high traffic densities inthe Nation’s busiest corridors, but sections ofradar coverage can be systematically blocked outfrom the Air Force data link to allow the Air Forcesystem to concentrate on only those sectors thatare important to it. In the future, all Service ATCequipment and computers will be comparablewith FAA equipment.

The FAA plans also for the gradual adoption ofan automatic system to help pilots avoid mid-aircollisions. Currently, aircraft pilots have visualinformation, on-board radars, and secondary in-formation relayed up from ground radars. Groundradars can interrogate the transponders carried onaircraft but currently the aircraft cannot interro-gate each other’s transponders. The Traffic Alertand Collision Avoidance System (TCAS) willallow aircraft to use Mode-S transponders to getinformation from other aircraft in the area whichwill allow on-board computers to calculate andrecommend collision avoidance maneuvers. TCASwill be in place for airlines by the end of 1993.TCAS may eventually place some requirements

17 Frti COISO~ OSD liaison to FAA, persOnal brief.

1S Tom McNiff, *‘Air Force, FAA Working on New Radar System, ’ JournaZ of Commerce (Sept. 21, 1992), p. B3, (and personal briefingsfrom Richard Lay, FAA in-route radar manager)


on military aircraft even if none are in place now.The first phase of TCAS for general aviation orcommuter carriers is planned for implementationin 19950

19If cooperative IFF systems are to be used by

allies or even by U.S. forces operating in alliedcountries, then frequency allocation problemsmust be resolved. For example, the current MarkXII transponders can cause interference in Ger-many, where their use comes under some restric-tion. The converse problem is getting everyoneonto the same frequency. For example, communi-cations between military and civilian aircraft arenot always easy since military fighters primarilyuse UHF frequency bands for communicationwhile commercial airliners use VI-IF bands.

IDENTIFICATION OF SHIPSThe problems of identification of friendly and

enemy ships are more like those of airbornetargets than they are like those of land surfacetargets. Ships use a combination of tacticalinformation and direct target identification toseparate friends from foes.

Navy ships now have both Mark XII interroga-tors and transponders that allow positive friendlyidentification. The ships are equipped with so-phisticated navigational and communications equip-ment, thus antifratricide efforts among shipsthrough sharing of tactical and navigationalinformation comes naturally.

Chapter 2 related several historical incidents offriendly fire between ships, and these are notunknown today .20 With the strong emphasis oncarrier-based airpower, however, U.S. ships prob-ably face a greater fratricidal danger from friendlyaircraft. The consequences would be much moresevere than for attacks on other aircraft butfortunately the likelihood may not be as great.Since the United States and its European allieswill have total control over the sea in most

foreseeable conflicts, there may be theaters inwhich there is no reason for a friendly aircraft toattack any ship.

There are several reasons to believe thatavoiding fratricide of ships will be easier, or nomore difficult, than avoiding fratricide of aircraft.Modem warships are not as widely proliferatedaround the world as are aircraft. Ships can affordto carry more powerful transponders, surveil-lance, and communications equipment. Ships arelarger and slower than aircraft so some ap-proaches to noncooperative identification of air-craft, for example, high resolution radar, shouldwork at least as well. Some noncooperativetechniques, for example, analysis of radio emis-sion should work for ship identification just asthey do for aircraft identification, although shipsand aircraft will have different rules about radioand radar silence. Other techniques, jet enginemodulation, for example, are simply not applica-ble. But still others might work better for shipsthan aircraft: examples include, synthetic aper-ture radars, laser radars, and high resolutioninfrared imaging. Because the country has farfewer ships than aircraft, cost per platform is lessimportant for ships than for aircraft. Furthermore,while perhaps not spacious, ships have much lessof a problem with packaging, power supply,waste heat removal, and antenna placement thando fighter aircraft.

9 SubmarinesThere is little current concern about fratricide

of submarines. Few nations possess them andmost that do are U.S. allies, so all submarinescould be assumed friendly in most limited con-flicts. Furthermore, U.S. submarines are substan-tially different from those of other nations. In thefuture, the problem may become worse, however.Some countries hope to export small diesel-electric submarines. One can easily imagine a

19 U.S. Dep~ent of Transportatio~ Federal Aviation ~“ “stration, “Introduction to TCAS II,” March 1990.ZO Dtig NATO exercises in october 1992, the U.S. aircraft carrier Saratoga accidentally fwed missiles at the Thrkish destroyer MU~VefIeL

killing five and wounding at least 15. See Eric Schmitg “U.S. Missiles Hit ‘Ih.rIcish Ship, Killing 5,” New York Times, Oct. 2, 1992, p, A~.

Chapter

conflict like that in the Persian Gulf War, takingplace perhaps a decade hence, in which both localallies and local enemies are equipped with thesame German- or Russian-built submarines.

In principle, submarines could use cooperativeand noncooperative IFF systems analogous tothose used on aircraft. But whereas the aircraftsystems would depend primarily on radio-frequency electromagnetic energy, submarineswould use sound waves, magnetic fluctuations, orsome other signature. But submarines are differ-ent. The hard part of killing a submarine is notattacking and destroying it—although that can bea challenge under some circumstances; the hardpart is finding it. Submarine designers devoteconsiderable attention to further increasing thedifficulty of detecting submarines, largely bymaking them as quiet as possible. Submarinerswould strenuously resist any effort to havesubmarines actively broadcast in response tosome identification hail, or to somehow increasetheir boat’s signature to make cooperative-passive identification easier.

For the foreseeable future, therefore, avoidingsubmarine fratricide will be based on the currentprocedures for careful command and control ofareas of operation. Submarines now patrol indi-vidual, well-defined areas; any other submarinethat enters that area is considered hostile and opento attack. Moreover, within assigned patrol areas,surface ships do not engage in antisubmarineactivity (that is left to the resident submarine), butany submarine found outside of assigned areas isassumed hostile and could be attacked by surfaceand airborne forces. This approach works but isvery flexible only with good communicationbetween submarines and surface ships. If as-signed patrol areas can be reliably updated severaltimes a day, ships above and below the surfacecan cooperate smoothly.

