an experimental gps navigation receiver for general aviation
Post on 09-Feb-2022
3 Views
Preview:
TRANSCRIPT
FAA-RD-83/26
Project ReportATC-121
An Experimental GPS Navigation Receiver for General Aviation: Design and
Measured Performance
S. D. CampbellR. R. LaFrey
27 September 1983
Lincoln Laboratory MASSACHUSETTS INSTITUTE OF TECHNOLOGY
LEXINGTON, MASSACHUSETTS
Prepared for the Federal Aviation Administration, Washington, D.C. 20591
This document is available to the public through
the National Technical Information Service, Springfield, VA 22161
This document is disseminated under the sponsorship of the Department of Transportation in the interest of information exchange. The United States Government assumes no liability for its contents or use thereof.
1. Report No.
FAA·RD·83/26
2. Government Acce..ion No.
Technical Report Documentation Page3. Recipient's Cetalog No.
,
•
4. Tide end Subtitle
An Experimental GPS Navigation Receiver for General Aviation:Design and Measured Performance
7. Author(s)
Steven D. Campbell and Raymond R. LaFrey
9. Performing Or.eniz.tion Nlme Ind Addrl..
Massachusetts Institule of TechnologyLincoln Laboratory, M.LT.P.O. Box 73Lexington, MA 02173-0073
12. Sponsoring Agency Nlml Ind Addre..
Department of TransportationFederal Aviation AdministrationSystems Research and Development ServiceWashington, DC 20591
15. Suppllmentery Nolls
5. Report Dete
27 September 1983
6. Performing Or.enizllion Code
8. Plrformin. Or.enizetion Rlport No.
ATC-121
10. Work Unit No.
11. Contrlct or Grlnt No.
DOT-FA79-WAI-091
Project Report
14. Sponsoring Agency Code
,
•
The work reported in this document was performed at Lincoln Laboratory, a center for research operatedby Massachusetts Institute of Technology, under Air Force Contract F19628-80·C-0002.
16. Abstrlct
This report describes work performed by M.I.T. Lincoln Laboratory, between 1 October 1979 and 1 March1983, to evaluate the use of the Global Positioning System (GPS) for low-cost civil air navigation.
The report describes a GPS Test and Evaluation System developed jointly by M.LT. Lincoln Laboratory,Stanford Telecommunications, Inc., and Intermetrics, Inc., using techniques that could lead to low-cost commercial avionics. System performance results obtained in the laboratory and during flight tests are providedwhich demonstrate compliance with current and future navigation accuracy requirements for enroute, terminal and non-precision approach flight paths. The report also includes functional specifications for a low-costGPS navigation system for civil aircraft.
The GPS Test and Evaluation system design was based on two important features: 1) automatic tracking ofall visible satellites (rather than a minimum set of four) and 2) a dual-channel GPS C/A code receiver. Tracking all visible satellites allows the system to maintain continuous navigation when a satellite sets or is momen·tarily masked during aircraft maneuvers. The dual·channel receiver dedicates one channel to pseudo-rangemeasurements, and the other channel to acquiring new satellites as they become visible. These two features,validated by flight test, allow the system to provide continuous navigation updates during critical aircraftmaneuvers, such as non-precision approaches, and during satellite constellation changes.
17. KlY Words
NAVSTARGPSglobal positioning systemair navigationcivil aviation
general aviationC/A code receiverdual-channel receiver
18. Diltribution SlItlment
Document is available to the public throughthe National Technical Information Service,Springfield, Virginia 22151
19. SlCurity ClllSif. (of this rlport)
Unclassified
Form DOT F 1700.7 (8-72)
20. SlCurity ClllSif. (of this PIli)
Unclassified
Reproduction of completed page authorized
21. Na. af PillS
254
22. Price
•
•
CONTENTS
1.0 INTRODUCTION AND SUMMARY
2.0 PROJECT OVERVIEW2.1 Background and Objectives2.2 Current Civil Navigation Requirements
2.2.1 Accuracy2.2.2 Reliability2.2.3 Integrity2.2.4 Compatibility
2.3 Future Civil Navigation Requirements2.4 Assumptions2.5 General Approach
2.5.1 Development2.5.2 Evaluation
3.0 GPS TEST AND EVALUATION EQUIPMENT DESIGN3.1 System Design
3.1.1 System Timing3.1.2 Alternative Startup Modes
3.2 Antenna3.2.1 General Considerations3.2.2 Multipath Delay Greater Than 1.5 Microseconds3.2.3 Multipath Delay Less Than 1.5 Microseconds
3.3 Preamplifier3.4 Dual Channel Receiver Hardware
3.4.1 Architecture and Design Features3.4.2 Receiver States
3.5 Position Software3.5.1 Overview
3.5.1.1 System Mode Control3.5.1.2 System Time Management3.5.1.3 Satellite Selection3.5.1.4 Receiver Management3.5.1.5 GPS Navigation Data Management3.5.1.6 Measurement Processing3.5.1.7 Navigation Tracker3.5.1.8 DAU Interface
3.5.2 Acquisition Strategy3.5.3 Transition Strategy3.5.4 Navigation Strategy3.5.5 Position Estimation3.5.6 Performance Monitoring3.5.7 Fail-Soft Techniques
3.6 Navigation Software3.7 Control and Display Unit3.8 Instrumentation
iii
1
33335555999
10
12121215151517172629292944444444444444494949495050515155556368
CONTENTS (CONT'D)
4.0 MEASURED PERFORMANCE4.1 Test Facilities
4.1.1 Ground Test Laboratory4.1.2 Aircraft Installation4.1.3 Mode S Experimental Facility4.1.4 Analysis Software
4.2 Laboratory Tests4.2.1 Antenna4.2.2 Receiver Channel Performance4.2.3 Static Acquisition Performance4.2.4 Static Position Accuracy and St.ability4.2.5 Interference Effects4.2.6 Performance Monitoring
4.3 Flight Tests4.3.1 Engineering Flight Tests
4.3.1.1 Engineering Tests With Preliminary PositionSoftware
4.3.1.2 Engineering Tests With Final PositionSoftware
4.3.1.3 Summary of Engineering Test Results4.3.2 Operational Flight Tests
4.3.2.1 Burlington. VT Tests4.3.2.2 Logan International Tests4.3.2.3 Hanscom Field/Manchestt~r Airport Tests4.3.2.4 Summary of Operational Flight Test Results
5.0 FUNCTIONAL REQUIREMENTS OF A GENERAL AVIATION RECEIVER5.1 General Requirements
5.1.1 Reliability5.1.2 Integrity5.1.3 Compatibility
5.2 Technical Requirements5.2.1 Link Margin5.2.2 Startup5.2.3 Time-to-First-Fix5.2.4 Continuous Navigation Solutions5.2.5 Position Estimation5.2.6 Performance Monitor
REFERENCES
APPENDIX A - POSITION ESTIMATION ALGORITHM
APPENDIX B - POSITION ESTIMATION FILTER
APPENDIX C - COORDINATE CONVERSION ALGORITHM
iv
8484848484949494
102118118122122122122
124
135148148149172185228
230230230230230230230230231231231231
232
A-1
B-1
C-1
•
..