4—Avoiding Fratricide of Air and Sea Targets 67

A COMPREHENSIVE IDENTIFICATIONSYSTEM

No cooperative IFF system is perfect, so pilotsand ships’ crews hesitate to fire based solely onthe assumption that lack of an IFF reply means atarget is an enemy. Any shooter would prefer tohave positive evidence that a target is an enemy,hence the motivation for noncooperative IFFsystems. Yet neither are noncooperative systemsperfect. Exercises on test ranges indicate that withcurrent technology a few percent of targetsidentified as hostile are, in fact, friendly. Thiserror rate may be acceptable in a struggle forsurvival, as was expected between NATO and theWarsaw Pact. However, if in the future the UnitedStates is engaged in a lesser conflict with over-whelming force ratios, then most of the aircraft inthe air at any given time will be friendly and evena small rate of mistakenly identifying friends asfoes can result in the loss rates from fratricidebeing as high as those from enemy action.

Cooperative and noncooperative systems canwork together to make each more effective thaneither could be alone. Assume that a cooperativequestion-and-answer type system were, say, 90percent reliable. If this system were used to sortall targets, then 10 percent of targeted friendlyforces would be classified as hostile, clearly anunacceptably high value. But if a noncooperativesystem were also 90 percent reliable, that is, itclassified friendly vehicles as hostile 10 percentof the time, and all hostile-classified targets werealso queried by the cooperative system, then only10 percent of 10 percent, or 1 percent, would slipthrough both, assuming the two systems areindependent. 21 These numbers are picked only tobe illustrative, whether a 1 percent error rate isacceptable will depend on the tactical situationand political and military judgment.

Cooperative systems can even help identifyenemies, if operated in conjunction with nonco-operative systems. Many noncooperative identifi-

21 In the red world, the systems may not be independent. Batde damage, for example, might both change the noncooperative signature andput IFF transponders out of commission. In war, nothing is neat.


cation techniques will discriminate friend fromfoe on the bases of subtle differences, but willrequire expensive, hence perhaps scarce, equip-ment. If this equipment is unable to examine allpossible targets, then cooperative question-and-answer systems could concentrate efforts onambiguous targets. For example, only thosetargets that do not respond to a cooperative IFFqueries would be examined by noncooperativetechniques.

Currently, the technology required for coopera-tive question-and-answer IFF systems is moredeveloped than that for most noncooperativesystems. The Services see the advantages ofnoncooperative IFF but that longer term goalshould not erode efforts to improve the reliabilityof cooperative question-and-answer IFF by theapplication of technology that is near-term orin-hand.

AvoidingFratricide of

Land SurfaceTargets 5

A11 of the U.S. fratricidal casualties in the Persian GulfWar were among land forces. Near-absolute dominanceof the air and Iraq’s inability to project power to seaallowed conservative rules of engagement in those

media, which avoided accidental attacks on friendly forcesexcept in a couple of cases that did not result in casualties.

The fratricide of the Persian Gulf War was unusual comparedto that of past wars. As pointed out in chapter 2, the most strikingaspect at first was the apparently unprecedented high fraction oftotal casualties resulting from fratricide. In addition, however,the importance of each type of fratricide was different from otherlarge mechanized land battles. In World War II, for example, themost deadly fratricide were the result of aircraft bombingfriendly troops. Surface-to-surface fratricide resulted most oftenfrom indirect fire in which artillerymen fire at a target that theycould not see, The Persian Gulf War, in contrast, had anunusually high fraction of fratricides from direct-fire weapons,such as tanks, shooting mistakenly at other land targets.

Most of the U.S. personnel were mounted in vehicles; thus, notsurprisingly, most of the fratricide occurred when vehicles werehit, so that current emphasis is on protection of vehicles, not on [

that of individual infantry. Helicopters are included in this ~

chapter; although helicopters are aircraft, their operation andemployment gives them more in common with surface vehiclesthan fixed-wing airplanes, at least as far as fratricide technologyand equipment are concerned.

The next sections discuss general approaches to avoidingfratricide of ground targets, a number of specific technologiesavailable to implement these approaches, and some of theadvantages and disadvantages of each approach and technology,

69

. . . . . .


● ✎ ● ☛☛ ✍ ‘ “ *

..- < “.’

% -\ .. ‘-:&’ - - .-% ● . * - . *- 6

. , .;- -, .

. -.%’ “‘.

\ . .-,.- ***

“-. a..-. *..%“ 2- % k - *. “-

US. armored vehicles destroyed by friendly fire during the Persian Gulf War were recovered and collected,

INCREASING KNOWLEDGE OF THETACTICAL ENVIRONMENT

Reviews of past cases of fratricide, includingthose from the Persian Gulf, show that a primeculprit is the shooter’s poor understanding ofwhere friends are, and even where he is himself.A unit spreads further confusion among neigh-bors when—instead of being flatly lost—it be-lieves it knows exactly where it is and reports anincorrect location to other units. Thus, navigationand communication are two vital keys to avoidingfratricide.

B NavigationCurrent multimillion dollar U.S. tanks do not

have compasses. This may seem astonishing untilconsidering that magnetic compasses are uselessinside 60-ton metal boxes. Alternatives-like

—have been prohibitively expen-gyrocompassessive in the past and were often unreliable in therough environment of a tank. Nevertheless, thefirst step for improving tactical knowledge of

mobile units requires improving their naviga-tional tools. Fortunately, new technical develop-ments make this easier.