..
•
•
•
•
3-13-23-33-43-53-6
3-7
3-8
3-93-103-113-123-133-143-153-163-173-183-193-203-213-223-233-243-253-263-273-283-293-30
3-313-323-333-343-353-363-37
4-14-24-34-4
ILLUSTRATIONS
GPS Test and Evaluation SystemSystem TimingAltitude vs. Elevation Angle for 1.5 Chip Multipath DelayEffect of Multipath Delay Greater than 1.5 ChipsEffect of Multipath Delay Less than 1.5 ChipsExpected Noncoherent DLL Tracking Error, 20 dB Carrier
Tracking Margin (Ref. 8)10 Noncoherent DLL Tracking Error, 20 dB Carrier Tracking
Margin (Ref. 8)Typical Vertically Polarized Signal Reflection
Characteristics (Ref. 2)Initial Antenna Gain vs. Elevation Angle RequirementRevised GPS Antenna CharacteristicGPS Antenna, Outline DimensionsExperimental Dual Channel ReceiverFunctional Schematic of ReceiverAFC and CINo EstimatorReceiver Acquisition State DiagramReceiver Transition State DiagramReceiver Navigation State DiagramNavigation State 220 MS Interval Modes and Time LinesSatellite Addition State DiagramDual Channel ReceiverDual Channel Receiver - Down ConverterDual Channel Receiver - Top ViewSystem SoftwarePosition Software System Block DiagramReceiver ModesLoss of Lock AlgorithmPosition Solution and Measurement ProcessingPosition Software Performance MonitorPosition Software Performance MonitorTiming Diagram for a 2.4 Second Cycle Single Channel
SystemNavigation Software, Functional Block DiagramRNAV GeometryCourse Deviation Indicator/omni-Bearing SelectorControl and Display Unit (CDU)DAU-eDU InterfaceSample Flight Plan: Atlantic City to Hanscom FieldGPS CDU Front Panel
Aerocommander Test AircraftGPS Antenna InstallationGPS Test Aircraft Internal LayoutGPS Flight System in Passenger Compartment ofAerocommander Aircraft
v
1314181920
22
23
24252728303134363738394041424345474852535657
5860616264656671
858687
88
4-54-64-74-84-94-104-11
4-12
4-13
4-144-154-164-174-184-194-204-214-224-234-244-254-26
4-27
4-284-294-304-314-32a4-32b4-33a4-33b4-344-354-36a4-36b4-374-384-394-404-414-42a4-42b4-43
ILLUSTRATIONS (CONTIfD)
Instrumentation Rack 89Receiver Rack 90Aerocommander N333BE, Cockpit View 91Post Flight Analysis Software 95GPS Position Estimation Error Algoritrlm 96Curved Groundplane 97Gain vs. Azimuth for Elevation Cuts at: -5° and 0°; Chu Model
CA-3224 Volute Antenna 98Gain vs. Azimuth for Elevation Cuts at: +60 and +17°; Chu
Model CA-3224 Volute Antenna 99Gain vs. Azimuth for Elevation Cuts at: +30° and +50°; Chu
Model CA-3224 Volute Antenna 100Volute Antenna Gain, Elevation Profile! 101Dual Channel Receiver Acceptance Test Configuration 103Acquisition Times of a Single Satellite 105Receiver AFC Performance in Acquisition Phase 106Receiver C/No Estimator Performance in Acquisition Phase 107Receiver Pseudo-Range Performance in Phased Transition 108Receiver AFC Performance in Phased Transition 109Receiver C/No Estimator Performance in Phased Transition 110Receiver Pseudo-Range Performance in the Navigation State 111Receiver AFC Performance in the Navigation State 112Receiver C/No Estimator Performance in Navigation State 113Receiver Performance During Satellite Pass 114Comparison of Theoretical and Measured Receiver Frequency
Tracking Performance 115Comparison of Theoretical and Measured Receiver Code Loop
Performance 116GPS Static Test, 19 Nov 82 121Test Flight Run la, 28 Oct 82 125Satellite Visibilities for Run la, 28 Oct 82 126Dilution-of-Precision, 28 Oct 82 Flight 127Position Fix Performance, 30° Bank Turn 129Position Estimate Performance, 30° Bank Turn 130Position Fix Error During 30° Turns 132Position Estimate Error During 30° Turns 133Satellite C/No During 30° Turns 134Flight Test, 21 Dec 82 136Position Fix Errors, 21 Dec 82 Flight Test 138Position Estimate Errors, 21 Dec 82 FHght Test 139Pseudo-Range Residuals, 21 Dec 82 Flight Test 141Five 360° Turn Series, 25 Jan 83 142Pseudo-Range Residuals for Five 360° ~Jrns 143Ground Track for Multipath Test, 25 Jan 83 144Altitude Profile for Multipath Test, 25 Jan 83 145Satellite Visibilities for Multipath T4:!st, 25 Jan 83 146Pseudo-Range Residuals for Multipath ~:!st, 25 Jan 82 147Burlington, VT Approaches: NDB RWY 15 i!nd VOR RWY 1 150
vi
•
•
•
•
•
•
4-444-454-464-474-484-494-504-514-524-534-544-554-564-574-584-594-604-614-624-634-644-654-664-674-684-694-704-714-724-734-744-754-764-774-784-79
4-804-814-824-834-844-854-86a4-86b
ILLUSTRATIONS (CONT'D)
Burlington, VT Operational Test, Run 2a, 27 Jan 83Altitude and Ground Speed, Run 2a, 27 Jan 83GPS SV PositionsSatellite C/No• Values, Run 2a, 27 Jan 83Pseudo Range Residuals, Run 2a, 27 Jan 83Burlington, VT VOR Approach, 27 Jan 83Altitude and Ground Speed, Run 2b, 27 Jan 83Satellite Visibilities, Run 2b, 27 Jan 83Satellite C/No• Values, Run 2b, 27 Jan 83Pseudo-Range Residuals, Run 2b, 27 Jan 83Burlington, VT VOR Approach, Run 3b, 27 Jan 83Altitude and Ground Speed, Run 3b, 27 Jan 83Satellite Visibilities, Run 3b, 27 Jan 83Satellite C/No • Values, Run 3b, 27 Jan 83Pseudo-Range Residuals, Run 3b, 27 Jan 83Burlington, VT NOB Approach, Run 4, 27 Jan 83Altitude and Ground Speed, Run 4, 27 Jan 83Satellite Visibility, Run 4, 27 Jan 83Satellite C/No • Values, Run 4, 27 Jan 83Pseudo-Range Residuals, Run 4, 27 Jan 83Logan International Test, 25 Jan 83, Run 2Altitude and Ground Speed, 27 Jan 83, Run 2Satellite Visibility, 27 Jan 82, Run 2Dilution-of-Precision, 25 Jan 83, Run 2Satellite C/No.' 25 Jan 83, Run 2Pseudo-Range Residuals, 27 Jan 83, Run 2Logan International, 28 Jan 83Altitude and Ground Speed, 28 Jan 83, Run 2aSatellite C/No.' 28 Jan 83, Run 2aPseudo-Range Residuals, 28 Jan 83, Run 2aManchester Airport, 25 Jan 83, Run 3Altitude and Ground Speed, 25 Jan 83, Run 3Satellite Visibilities, 27 Jan 83, Run 3Satellite C/No.' 25 Jan 83, Run 3Pseudo-Range Residuals, 27 Jan 83, Run 3Test Scenario for Hanscom Field/Manchester Airport