During the Persian Gulf War, tanks usedequipment—including some off-the-shelf com-mercial equipment—that calculated latitude andlongitude from data received via radio pulsesfrom a network of satellites, the so-called GlobalPositioning System or GPS. If desired, the sameprinciple could be applied on a local tactical scaleusing airborne transmitters.1 Gyroscope com-passes provide direction, and some, in particularring-laser gyroscopes, are improving in accuracyand coming down in cost. Current plans call forbroad use of GPS and compasses, or azimuthindicators, as the Army calls them.

Chapter 2 showed location uncertainty to be aprime cause of fratricide from artillery. Global orlocal positioning systems could provide artillerybatteries and their forward observers with accu-rate, consistent coordinates. Some proposals call

1 Briefing “Very Imw Frequency Identification Friend or Foe, ” Willie Johnsou U.S. Army Communications and Electronics Command,Mar. 10, 1992.

,

— —

Chapter 5-Avoiding Fratricide of Land and Surface Targets 71

for use of positioning systems to increase fireaccuracy and perform last-instant IFF.

E CommunicationKnowing location is just part of the job of

avoiding fratricide; each individual and unit mustinform nearby units of its location. One approachis for each unit or vehicle to send its location tosome central distribution point. This might be aground-based or airborne communications node,or it could be a satellite, perhaps the same one thatprovides the location broadcasts as illustrated infigure 5-1.

Communication of location could also takeplace through shorter range networks, as illus-trated in figure 5-2. Each vehicle could be a nodein a interwoven network of communicating vehi-cles. Each vehicle would transmit informationabout itself as well as its neighbors. Thus, notevery pair of vehicles would need to be inconstant direct communication, In the figure, forexample, tanks numbered 3 and 5 might not beable to communicate directly because of interven-ing terrain, but they can communicate by mes-sages relayed via numbers 1 and 4. When tanks 3and 5 then come into sight of one another aroundthe ridge, neither should be surprised. Systemsusing these principles are now in development.

As one analyst has pointed out, this is a reversalof how sensor networks have operated in the past:sensor systems have been developed to findinformation about the enemy, they collect infor-mation about the enemy and all the friendly forceswithin range, and then-with only slight exagge-ration—the information about friends is thrownaway.2

Networks would allow the propagation ofinformation about foes as well as friends. Thus,information about a hostile or ambiguous vehiclesighted by any member of the network would beavailable to each member of the network. In

and

‘d


principle, this type of communication allowsobservers to ‘‘compare notes, ’ develop a com-posite picture, and thus get a better overallidentification. For example, one observer mightsee an ambiguous vehicle, then try, and fail, to geta reply to a IFF query. Another might have limitedinformation on its appearance, another detectedsome suspicious radio signals, and so on. No pieceof information by itself is definitive, but all of ittogether might be. The range of the communicationlinks is not a problem; indeed, considered strictlyfrom the point of view of avoiding fratricide, therange does not need to be much more than, say,double the range of the vehicles’ weapons.

A simple navigation-communication systemcould be based on the current standard Armytactical radio, the Single Channel Ground andAirborne Radio System (SINCGARS).3 Moreelaborate data exchanges will require radios withhigher data rates. Almost any system would bemost useful with some way to display the locationinformation inside the vehicle graphically.

2 Briefing entitled, “Fratricide Prcventiou ” Mark Fine, U.S. Army Communications and Electronics Command, Mar. 10, 1992.3 See ‘‘Combat Identification, Data Pose Battlefield Awareness, ” Signa/, November 1992, pp. 37-39 and Mark Tapscott, “SINGARS

Upgrades Moving to Full Battlefield Communications, ’ Dqfense Electronics, vol. 25, No. 1, January 1993, pp. 29-32.

- . - — - . . . . - . . . - . . . . — - . . -.


Figure 5-2—A Ground-based CommunicationsNet for Exchange of Location and

Identification Information

@


M IFF Through Exchange of LocationInformation

Navigation and communication systems can beused as a type of IFF device, one in whichidentification is based not on vehicle characteris-tics or the exchange of codes, but on locationinformation. Take as an example a direct-fireweapon like a tank. A navigation system couldprovide each tank with an accurate position, anda communication system could provide the posi-tions of all friendly tanks within range. As part ofthe current firing sequence, the tank gunner usesa laser range-finder to determine the target’srange. While current standard equipment on tanksdoes not provide gun bearing, that is, the direction

in which it is shooting, the navigation systemcould provide that information as well.

The tank gun’s computer or free-control sys-tem, knowing its own location and the range andbearing of the target, can quickly calculate theexact location of the target. This location can becompared to the locations currently reported byfriendly vehicles and, if there is no match, thetarget could be assumed hostile. (One might wantfurther conflation in case it is a friend withmalfunctioning equipment.)

If the current register of locations shows amatch, the answer is somewhat ambiguous. Amatch only means that a friendly vehicle is at thesame location, within the accuracy of the system,not that the target in the gun sights is a friend. Theaccuracy may be dozens of meters and the truefriend may be obscured behind a clump of trees.

Other approaches following this general themeare possible. For example, just before the gunfrees, a radio or laser signal could be sent outcontaining the message: “Attention! I am aboutto shoot at a target at the following coordinates,if you are sitting on that spot, you should tell menow.” Obviously, the message, and reply, wouldhave to be encrypted to keep the enemy fromexploiting the system. The speed of the system isalso important. Tank crews are trained to get around off within ten seconds of first detecting atarget; good crews can do it within six seconds.Any IFF procedure that takes “only” a secondsignificantly increases engagement time.

IDENTIFICATION SYSTEMSEven a very effective navigation-communica-

tion system probably will be unable to offerperfect friend and foe identification; an additionalsystem devoted to IFF may also be needed. Thesimplest cooperative IFF system allows the ob-server to remain passive, as illustrated in figure5-3a. The target vehicle might use characteristicmarkings to identify itself as friendly. In thePersian Gulf War, the vehicles were marked withan inverted ‘V’ made with an infrared-reflective

——— — ——— -- —. —— —— —


Figure 5-3-Approaches to Identification ofFriend and Foe

‘“*

“ *

c’*

‘“*SOURCE: U.S. Army.

4

r, ,7“ \ \

‘ /r

tape or paint that showed up on infrared imagingdevices while being fairly unobtrusive visually.

1 BeaconsThe target vehicle might instead broadcast a

signal in all directions, identifying itself asfriendly, as illustrated in figure 5-3b. This is theprinciple of the so-called “DARPA Lights” and“Budd Lights.”4 Attaching flashing lights tocombat vehicles may seem unwise, since theyreveal the vehicles’ position to the enemy as well,

but these particular lights are only visible in thenear-infrared. In the Gulf War, the Allied forceswere widely equipped with infrared image-intensifiers and the Iraqis were not. In addition,the DARPA Light is fitted with an adjustableshroud that blocks ground observers’ views of thelight while keeping it visible from aircraft. Thissimple security measure was useful, again, be-cause of limitations of the enemy: the Allies hadtotal air dominance leaving no Iraqi airborneplatforms to observe DARPA Lights from above.(DARPA Lights were acquired and ready todeliver to the Gulf, but the conflict ended justbefore shipment.)

The Thermal Identification Device, or TID, isanother beacon system in limited current produc-tion. The TID is a simple bent metal sheet in theshape of a roof. The sheet is heated electrically onone side to generate a thermal signal. The otherside reflects the cool thermal image of the sky.The sheet is mounted on a short mast and anelectric motor rotates the sheet. Thus an observerseeing the TID through an infrared image will seean alternating bright and dark spot.5 See figure5-4. The disadvantages of the TID include lack of

The ‘‘DARPA Light’ is a simple infrared beaconvisible only through night-vision goggles.

4 me -e @@t ~ve ~~ck &,mu5e of its s~~~ t. tit of a pop~ ~t bevemge, but the trmmitter was ori@mlly so named titer

its inventor, Henry ‘‘Bud’ Croley at the U.S. Army Night Vision Laboratory who, along with Wayne Antesberger, holds U.S. patent number4862164 on the device.

s Bnefmg entitled, ‘‘Thermal Identi17cation Device (TID) for Combat Identificatio~’ John R. Gres@ Night Vision and Electro-optics,Night Vision Laboratory, Ft. Belvoir, Virginia (undated).

. ——. .- . . . . . . . - . . . . . . . . .-


Figure 5-4-The Thermal Identification Device (TID)

SOURCE: U.S. Army.

security (any enemy with the proper thermalsights can see it as well as other friends can) andunreliability (apparently the exhaust plume fromthe engine of the M-1 tank can obscure the TIDfrom certain angles).

Near-infrared blinking lights, such as theDARPA and Budd Lights, have two relateddrawbacks. First they are visible to anyone, friendor foe, with near-infrared night-vision image-intensifiers, which are cheap and widely available

Manual control

Laser activation

through the mail for about a thousand dollars.Thus, any enemy might have at least a fewdevices that could detect the beacons. Second,they are not visible to the far-infrared imagingdevices used to aim the guns and other weaponson U.S. armored vehicles. The TID avoids theseshort-comings. Whereas the near-timed imageintensifiers are relatively inexpensive, far-infrared viewers are a hundred thousand dollars ormore, and correspondingly fewer countries have

Chapter 5-Avoiding Fratricide of Land and Surface Targets [ 75

the devices widely deployed. The TID alsoradiates at the wavelengths at which the far-infrared targeting viewer is most sensitive.

Future systems operating on the same principlecould use continuous omnidirectional radio trans-missions. Signals sent by radio would have longerrange than visible or infrared light, especially infog, rain, dust, or smoke. Of course, specialprovisions must be made to keep an enemy fromintercepting and exploiting the signals. Possibili-ties include spread-spectrum signals-describedin the previous chapter--or time-synchronizedsignals. Either approach forces an enemy to listento radio noise over a wide range-of frequencybandwidth or time-and filter out a signal, whilefriendly forces, knowing the waveforms or thetime sequence pattern, can listen only for thesignal.

I Question-and-Answer SystemsHaving friendly targets broadcast only when

interrogated by another friendly shooter providesadditional security against enemy intercept andexploitation of identification signals. This ap-proach is illustrated schematically in figures 5-3cand 5-3d.

QueriesThe query signal could be a laser or some radio

pulse. The query pulse should be directional,otherwise on a complex battlefield each tank’squery would set off all tanks’ responses andmatching up queries and responses would behopelessly complicated.

The queries should also be encrypted to authen-ticate them as coming from friends. This keepsthe enemy from exploiting the system. The enemymight, for example, suspect that U.S. vehicles liecamouflaged along a tree line but not knowwhere. If sweeping the whole treeline with a laseror radio pulse caused all the U.S. weapons to sendout IFF replies, they would reveal their positions,

so the response must come only after receipt of avalid, encrypted query.

A potentially cheap IFF approach would beadapting the existing laser range-finders nowfitted on U.S. tanks. A danger of this approach isthat the laser signal will not reliably penetratesmoke, fog, and dust. This may not cause aproblem for the laser’s range-finder function—after all, ranging the target, the dust cloud aroundit, or the tree next to it is good enough to get a hitwith a flat trajectory weapon like a tank gun. Butif the laser did not penetrate intervening dust, thenthe target would never get the query and wouldnever send off a friendly response.

The query signal would be sent through asimple directional antenna aligned with the weapon.It could be, for example, fixed to the barrel of thetank gun or to the turret much as the laserrange-finder is now. The same simple directionalantenna would then naturally be lined up to pickup the signal from the particular target in questionand not any other.