Operational FlightsOperational Test, 4 Feb 83, Run 1aAltitude and Ground Speed, 4 Feb 83, Run 1aSatellite Visibilities, 4 Feb 83, Run 1aDilution of Precision, 4 Feb 83, Run 1aSatellite C/No.' 4 Feb 83, Run 1aPseudo-Range Residuals, 4 Feb 83, Run 1aPosition Fix Error, 4 Feb 83, Run 1aPosition Estimate Error, 4 Feb 83, Run 1a
vii
151152153154155157158159160161162163164165167168169170171173174175176177179180181182183184186187188189190
191193194196197198199200201
4-874-884-894-904-91
4-924-934-944-954-964-974-984-994-994-1004-1014-1024-1034-104
ILLUSTRATIONS (CONT'D)
Holding Pattern, 4 Feb 83, Run la: Position Fixes 203Holding Pattern, 4 Feb 83, Run la: Tracker Estimates 204Position Fix Errors for Holding Pattern, 4 Feb 83, Run la 205Tracker Estimate Error for Holding Pattern, 4 Feb 83, Run la 206Altitude and Ground Speed for Holding Pattern, 4 Feb 83,
Run la 208Operational Flight Test, 9 Feb 83, Run la 210Altitude and Ground Speed, 9 Feb 83, Run la 212Satellite Visibilties, 9 Feb 83, Run la 213Dilution of Precision, 9 Feb 83, Run la 214Position Fix Error, 9 Feb 83, Run la 215Tracker Estimate Error, 9 Feb 83, Run la 216Operational Test, 9 Feb 83, Run Ib 219Altitude and Ground Speed, 9 Feb 83, Run lb (1 of 2) 220Altitude and Ground Speed, 9 Feb 83, Run Ib (2 of 2) 221Satellite Visibilites, 9 Feb 83, Run Ib 222Dilution-of-Precision Values, 9 Feb 83, Run Ib 223Satellite C/No. Values, 9 Feb 83, Run Ib 224Position Fix Error, 9 Feb 83, Run Ib 225Tracker Estimate Error, 9 Feb 83, Run Ib 226
•
•
A-IA-2
B-1
Position Fix AlgorithmGPS Position Fixing Geometry
Position Estimation Trackervi
viii
A-2A-4
B-2
•
•
•
•
2-12-2
2-32-4
3-13-23-33-43-53-6
3-73-83-93-103-113-123-133-143-153-163-173-183-193-203-213-22
4-14-24-34-44-54-64-74-84-94-104-114-124-13
TABLES
AC90-45A 2D Navigation RequirementsGPS Test and Evaluation Equipment Accuracy and Update
RequirementsGPS Test and Evaluation Equipment Performance GoalsFRP Navigation Accuracy to Meet Projected Future
Requirements
Alternative Start-Up ModesInitial Antenna Gain vs. Elevation Angle RequirementRevised GPS Antenna SpecificationReceiver Channel - Position Software Functional AllocationReceiver Loop CharacteristicsReceiver AFC Lock Detector and CINo Estimator
CharacteristicsSoftware Functional AreasPerformance Monitor ParametersFail Soft TechniquesStored Waypoint Data File: Atlantic City to Hanscom FieldActive Waypoint Data FileDisplay Field by Key Entry (aPR Mode Only)Data Recording MessagesBlock #10 Receiver Channel Configuration DataBlock #30 Raw GPS Nav Message Data WordBlock #40 Acquisition Command DataBlock #70 Measurment DataBlock #150 Raw Position Fix DataBlock #160 Tracked Position Fix DataActive Waypoint Data Block #200 (From CDU)Area Navigation Data Block #220 (To CDU)Real-Time Aircraft State Data
Aerocommander Flight CharacteristicsMode S Experimental Facility Surveillance CharacteristicsReceiver Hardware PerformanceComparison of Theoretical and Measured Receiver PerformanceUser Received Power for CIA Code OnlyExpected Link MarginStatic Acquisition ExampleAirports for GPS Flight TestsPosition Accuracy Statistics for Run la, 28 October 1982Position Accuracy Statistics for 30° Bank Angle TurnsPosition Accuracy Statistics, 21 December 1982 TestPosition Accuracy Statistics, 4 Feb 1983, Run lAPosition Accuracy Statistics for Holding Pattern,
4 February 1983, Run lA
ix
4
67
8
1621263233
3346545967697072737374767879808182
9293
104117117119120123128131137202
209
TABLES (CONT'D)
4-14 Position Accuracy Statistics for 9 February 1983 Run lA(43085 - 45770) 217
4-15 Position Accuracy Statistics for 9 February 1983 Run IB(46030 - 48580) 227
4-16 Position Accuracy Statistics for Operational Flight Tests4-10 February 1983 229
A-I Signal Propagation Delay Compensation Models A-S
x
,
•
•
1.0 INTRODUCTION AND SUMMARY
This report describes the work performed by M.I.T. Lincoln Laboratory,between 1 October 1979 and 1 March 1983 to evaluate the use of the GlobalPositioning System (GPS) for low-cost civil air navigation. The effort wassupported by the Federal Aviation Administration through Interagency AgreementDOT-FA79WAI-091 between the FAA and the United States Air Force.
M.I.T. Lincoln Laboratory was tasked by the FAA to develop a GPS Test andEvaluation System and to flight test the system in a typical general aviationaircraft. The system was required to meet FAA accuracy requirements for areanavigation systems and to use techniques that could lead to low-costcommercial avionics. This report describes the system design, and providessystem performance results from laboratory and flight tests. The report alsoincludes functional specifications for a low-cost GPS navigation system basedon validated design concepts.
The GPS Test and Evaluation system design was based on two importantfeatures: 1) automatic tracking of all visible satellites (rather than aminimum set of four) and 2) a dual-channel GPS CiA code receiver. Trackingall visible satellites allows the system to maintain continuous navigationwhen a satellite sets or is momentarily masked during aircraft maneuvers. Thedual-channel receiver dedicates one channel to pseudo-range measurements, andthe other channel to acquiring new satellites as they become visible. Thesetwo features allow the system to provide continuous navigation updates duringaircraft maneuvers such as non-precision approaches and during satelliteconstellation changes.