The Army’s current near-term solution will usea millimeter wave radio beam as the query signal.Millimeter waves can penetrate obscurants, likedust, and they can be fairly narrowly focused.6

Replies

The reply signal could be any of those de-scribed above for a passive observer with muchthe same advantages and disadvantages. Like thequery, the reply signal would have to be en-crypted in some way that the enemy could notreproduce,

Since the reply would be sent out only when thesystem is properly queried, somewhat less atten-tion could be given to avoiding intercept by theenemy. For example, some proposed systemswould determine the general, but not exact,direction of the query and return a bright, broadlaser beacon to show that the target is friendly,accepting that the beacon would sometimes be

6 U.S. Army, Combat Identification System Program Office, Ft. Meade, MD, ‘‘Report to the Committees of the United States Senate andHouse of Representatives on the Combat Identification Program, ” January 1993.

— . . -. —.- . - . . . .- . . .


seen by enemies. Alternately, omnidirectionalradio beacons, illustrated in figure 5-3c, couldannounce that the target is friendly. Omnidirec-tional broadcast is much simpler because itremoves any requirement for steering an an-tenna—very quickly!—in a particular direction.

A directional reply, shown schematically infigure 5-3d, would provide additional security bymaking enemy intercept less likely. One cleverapproach would use the interrogating laser as itsown reply signal by reflecting the laser back to itsorigin using a comer reflector. (A corner reflectoris just a set of reflective surfaces in the shape ofthe inside of the comer of a cube. The geometryof the surfaces is such that no matter whichdirection a light beam enters, it is reflected backout in exactly that direction.) The reply would beauthenticated by modulating or chopping thequery pulse in some way, for example, by turningon and off liquid crystal windows covering thecomer reflector. This particular approach, usinglasers, would be unreliable in smoke or dust.Radio frequencies could penetrate better andcould use the same principle with the authentica-tion provided by vibration of the reflectivesurfaces to produce a detectable Doppler shift inthe return signal or by rapidly changing theimpedance of the reflecting antenna-similar to

the pre-World War II proposals for varyingdipoles on aircraft.7

DISMOUNTED INFANTRYMost of the effort for IFF devices has centered

on the identification of vehicles, not people. Atpresent, programs examining exclusively theproblem of identification of dismounted infantryare in the planning stage. The Army intends tofund dismounted infantry IFF programs in thecoming and subsequent fiscal years. Infantryalmost always work closely with vehicles of somesort and any vehicle-mounted system will alsohelp prevent mistaken attacks on friendly infan-

These three images are of the same tank on a testrange. The top is a close-up taken in visible light atmidday. The tank as seen at night through infrared‘‘image intensifiers’ at a range of 500 meters is shownin the middle image. The bottom image is the tank seenthrough the same device but at 1500 meters. Modernoptics and electronics allows the detection of vehiclesbeyond the range at which they can be reliablyidentified.

T Briefing entitled, “Achieving Covert Communications and Ground-Combat Identification Using Modulated Scatterers,’ E.K. Miller andD.M. Metzger, Mechanical and Electronic Engineering, Ims Alamos National Laboratory (Mar. 11, 1992).


try. Moreover, navigational and communicationsimprovements that increase tactical knowledgewill reduce fratricide of infantry as well asvehicles, even without specific new IFF capabili-ties.

One of the quick fixes used in the Persian Gulfand mentioned already was the small, battery-powered blinking infrared light called the BuddLight. Its flashing is invisible to the unaidedhuman eye but shows up clearly at night throughnight-vision goggles or other near-infrared image-intensifiers. The device is cheap and easilyproduced by components from retail electronicsstores. Being so widely available, it hardly countsas an IFF device but it can be a useful commandand control device. For example, one combat unitmay know that another unit is somewhere in frontof it in a group of trees but not know exactlywhere. Getting the two units coordinated bydescribing features at night over the radio isdifficult, but a request to turn on the infraredblinkers for five seconds could make relativepositions clear in an instant, and thus help avoidfratricide.

Future research will include work on tech-niques specifically directed toward infantry iden-tification. Some of these ideas are modern incar-nations of World War I infantry identificationefforts that included sewing mirrors on the backsof trench coats to aid identification by friendlyaircraft overhead. Notional proposals include:fabrics that reflect millimeter waves or fabricsand dyes that reflect only very specific wave-lengths chosen to match up with the wavelengthsdetected by targeting sensors; fabrics and dyesthat luminesce under specific laser illumination;retro-reflectors on combat uniforms; and activeinfrared displays on the uniforms.8

‘‘BuddLights’ are invisible to the naked eye but showup clearly through night-vision goggles.

NONCOOPERATIVE IDENTIFICATIONDevelopment of noncooperative identification

of ground targets is perhaps at an even earlierstage than it is for air targets, but some ideas havesurfaced. Just as for air targets, the simplestidentification technique is based on the outwardappearance of the vehicle. Again, as for airtargets, this simple approach is made very com-plex by the proliferation of U. S., British, French,Russian, and other weapons throughout the world.

A straightforward improvement in identifica-tion capability would come from improvement inimaging devices, for example, higher resolutionfor infrared detectors, laser radars that providethree dimensional images, and so on.

More subtle clues to identity might be providedby, for example, vibrations of the surface of thevehicle, detectable by Doppler radar. Dopplerradar shows some promise for the identificationof helicopter because of the characteristic Dop-pler shifts caused by the rotating blades.9

Sensors might be able to pick up characteristicsof the vehicle exhaust, even detecting differencesin fuel type,10 or sensors could look at the

8 ~ ~I~Omatlon paper: Identification Friend or Foe for the Individual soldier, ’ Robin St. Pere, U.S. Army, Natick RD&E Center, Oct. 6,

1992.

s Briefiig entitled, “Radar Helicopter Ident~lcation, ” Gcrardo Melcndez, EW/RSTA Directorate, Us. -Y CO~unications ~dElectronics Command (undated).