The major elements of the GPS Test and Evaluation System are:
• Dual Channel Receiver - a unit developed by StanfordTelecommunications, Inc. that sequentially tracks satellites on onechannel while the other channel acquires new satellites anddemodulates data; provides pseudo-range estimates and satellite datato the Position Processor.
• Position Processor - a microcomputer with software developed byIntermetrics, Inc. that manages the receiver channels and computesgeodetic position estimates.
• Navigation Processor - a microcomputer that translates geodeticestimates into displays suitable for pilot navigation.
• Pilot Displays - a standard Course Deviation Indicator/Omni-BearingSelector (CDI/OBS) and a Control and Display Unit (CnU), designed toprovide navigation data to the pilot in a format consistent withcurrent civil navigation practices.
The system also includes extensive instrumentation and data recordingcapabilities for post-flight data analysis.
1
The performance of the GPS Test and Evaluation System was measured duringlaboratory tests and during flight tests in a typical twin-engine generalaviation aircraft, a Rockwell Aero Commander. The laboratory test showed thatthe typical static RMS horizontal position error was 93 feet. The flighttests showed a RMS horizontal error of 189 feet and a 95% confidencehorizontal error of 333 feet in typical general aviation operations. As shownbelow, the measured accuracy was well within the requirements of FAA AdvisoryCircular 90-45A for two-dimensional area navigation and essentially met theproposed 328 ft., 95% Federal Navigation Plan requirement for non-precisionapproach.
Navigation Accuracy Summary
Horizontal (95%)(ft)
•
FAA Advisory Circular 90-45AApproach, ascent/descent
Federal Radionavigation PlanNon-precision approach
Accuracy AchievedLevel and turning mix
1824.
328.
333.
Operation of the GPS Test and Evaluation System was evaluated in areas ofmountainous terrain, at a large urban airport and in typical general aviationoperations. The system was shown to provide continuous navigation serviceduring 30° bank-angle turns and was able to track satellites with elevationangles as low as S°. Finally, the system appeared to be compatible withexisting air-traffic control procedures and air-crew practices.
Although the experimental GPS was found to perform well as a practicalgeneral aviation navigation system, it was physically large. This wasnecessary in order that the system be construct'ed from readily availablecomponents ana that it provide for experimental flexibility. However, theincreasing use of VLSI digital techniques is likely to result in a systemdesign which can be produced in an avionics package comparable in size to thatof present day commercially available area navigation equipment. The cost, asestimated by ARINC Research, Inc., of a GPS navigator based on the Lincolndesign was $8500 in 1982 dollars assuming the use of circa-1990integrated-circuit technology.
The remainder of this report is divided into four sections. Section 2.0provides an overview of the project and details the performance requirementsand goals. Section 3.0 describes the system de:~ign, including the hardwarecharacteristics and software architecture. Section 4.0 provides the systemperformance results, including ground and flight test measurements.Section 5.0 gives functional requirements of a general aviation GPS navigationsystem as determined from the results of this project.
2
•
2.0 PROJECT OVERVIEW
2.1 Background and Objectives
The NAVSTAR Global Positioning System (GPS) is a satellite-based systemcurrently being developed by the Department of Defense to provide positioninformation to suitably equipped military users. Such users will receiveworld-wide t continuous t real-timet all weather t precision navigation data froma constellation of 18 satellites in 12 hour orbits (Ref. 1). Six satelliteshave been placed in orbit and ground support facilities have been developed tosupport user equipment development.
Previous studies (Refs. 2, 3) suggested that a GPS navigator meeting FAAperformance requirements could be developed at a cost consistent with generalaviation use. However t the performance of low cost GPS receiver designs hadnot been fully determined by field measurement. To fill this gap, M.I.T.Lincoln Laboratory was tasked by the Federal Aviation Adminstration toa) develop GPS Test and Evaluation Equipment functionally consistent with alow cost design and FAA two-dimensional (2-D) navigation requirements tb) evaluate the design by field measurement t and c) develop a functionalspecification for a general aviation GPS receiver. Lincoln Laboratorycontracted with Stanford Telecommunications t Inc. t to develop a dual channelreceiver, and Intermetrics, Inc., to develop receiver management and positionestimation software. To establish cost t the FAA tasked ARINC Research t Inc. tto estimate the cost of a commercial version of the design developed atLincoln Laboratory; the results of that cost study have been reported by ARINCResearch in a separate document (Ref. 4).
2.2 Current Civil Navigation Requirements
In order that a navigation system be acceptable for civil aviation itmust satisfy needs within four major areas of concern: accuracy, reliabilityintegrity, and compatibility.
2.2.1 Accuracy
The basic output of the typical GPS navigation system is a geodetic(latitude-longitude) position estimate. This is to be contrasted with lowcost VOR/DME navigation receivers which produce rho-theta (range-bearing)estimates with respect to a ground facility. The relevant specifications fromFAA Advisory Circular 90-45A t (Ref. 5) for GPS are shown summarized inTable 2-1.
Position accuracy should not be considered apart from update rate.Equipment intended for use in non-precision approaches should providenavigation information that is essentially continuous with interruptions nolonger than would result from switching from one pre-programmed waypoint toanother. Also t navigation service should be continuously available duringflight in any direction/climb/descend profile approved by the aircraftmanufacturer. The navigation service shall t per AC90-45A, be restored (iftemporarily lost due to bank-induced fades) within 5 seconds after thecompletion of any allowable maneuver. Also, the time lag between datameasurements and displayed position must not be operationally significant.
3
TABLE 2-1
AC90-45A 2-D NAVIGATION REQUIREMENTS
Position Accuracy Requirements - 95% Conf.(Averaged Over One Update Cycle)
Cross-Track I Along-Track
IFlight Technical Error 1 (FTE)
Enroute2 2.0 runi --Termina13 1.0 nJui I --Non-Precision Approach4 0.5 nJni --
Navigation Equipment Error
Enroute 1.5 mlli 1.5 runi
Terminal 1.1 mlli 1.1 nmi
INon-Precision Approach 0.3 nDli 0.3 runi
Total ErrorS
Enroute 2.5 nnl1 1.5 runiI
ITerminal 1.5 nOli 1.1 nmi
I INon-Precision Approach 0.6 nmi 0.3 nmi
1. FTE accounts for deviations due to display interpretation, pilotresponse and aircraft response.
2. Enroute = Cruising flight between terminal areas.
3. Terminal = Between enroute and approach; normally below 18000 feet andwithin 50 miles of the airport.
4. Non-Precision Approach • Between final approach waypoint and airport.
5. Total Error • RSS combination of FTE and navigation equipment errors.
4
•
As part of the test and evaluation equipment design development, theaccuracy and update specifications in Table 2-2 were established. The dynamicrequirement is the Non-Precision Approach approach tolerance of AC90-45A. Thestatic requirement is derived from a generic low-cost receiver model. Anadditional requirement, Time-to-First-Fix, was established based on thetypical time between first turn-on of the avionics and when the generalaviation aircraft is positioned near the departure runway.
2.2.2 Reliability
The GPS navigation system encompassing the user equipment, satellitevehicles and ground support equipment, must have a combined reliability equalto or surpassing that of alternative navigation systems. The GPS userequipment must therefore provide navigation data without operationallysignificant outages due to fades or multipath, assuming that the NAVSTARconstellation provides acceptable coverage.
2.2.3 Integrity
The GPS navigation system must not provide misleading information underany conditions of operational significance. It is therefore necessary thatthe GPS receiver continually monitor its own performance and indicate to thepilot when the navigation information displayed is no longer in compliancewith the accuracy requirements.
Further, provisions must be incorporated to allow the pilot to verifythat the navigation information is accurate using either built-in testequipment, an auxiliary test system, or a procedural check.
2.2.4 Compatibility
The GPS navigation system must provide a pilot interface which iscompatible with existing air navigation systems. This requirement stems fromthe way that pilots are accustomed to using aircraft navigation systems.
2.3 Future Civil Navigation Requirements
To anticipate the more stringent navigation requirements likely in thefuture, the accuracy and update rate goals in Table 2-3 were established forthe GPS Test and Evaluation (T and E) equipment.
Later, after the CPS T and E Equipment specifications were frozen inprocurement specifications, the Federal Radio Navigation Plan (FRP) (Ref. 6)was completed and published. The FRP established new future navigationrequirements that are summarized in Table 2-4.
The rationale for increasing the accuracy is to provide a serviceequivalent to that provided by on-airport VOR's which now exist atapproximately 30% of all U.S. airfields.
5
TABLE 2-2
GPS TEST AND EVALUATION EQUIPMENT ACCURACY AND UPDATE REQUIREMENTS
Horizontal Dynamic Position Error1
Horizontal Static Position Error2
Update Rate
\Time to First Fix3
Assumptions:
Velocity < 200 KTS
Acceleration < 0.5g
Ll CiA only
0.3 NM, 95%
335 ft., 95%
0.5 Hz, min
6 minutes, max
•
No artifical degradation of the GPS navigation message dataor intentional electromagnetic interf€~rence
1. For HDOP* < 5, bank angle> 30°.
2. For HDOP < 2.0 and a 2 minute integration.
3. From power-on until cockpit display of position fix.
*Horizontal Dilution of Precision.
6
TABLE 2-3
GPS TEST AND EVALUATION EQUIPMENT PERFORMANCE GOALS
Horizontal Position Error1
• Level Flight
Turni ng Flight
Vertical Position Error2
Enroute level flight
Terminal level flight
Approach level flight
Enroute Ascend/Descent
Terminal Ascend/Descent
Approach Ascend/Descent
Update Rate
500 ft., 95%
1000 ft., 95%
163 ft., 95%
163 ft., 95%
88 ft., 95%
275 ft., 95%
205 ft., 95%
116 ft., 95%
1 Hz, min
1. For HDOP < 5, bank angle < 30°, ground speed < 200 KTS.
2. From AC90-45A (Total error less FTE, Table B, Appendix A); sameconditions as Horizontal Goals.
7
TABLE 2-4
FRP NAVIGATION ACCURACY TO MEET PROJECTED FUTURE REQUIREMENTS
Phase
Enroute
Terminal
Non-PrecisionApproach
Altitude Horizontal Position Error
500 ft. to FL 180* 1000 meters (2 drms)(3280 ft.)
500 ft. to FL 180 500 meters (95%)(1650 ft.)
250 to 3000 ft. above surface 100 meters (95%)(328 ft.)
I *FL 180 means flight level 18,000 feet.
8
•
2.4 Assumptions
The following assumptions were made for th~ development and testing ofthe GPS T and E Equipment:
a. The satellites will radiate the L1 signal such that the receivedsignal at the civil users GPS antenna input will meet or exceed thelevels specified in the space segment User Interface lCD, (Ref. 7).
b. The radiated signals from the GPS satellites will not be artificallydegraded to reduce accuracy.
c. The T and E Equipment will not be exposed to intentional radiofrequency interference •
2.5 General Approach
2.5.1 Development
The development of the GPS T and E Equipment was, in part, based on pastwork (Ref. 7) and on the DOD GPS user equipment developments that suggested:
a. Designs which tracked all satellites in view (a significantadvantage) could be realized at reasonable cost,
b. Designs which convert functions from hardware to firmware usingmicrocomputers would reduce cost, and
c. An L1 CiA-only receiver would be able to provide acceptable accuracyfor civil general aviation users.
The need for continuous navigation service, especially duringnon-precision approaches, led to the requirement that the T and E Equipmentcontain two independent channels. The need to evaluate performance under avariety of stressing situations resulted in the requirement for extensiveinstrumentation. The requirement to evaluate operational issues lead to thedefinition of a general purpose cockpit display consistent with current civilnavigation procedures.
It was assumed that during the initial general aviation use of GPS, thepilot interface would be consistent with the VOR/DME system, with RNAVfeatures as added options. This implied that the following features would beavailable on initial GPS receivers and therefore should be part of the T and EEquipment:
a. Automatic Operation. The receiver should automatically acquire andtrack satellites, estimate position and continually monitorperformance without operator intervention.
b. Course Deviation Indicator (CDI) Interface. The receiver shouldaccept desired course bearing from an Omni Bearing Selector (OBS) andprovide signals to drive the Course-Deviation Indicator (CDI).
9
c. Magnetic Bearing. A map should be provided to apply local magneticvariation corrections (to within ± 1 degree); the map to be stored ina non-volatile memory.
d. Way Points. The system should accept w.ay points in the form ofunambiguous designators such as the current letter codes. Thereceiver could then maintain several thousand waypoints withtheir associated geographical coordinat1es in a non-volatile memory.
The concern for integrity led to the inclusion of performance monitoringand fail-soft techniques. Tests such as satellite health checks, pseudo-rangeconsistency tests, position estimate variance tests and geometry tests wereprovided.
Finally, low cost was emphasized by designing to requirements that areconsistent with civil aviation navigation and by using technology which can beexpected to become inexpensive in the current de'cade.
The design and construction of the equipment was accomplished during1980-82. As mentioned previously, Standard Tele.communications, Inc.,developed the dual channel receiver and Intermetrics, Inc., developed softwarewhich managed the receivers and estimated position. Lincoln Laboratorydeveloped the navigation software, instrumentation, Control and Display Unit(CDU), and ground support facilities, conducted ground and flightmeasurements, and evaluated system performance.