10 Briefing entitled, ‘ ‘Laser Radar Effluent Sensor,” Guil Hutcheson, Ims Alarnos National Laboratory (undated).

— — — — . . . . . . . .- - - . . . - . . .


reflectivity of the vehicle in several spectralranges. ll For ground combat application, each ofthese techniques that depend on detection ofsubtle spectral differences should be approachedwith a healthy skepticism, since transmission ofradiation through the dirty, dusty, smoky, foggyair of the battlefield will often be the fundamentallimit to their performance. In addition, one mustconsider the effects of trees, buildings, andintervening terrain.

Some sensor systems have been developed todetect enemy vehicles on the battlefield. Forexample, the Remotely Monitored BattlefieldSensor System, or REMBASS, uses a combina-tion of acoustic, seismic vibration, magnetic, andinfrared sensors to detect vehicles. With someadded sensitivity, each of these techniques mightbe able not just to detect, but identify, vehicles.These sensors are typically short range and fixedso they are probably better at providing tacticalintelligence about enemy vehicles to a communi-cations network than at providing the informationdirectly to a shooter.

CURRENT PROGRAMSA quick review is worthwhile to repeat what is

hypothetical and what is under active considera-tion or development. The Army divides itsdevelopment efforts into four time categories.The “Quick Fix” was intended to get somethinginto the hands of troops immediately. Quick fixesinclude the Budd and DARPA Lights, receiversfor satellite transmissions of global positioningdata, and the infrared-reflective “thermal” tapethat were rushed into service during the PersianGulf War. The Thermal Identification Devicedescribed above is another quick fix.

After the Persian Gulf War, the Army acquired20,000 Budd lights, which are to become logistic

stock items, and contracted for 120,000 squarefeet of thermal tape. The Army also has the 8,000DARPA Lights acquired during the Persian GulfWar. The Army will acquire 300 TIDs, or theparts required for rapid assembly. Production willbeat Tobyhanna Army Arsenal. This will provideenough TIDs for an armored brigade. (Note alsothat the extreme simplicity of the TID shouldallow rapid surge production.) The Army also hasbought, or is in the process of buying, over 10,000global position receivers called Small Light-weight Global Positioning Receivers, or SLGRs(pronounced “slugger”).

The near-term program aims to get somethingin the field in five years or so. The ideal near-termsystem should require little or no additionalresearch and development. A request for propos-als sent to industry solicited numerous responses.A test at Ft. Bliss, Texas examined many of theproposed systems using actual combat vehiclesunder realistic desert conditions.

12 The result of

that test was selection of a millimeter wavequestion-and-answer system for near-term devel-opment and deployment. One of the disadvan-tages of a millimeter wave system is that it will bedifficult to incorporate into airplanes, but compat-ibility between aircraft and surface vehicles isonly a desirable characteristic for near-termsolutions and a requirement only for far-termsolutions. The Army has set a limit of $100million for total costs to produce approximatelyenough identification devices for 1,500 vehicles,not enough to outfit the whole army but enoughfor a substantial contingency force.

Mid- and far-term solutions are intended toprovide a more permanent solution in seven ormore years. These programs are now in theexploratory stage.

11 ~e~e ~otio~ approaches are taken for a brief~ presented by Wayne Gr~6 “Electro-Optics (E-O) lkchnology for Positive TargetIdentification%” United States Arrny, Night Vision and Elecro-Optics Directorate, Fort Belvoir, VA (un&ted).

IZ Tfiing and Doctrine Co remand, Combut ID Tech Demo, June 9, 1992. Because of limitations of the desert test sites, performance infog, ra@ or forest could not be tested.

Chapter 5--Avoiding Fratricide of Land and Surface Targets 179

TACTICS, DOCTRINE, AND TRAININGThe Army has not given as much explicit

attention to avoiding fratricide in the past asmight be expected, but this is because troop safetywas a natural, integral part of coordination andplanning. The Army now believes, however, thatfriendly fire requires special treatment in tactics,doctrine, and training.

Some changes in tactics are possible. Forexample, most of the Persian Gulf fratricidesoccurred under conditions of reduced visibility,either darkness or dust and haze. If the goal wereto reduce fratricide, then one easy solution wouldbe to never attack except under clear daylightconditions. Remember, though, that low visibilityfor U.S. forces often translates into no visibilityfor enemy forces not equipped with infrared orimage-intensifying viewers. Night attacks reducethe enemy’s effectiveness, including his ability toinflict casualties, and thus reduce casualtiesoverall.

Nevertheless, technical advantages enjoyed bythe United States, compared to a wide spectrum ofpotential adversaries, could allow changes intactics. For example, in the Persian Gulf, U.S.forces could often see, hit, and kill Iraqi targetsthat could not even see the forces shooting atthem. Under these conditions, more cautious,deliberate attacks might be able to keep bothfratricidal and enemy-inflicted casualties down.

Training is extremely important to avoidingfratricide in future conflicts. The Army’s trainingcenters now pay special attention to fratricideincidents and collect the information for anon-going “lessons-learned’ study. Firing rangesare now equipped with both enemy and friendlytargets to practice ‘ ‘Don’t Fire! situations.

Simulation is an important part of moderntraining. In the past, simulators have been muchbetter at depicting the day-lit world than the nightworld seen through infra-red viewers. Yet most ofthe fratricide in the Persian Gulf occurred under

some sort of reduced visibility. Simulation fortraining is a rapidly progressing field and simula-tion of poor-visibility conditions must be sup-ported in future programs.

CONCLUSIONSGround combat illustrates most clearly the

vital importance of tactical knowledge of thebattlefield to avoiding fratricide. Indeed, by someestimates, the majority of fratricide could beavoided by improvements in navigation andcommunication without a dedicated IFF system.For the foreseeable future, however, IFF deviceswill probably also be desired to compensate forgaps in tactical knowledge.

Navigation-communication systems and IFFsystems will be pursued in parallel but, whencomparing costs, one should keep in mind thatavoiding fratricide is just one advantage ofimproved knowledge of the battlefield. Betterinformation will increase maneuverability, flexi-bility, control and coordination of units and fire,and, hence, overall combat capability.

The Army’s preferred near-term solutions willbe difficult and expensive to incorporate intofixed-wing aircraft, This alone might make itunacceptable as a permanent solution if fixed-wing aircraft are expected to continue to provideclose air support. One of the important policyquestions is how to allocate resources betweensolutions that provide quick, but limited, resultsand more permanent solutions, which admittedlywill take longer to implement.

The Army’s technology and equipment toavoid ground combat friendly fire is primitivecompared to Navy and Air Force equivalents.Army programs may need preferential funding forseveral years just to catch up to the level enjoyedby its sister Services today. The distribution offratricides in the Persian Gulf argues for thisrelative shift in effort, at least until imbalances areless pronounced.

.— —— ——

Absolute position, 35

“Accidental self-destruction,” 8Accidents, fratricide due to, 9-10Active systems, 39-40. See also Cooperative question-and-

answer systems; Noncooperative question-and-answersystems

Air targets, avoiding fratricide of, 49-66

Aircraft identification, history, 54-55

Allied operations, 4, 29encryption keys and, 44

Amicicide, 8Antifratricide systems

criteria, 45-46cross-Service coordination, 6doctrine, 47emergence, 2short-term and Iong-term goals, 5-6tactical environment awareness vs. IFF, 46training and simulation, 46-47

Army Training and Doctrine Command (TRADOC)antifratricide measures, 46fratricide definition, 9

ARSA-4 system, 65Attack sequence, 32-33Avoiding fratricide. See also Antifratricide systems

in air, 49-66

general considerations, 29-48on ground, 69-79at sea, 66-67

AWACS, 50

Battle management, 33,49-50Beacons, 73-75Black boxes, 46Bradley, Gen. Omar, 13, 15Bryan, C. D. B., 8Budd Lights, 73-75,77,78

Index

CALL. See Center for Army Lessons LearnedCallaham, Lt. Commander, 20Casualty report forms, 8Casualty surveys. See Combat casualty surveysCauses of fratricide

findings, 3historical review, 7-28

Center for Army Lessons Learned (CALL), 19, 25-26CID. See Combat IdentificationCivil airborne transponders, 62-63

Mode S systems, 63-65Civil-military coordination, 62-66Classification, attack sequence step, 32Combat casualty surveys, 1,21-23

data, 23Combat Identification (CID), 33Command and control failures, 10-15, 28Commercial carriers, Mode S system and, 64Communication

in battle management, 49-50in tactical environment, 34, 71, 79

Communication failures, 28Commuter carriers, Mode S system and, 64Congressional issues, 5-6

Contests of less than national survival, considerations, 1,31

Cooperative question-and-answer systemsadvantages and disadvantages, 42aircraft identification and, 56-57limitations, 59-62passive cooperative IFF measures, 59passive measures, 59queries, 75replies, 75-76

Coordination failures, 28Cross-Service coordination, 4,6

81

. ——


DARPA Lights, 73-75,78Decision to attack, attack sequence step, 33Definition of fratricide, 9Denial, 43,61Detection, attack sequence step, 32Direct system of information, 39Dismounted infantry IFF programs, 76-77Doctrine, 33,47, 79Doppler radar, 59,77Dye-loaded shells, ordnance identification, 21

Eisenhower, Gen. Dwight D., 14Electromagnetic spectrum, 40,41Encryption. See Information securityEngagement ranges, 30,37Equipment design, 10Europe, Mode S system and, 63Examples

accidents, 9-10Civil War, 17-18command and control failures, 10-15fear of fratricide effects, 19-20fire discipline failures, 15-16identification failures, 17-18navigation failures, 16-17World War II, 12-15

Explicit rules of engagement, 37-38Exploitation, spoofing, and denial, 43,60-61

FAA. See Federal Aviation AgencyFatalities. See LossesFear of fratricide, tactical constraints due to, 18-20,Federal Aviation Agency (FAA), joint military ATC

operations, 65-66FEZs. See Fighter Engagement ZonesFighter Engagement Zones (FEZs), 53Findings, 2-5Fire discipline failures, 15-16,28Force ratio, 30Fratricide Incident Report, 46Free-fire zones, 12Frequency allocation, 65-66Frequency trends, 30-31Friendly Fire, 8Ft. Bliss, Texas, testing, 78

General aviation, Mode S system and, 64General Officers Steering Committee, 6Global position receivers, 78Goto, Adm., 19Ground targets, avoiding fratricide of, 69-79Guderian, Gen., 15

High-resolution radars (HRR), 51-53Historical review of causes

combat casualty surveys, 1, 21-23general considerations, 7-9Persian Gulf War, 7-8,26-28

prevalence of fratricide, 20-22Summary, 26-28tactical constraints due to fear of fratricide, 18-20,30training center data, 24-26types of fratricide, 9-18

Hooker, Gen. Joseph, 17Hopkins, Dr. James, 21HRR. See High-resolution radarsHuman error, accidents due to, 10

Identificationcommand and control failures and, 11-12findings, 4improved, 30

Identification failuresexamples, 17-18, 37types, 35-36

Identification of Friend and Foe (IFF), 33advantages and disadvantages of systems, 42-43exchange of location information, 72exploitation, spoofing, and denial, 60-61frequency allocation and, 66information security, 43-45and knowledge of tactical environment, 34-35Next Generation IFF, 59-62range and, 38rules of engagement, 35-38sources of information, 39-40system jamming, 61