2.5.2 Evaluation
In keeping with the principal concerns, field evaluation of theGPS T and E equipment focused on a) link margins and position estimationtechniques, b) performance monitoring and fail-soft techniques, andc) operational performance in typical general av:lation flight operations withemphasis on non-precision approaches.
Link margins were evaluated by determining the effects of:
a) the number and location of available satellites,b) received signal strength,c) multipath and terrain blockage,d) electromagnetic interference (EMI),e) antenna shielding, andf) receiver acquisition and tracking
on the acquisition time, positional accuracy and update reliability. Accuracywas measured using truth derived from a ground tracker while flying theT and E Equipment in a general aviation aircraft.. The position estimationtechnique, including its smoothing filter, was e'7aluated to assess the effectof signal margin and aircraft dynamics on the ponition estimate variance.
10
•
•
Performance monitoring and fail-soft features were evaluated within thelimitations of the current GPS constellation.
Operational evaluation flights were conducted using a limited set ofcross-country, terminal and non-precision approach flight plans in the NewEngland area. The cockpit was configured to allow the test pilot tosimultaneously observe the GPS CDU and CDI, and a conventional VOR-drivenCDI •
11
3.0 GPS TEST AND EVALUATION EQUIPMENT DESIGN
3.1 System Design
The GPS T and E Equipment is shown in Fig. 3-1. This design provideshigh rate (5 Hz) sequential processing of all visible GPS satellites in orderto support a position estimate update rate of 1.0 Hz. Two channels areprovided in order for one channel to be dedicat~i to high rate sequentialpseudo-range processing while the other is avaiL:lble to demodulate satellitenavigation messages. This assures continuous navigation service duringnon-precision approaches.
The key features of the system design are:
a. Automatic Operation. The receiver automatically acquires allsatellites in view and provides a first fix to the pilot in less than6 minutes. If required (after being p01o1ered-off for more than 30days), it automatically acquires fresh i:llmanac data, which increasesthe time-to-first fix (TTFF) to 16 minutes. The receiver normallyrequires no operator intervention duri~~ use other than suchnavigation functions as waypoint entry.
b. Performance Monitoring. The receiver continually monitors satellitedata, pseudo-range estimates, receiver parameters, and positionsolution estimates in order to determinl~ compliance with FAAnavigation requirements.
c. Microprocessor Based Design. The receiver makes extensive use ofmicroprocessor technology to synthesize receiver loops, managereceiver operations and compute own post tion.
d. Intelligent Control and Display Unit. The control and displayunit is managed by a separate Z80 microprocessor in order toallow developmental flexibility in the pilot display interface designwithout affecting the receiver design.
e. Operational Compatibility. The design fncorporates modes ofoperation consistent with current VOR enroute terminal andnon-precision approach procedures. In ~mdition, it includes directrouting area navigation (RNAV) modes.
3.1.1 SystemTiming
The typical sequence of events following pOlorer-on is illustrated inFig. 3-2. During a brief initialization period, built-in diagnostics testmicroprocessor and receiver functions, and the age of the stored almanac arechecked. The posi tion sof tware uses the last stored position and almanac datato select four space vehicles (SV's) for initial acquisition. The first SV isthen assigned to both channels, one searching thE! 1023 chip code in a 1500 Hz
12
..
-I.F
..
RS
23
2C
ON
TRO
L
-,
CO
RR
ELA
TOR
Z8
00
0L
SI-
I1
/23
LS
I-I
1/2
3.~
AR
INC
AN
DD
ISP
LAY
~..
""..
,/
....4
29
..U
NIT
~za
o
LON
NA
VIG
ATI
ON
NO
ISE
SO
FTW
AR
E
PR
EA
MP
CO
UR
SE
DO
WN~
50
Hz
..D
EV
IATI
ON
CO
NV
ER
TER
SY
NTH
ES
IZE
R-
IND
ICA
TOR
PO
SIT
ION
SO
FTW
AR
ED
MA
A..
-O
MN
I-BE
AR
ING
y"
SE
LEC
TOR
,IID
ATA
RE
CO
RD
ING
-I.
F.
A..
.R
S2
32
SO
FTW
AR
E
CO
RR
ELA
TOR
zaoo
oI~
CA
LEN
DA
R~
...C
LOC
K
RS
23
2~
AIR
CR
AFT
DA
TAU
NIT
DU
AL
CH
AN
NE
Lza
oR
EC
EIV
ER
IIA
LTIM
ET
ER
TAP
E"
-..R
EC
OR
DE
R
Fig
.3
-LG
PS
TE
ST
AN
DE
VA
LU
AT
ION
SY
ST
EM
r---
----
----
AC
QU
ISIT
ION------------.
.I.....---T
RA
NS
ITIO
N---.+
-----
NA
VIG
AT
ION
2-3
MIN
UT
ES
<I
MIN
UT
EP
OW
ER
ON
..
/\
rntff
i[8
22
0..
.eP
ER
SA
TI
SE
C-I
SA
T4
SE
AR
CH
SA
TS
AT
SA
TS
AT
AN
DD
AT
AI
23
4S
AT
2S
EA
RC
HA
ND
DA
TA
SA
TI
SE
AR
CH
PA
RA
LL
EL
SE
AR
CH
FOR
FIR
ST
SA
T
SY
ST
EM
INIT
IALI
ZA
TIO
N
1\--
2.2
SE
C----
IP
OW
ER
ON
r-
SY
ST
EM
SA
TI
lSA
TI
SA
T3
SE
AR
CH
SA
T&
SE
AR
CH
SA
T6
SE
AR
CH
SA
T7
SE
AR
CH
INIT
IALI
ZA
TIO
NS
EA
RC
H1D
AT
AA
ND
DA
TA
AN
DD
AT
AA
ND
DA
TA
AN
DD
AT
A
RE
CE
IVE
RC
HA
NN
EL
-2(D
AT
A)
RE
CE
IVE
RC
HA
NN
EL
-I(N
AV
)
.1AC
TIV
AT
ES
TA
RT
PL
OT
DIS
PL
AY
.N
AV
MO
OE
AS
SIG
NS
AT
7
ST
AR
TA
SS
IGN
AS
SIG
NT
RA
NS
ITIO
NA
SS
IGN
SA
T4
SA
T&
MO
OE
SA
T6
AS
SIG
NA
SS
IGN
SA
T2
SA
T3
ti
ii
ii
CO
MP
UTE
FIX
EV
ER
Y4
SE
CC
OM
PU
TEF
IXE
VE
RY
2.2
SE
CS
EL
EC
TC
ON
TR
OL
AC
QU
ISIT
ION
SA
T'S
CO
NTI
NU
ES
AT
AC
QU
ISIT
ION
SO
NC
HA
NN
EL
-2
AS
SIG
NP
OW
ER
ON
SA
TI
11-0
------
------
----T
TF
F<
6M
IN
PO
SIT
ION
SO
FT
WA
RE
(LS
I1
1/2
3)
FIG
.3
-2.