Identification systems, 67-68,72-73IFF. See Identification of Friend and Foe

30 Implicit rules of engagement, 38Indirect fire weapons, 22Indirect system of information, 39Infantry, 75-76Information security, 43-45

frequency bands and, 62Mode 4, 57-59Mode S and peacetime intelligence loss, 65question-and-answer systems, 75-76

Interrogators and transponderscivil airborne transponders, 62-65Mark X protocol, 56-57Mark XII protocol, 56-62,66-67Mark XV protocol, 56-62TACIT, 60

Jackson, Gen. Thomas “Stonewall,” 17-18JADO. See Joint Air Defense OperationsJamming, IFF and, 61-62Jet engine modulation, 51JEZs. See Joint Engagement ZonesJoint Air Defense Operations (JADO), 53,56Joint Engagement Zones (JEZs), 53,56Jungle fighting, 16-17

Knowledge of tactical environment, 33,34-35,46-47,70-72,79. See also Communication; Navigation

Index I83

L-band system, 56,62Land targets, avoiding fratricide of, 69-79Lessons learned. See Center for Army Lessons LearnedLocation information exchange, 72Long-range weapons, 22

IFF and, 38Losses

comparison of numbers, 31fraction of deaths due to fratricide, 1,2, 31

Malfunctions, 9-10Mark X protocol, 56-57Mark XII protocol, 56-59

limitations, 59-62onboard ships, 66-67

Mark XV protocol, 56-59improvements, 59-62

McNair, Lt. Gen. Leslie, 14Medical records, 21Mercer, Lt. Col. George, 7,8MEZs. See Missile Engagement ZonesMid-air collision prevention, 65MILES. See Multiple Integrated Laser Engagement System“Misadventure,” 8Misidentification, 18, 28Missile Engagement Zones (MEZs), 53Mitchell, Gen. William “Billy,” 11Mock combat. See Training and simulationMode S systems, 63-65Multiple Integrated Laser Engagement System (MILES),

24-25

National Trainin g Center (NTC), 9,24-26simulations, 46

Navigation, in tactical environment, 34-35,70-71,79Navigation failures, 16-17,28Next Generation IFF (NGIFF), 59-62No-fire zones, 12Noncooperative systems

advantages and disadvantages, 42-43, 44future of, 50ground targets and, 77-78radar, 51-53radio--emission intercept, 50-51surface-to-air missiles, 53, 56

NTC. See National Training Center

Observer/target energy transmission routes, 39-40. See alsoCooperative question-and-answer systems; Noncoopera-

tivequestion-and-answer systems

Occurrence patterns, 8-28Operation COBRA, 12-14Operation Desert Storm. See Persian Gulf WarOperation QUEEN, 14Operation TOTALIZE, 14-15Opposing forces, mock combat, 24-26Ordnance identification, 21, 30-31

Passive systems, 39-40. See also Cooperative question-and-answer systems; Noncooperative systems

Patriot missile, 53,56Persian Gulf War

equipment, 70fratricide during, 1,2,7-8,26-28,69friendly fire incidents, 1991, 27

Positive identification, 38Prevalence of fratricide, 20-22Psychological effects, 1-2, 31-32Pyle, Ernie, 13

Radar, 51-53Radio-emission intercept, 50-51Record keeping, 8Remotely Monitored Battlefield Sensor System (REM-

BASS), 78Resource allocation, 5Rules of engagement, 5, 33, 35-38

S-band system, 62Sa’adah, Dr. David, 21, 22SARTIS. See Shipboard Advanced Radar Target Identifica-

tion SystemSchrader, Lt. Col. Charles, 8,22Scott, Adm., 19Sea targets, avoiding fratricide of, 66-67Selective Identification Feature (SIF), 57Sensor networks, 71Sensor systems, 77-78Severity, measuring, 31-32Shipboard Advanced Radar Target Identification System

(SARTIS), 56Ships, identification, 66-67Short-range weapons, IFF and, 38Short-term versus long term goals, 5-6SIF. See Selective Identification FeatureSimulations. See Training and simulationSingle Channel Ground and Airborne Radio System

(SINCGARS), 71Situational awareness. See Knowledge of tactical environ-

mentSmall Lightweight Global Positioning Receivers (SLGRs),

78Sources of identification information. See Identification of

Friend and Foe; Knowledge of tactical environmentSpoofing, 43,60Submarines, identification, 66-67Surface-to-air missiles, noncooperative IFF and, 53,56

TACIT See Time Authenticated CryptographicInterrogator/Transponder

Tactical constraints due to fear of fratricide, 18-20Tactical environment. See Knowledge of tactical environ-

mentTCAS. See Traffic Alert and Collision Avoidance SystemTechnical developments

beacons, 73-75


findings, 4infantry identification, 77information security, 43-44resource allocation for, 5review, 78weapon and ordnance improvements, 30-31

Tempo of battle, 30Thermal Identification Device (TID), 73-75,78Time Authenticated Cryptographic Interrogator/Transpon-

der (TACIT), 60TRADOC. See Army Training and Doctrine CommandTraffic Alert and Collision Avoidance System (TCAS),

65-66Training and simulation, 4-5

antifratricide systems, 46-47

NTC mock combat, 24-26review, 79

Transponders. See Interrogators and transpondersTypes of fratricide, 9-18

Washington, Col. George, 7,8WDMET. See Wound Data and Munitions EffectivenessWeapons

aim and ordnance delivery, 33misidentification due to similarities between, 18range, 30, 38types causing injury, 22

Wound Data and Munitions Effectiveness (WDMET), 22

who goes there: friend or foe? - unt digital library/67531/metadc... · who goes there: friend or...

Documents