SY
ST
EM
TIM
ING
bin centered on the expected frequency, and the second searching in anadjacent 1500 Hz bin. As shown in the example of Fig. 3-2, Channel 2 acquiredSV-l first and proceeded to demodulate the 1500 bit navigation message.Concurrently, the receiver management software directed Channel 1 to searchfor SV-2.
When four SV's are acquired, Channel 1 is then placed in the transitionmode. On entry to the transition mode, the satellites have ± 17 chip codephase and ± 750 Hz frequency uncertainties. After several four-secondtransition cycles, the code and frequency uncertainties narrow to ± 0.16 chipsand ± 200 Hz. Channel 1 is then assigned to the navigation mode where SVs 1-4are sequentially accessed at 220 msec per SV. The navigation mode operates ona ten slot cycle lasting 2.2 seconds. As additional SVs are acquired byChannel 2, they are immediately transferred to a slot in the Channel 1navigation mode cycle.
When six SVs have been acquired or six minutes has passed, a smoothedposition estimate, based on a batch processed least-squares linearizationalgorithm, is transferred to the pilot display. It is important to attempt tohave acquired at least six SVs prior to activation of the pilot display inorder to have two backup SVs during aircraft takeoff maneuvers. Typically sixor more SVs will have been acquired within six minutes.
3.1.2 Alternative Startup Modes
The startup sequence discussed in the previous section assumed that thecurrent time (+ 1 minute) and location (+ 3 miles) are available (i.e., wererecorded prior-to the previous power shutdown) and that a stored almanacexists whose age is less than 30 days. Should own position be unknown, or thebattery supporting the built-in calendar clock be low, one of the alternativestartup modes shown in Table 3-1 will be auto~atically selected.
Should the almanac be old (age> 30 days), an additional 12.5 minutes isnecessary to acquire a new almanac from any of the available SVs.
3.2 Antenna
An importantdefinition of theaircraft antenna.summarized below.
aspect of the low-cost GPS receiver development is therequired gain versus elevation angle characteristic for theResults of a study to determine this characteristic are
3.2.1 General Considerations
It is convenient to consider the above-horizon and below-horizon portionof the GPS antenna gain characteristic separately. The above horizoncharacteristic depends on the satellite slant ranges at low and high elevationangles. Since the free space path loss is 2 dB greater at 5° elevation thanat zenith, the antenna gain should be 2 dB greater at low elevations than athigh elevations. The boundary between these two regimes is taken as 30°,where the free space path loss is 1 dB greater than at 5°.
15
TABLE 3-1
ALTERNATIVE START-UP MODES
MODE CONDITIONS ASSUMPTIONS CHARACTERISTICS
30 ~eg banksuncertainty(3x10-6 )
I • TTFF - 3-4 mins
t -Predicted SV visibility• Narrow frequency
Isearches
• Full code 1st SVI search - limi ted sub-
• Know L.O. fr~quency [ sequentlyto ± 5 x 10- -"- • TTFF - 3_-_4_mi_n_s _
• Know position to ± 3miles
• Know time to ± 1 min
ICoid • Full range and L.O. • Prestored almanac • Trial-and-error SVStart frequency uncertainty (may be old) vi sibili ty determi-
• User position • Stationary user nationuncertainty • Wide frequency
searches• Full code searches• Constellation
I Ireview and 2ndacquisition pass
I I following first fix
• TTFF - 10-36 mins I
I"Time • Know time to ± 1 min • Prestored almanac • Predicted SV visi-only" bilitystarts • Know coarse position • Stationary user • Wide 1st SV frequency
(± 6.75°) search - narrowedsubsequently
• Full L.O. frequency I • Full code searchesuncertainty • TTFF - 4-7.5 mins
Time-and • Know time to ± 1 min • Prestored almanac • Predicted SV visibilityposition • Wide 1st SV frequencywideband • Know posi tion to ± 1° • User motion: search - narrowedstart 200 knot veloci ty subsequently
I • Full L.O. freq 0.5 acceleration • Full code searches
16
The below-horizon characteristic depends upon the effect of multipath onthe GPS receiver. Two types of effect result from multipath. The firsteffect is that deep fading of the GPS signal can result from the destructiveinterference of the direct and multipath signals. The magnitude of thiseffect is largely dependent on the GPS signal fade margin. In general, a deepfade will mean either a temporary loss of the signal or a delay in signalacquisition.
The second effect of multipath on the receiver is to cause an error inthe delay-lock-loop pseudo-range tracking. The nature of this error dependsupon whether the multipath delay is greater or less than 1.5 CiA code chips.The dependence of the multipath delay upon elevation angle and aircraftaltitude is shown in Fig. 3-3. As seen in the figure, the multipath delayalways exceeds 1.5 code chips for altitudes greater than 8600 feet andelevation angles greater than 5°. In general, tracking errors due tomultipath delays greater than 1.5 chips can be removed by software tests,whereas errors due to delays less than 1.5 chips cause tracking bias errorswhich cannot be removed. These effects are discussed more fully in thefollowing paragraphs.
3.2.2 Multipath Delay Greater Than 1.5 Microseconds
If the multipath delay is greater than 1.5 ~sec (1.5 CiA code chips), thedelay-lock-loop (DLL) discriminator characteristic can develop a second stableoperating point, as shown in Fig. 3-4. In this case, the DLL may track thepseudo-range of the multipath signal rather than the direct signal, resultingin a large (> 1500 ft) pseudo-range error. However, since the receiver issequential, reacquiring each satellite once per second, this error is likelyto be manifested as an occasional false lock with accompanying large rangechange. Thus, the receiver control software can eliminate this type of errorby tracking the pseudo-range and rejecting any unrealistically large rangechanges.
3.2.3 Multipath Delay Less Than 1.5 Microseconds
In the case where the multipath delay is less than 1.5 chips the effecton the DLL performance is more serious. Figure 3-5 shows that the effect ofmultipath in this case is not to create a second stable operating point, butrather to introduce a bias into the original operating point. This bias,furthermore, may not be discernable by the position software. The magnitudeof this error depends upon both the magnitude of the multipath relative to thedirect signal and magnitude of the multipath delay.
17
Altitude(ft. x 1000)
15
10
5
o
5 Deg. mask - 1. 5 chip intercept altitude
Delay > 1. 5 Chips
Delay < 1.5 Chips
o 5 10 20 30 40 50 60 70 80 90
Elevation Angle (del~')
FIG. 3-3. Altitude vs Elevation Angle For1.5 Chip Multipath Dehy
18
Am
pli
tud
e
t d>
1.5
Ch
ip
-1.0
-0.5
-1.0
-1.0
No
rmal
Op
era
tin
gP
oin
t
..M
ult
ipath
~Mul
tiPa
th(i
np
hase
wit
hre
sp
ect
tod
irect)
3.0
Co
de
Ph
ase
(Ch
ips)
FIG
.3
-4.
Eff
ect
of
Mu
ltip
ath
Dela
yG
reate
rth
an
1.5
Ch
ips.
Am
pli
tud
e
+1
.S
t d<
I.S
ch
ip
(in
ph
ase
wit
hre
spect
tod
irect)
Dir
ect
+M
ult
ipath
Cod
eP
has
e(C
hip
s)
2.0
1.0
1.S
Op
era
tin
gP
oin
tA
sS
hif
ted
by
Mu
ltip
ath
DLL
Tra
ck
ing
Err
or
Due
toM
ult
ipath
+1
.
-1.0
••
-2
.0
N o
-1.5
FIG
.3
-5.
Eff
ect
of
Mu
ltip
ath
Del
ayL
ess
Tha
n1
.5C
hip
s.
Hagerman (Ref. 9) has studied the effect of multipath on both thecoherent and noncoherent Delay Lock Loop. Two regions of DLL operation in thepresence of multipath have been defined. In Region I, the DLL is tracking thedirect signal, whereas in Region II the DLL is tracking the multipath signal.The results of Hagerman's analysis for the noncoherent DLL are shown inFigs. 3-6 and 3-7. (Note: Hagerman's results are in terms of the P-code andthus must be multiplied by 10 for the CiA code). Although the Region IIerrors are much larger than the Region I errors, they are less probable,especially in the case of a non-coherent delay lock loop. The worst-caseerror therefore can be taken from the results for Region I.
It is seen from these results that the worst-case error depends on therelative multipath amplitude. If the multipath amplitude can be limited to0.2, then the expected (mean) tracking error is 10 feet and the rms error is50 feet. The multipath amplitude depends in turn on the antenna gain vs.elevation characteristic and the reflection coefficient of the surface.According to Figure 3-8 (Ref. 10) the worst case reflection coefficient at 5°is 0.7. If the antenna gain is 0 dBIC at 5°, then the required antenna gainat -5° can be calculated from:
Relative multipath =amplitude at 5°
Antenna Gain (+5°)------------------- x Reflection coefficient at 5°Antenna Gain (-5°)
If it is assumed that the antenna gain at 5° is at least 0 dBIC, then therequired antenna gain at -5° should be no greater than -5.5 dB.
The results of the foregoing discussion are summarized in Table 3-2.
TABLE 3-2
Initial Antenna Gain vs. Elevation Angle Requirement
Elevation Angle
30° to 90°5° to 30°
-90° to -5°
Antenna Gain
> -2 dBIC> 0 dBIC< -5.5 dBlC
This characteristic is illustrated in Fig. 3-9. Also shown is thespecification developed by General DYnamics (Ref. 11). The single differencebetween the two characteristics is that General Dynamics specifies < -7 dBICbelow 10°, perhaps due to the difficulty of achieving -5.5 dBlC at 5°. Thereceiver performance characteristics, however, will be essentially equivalentfor both specifications.
21
:: 45.:
! 40
aw 35CIl:
II:0 30~
3CIl: 25CIl:w::z:...c0.i=oJ:)
2QW...UW0.)(W
a: 180
RELATIVE MULTIPATH AMPL.ITUDE
RELATIVE MULTIPATlo4 AMPLITUDE
FIG. 3-6. Expected Noncoherent DLLTracking Error. 20 dB Carrier
Tracking Margin (Ref. 15)
22
P CODE.DLL TRACKINODIRECT 81GNAL
P CODE.DLL TRACKING
MULTPATH 8IGNAL
9O~--r-"""T'-...,...-.....--.......--r---.--...,...-.....----.
~ 10iswCII: 60
!l;" SOCII:o:"0w
RELATIVE MULTIPATH AMPLITUDE P CODEDLL TRACKINGDIRECT SIGNAL
~ 10iswCII: 60
!l;" SOCII:o:40w
~ 30•~~ 20...J
Sl) 10-
RElA TIVE MUlTIPATH AMPLITUDE
P CODEDlL TRACKING
MULTIPATH SIGNAL
•
FIG. 3-7. 10 Noncoherent DLL TrackingError, 20 dB Carrier
Tracking Margin (Ref. 8)
23
1.0
0.8
t- Z II!0
.82 II
.II
.W 0 0
Ol'
llz 2
0"\.
lt-
0.4
61
'''0
",e"
Nw
.l""
... II.
W at
0.2
0.0
o1
02
0
AN
GL
EO
FIN
CD
EN
CE
(DE
G)
80
40
FIG
.3
-8.
Ty
pic
alV
ert
icall
yP
ola
rize
dS
ign
alR
efle
ctio
nC
hara
cte
rist
ics
(Ref
.2
).
Gai
n
N In
r:'..
ener
alD
ynam
ics
Sp
ecif
icati
on
>-2
db
ic
<5
.5d
bic
FIG
.3
-9.
INIT
IAL
AN
TE
NN
AG
AIN
VB
EL
EV
AT
ION
AN
GL
ER
EQ
UIR
EM
EN
T
However. the way in which groups of pseudo-range measurements (up to 10measurements. from as many satellites) are batch processed in the positioncomputation makes the gain characteristic for 1Il1ltipath suppression lessstringent. The batch least-squares position estimation method reduces thesensitivity of the position solution to an individual nultipath-corruptedpseudo-range measurement. With least-squares fit processing it was determinedthat it is possible to redesign for a random error of 120 feet. as long as theassociated bias error was limited to 50 feet, wor:st case. The below-horizonantenna gain requirements were recalculated to mel~t this revisedmultipath-induced range error bias limit assuming a ground bounce reflectioncoefficient of 0.7, maximum.
The revised specifications are depicted in Table 3-3 and in Fig. 3-10.Note that the below-horizon gain for elevation a~~les in the range minus 5 tominus 30 degrees need only be 2 dB less than the gain over the above-horizon 5to 30 degree sector.
TABLE 3-3. REVISED GPS ANTENNA SPECIFICATION
Elevation AntennaAngle, ( t1) Gain. G( tI)
30° to 90° ~ -2 dBIC5° to 30° ~ o dBIC
-5° to 5° " o dBIC-30° to -5° " -2 dBIC-90° to -30° " -4 dBIC
The measured performance of an antenna (Fig. 3-11) purchased from ChuAssociates using the revised specifications is reported in Section 4.2.1.
3.3 Preamplifier
The receiver preamplifier specification was based on considerations ofnoise figure and out-of-band rejection. A nominal link budget, based onreasonable fade margins, resulted in the need for a 4 dB receiver noisefigure. Since 3 dB must be allocated to the preamplifier and 1 dB to lossesassociated with filters and connectors, it is necessary that the preamplifierbe located at the antenna. The dual-channel receiver developed by STI furtherassumed a nominal gain of 50 dB with 10 dB cable loss between the amplifierand receiver.
A commercial amplifier, an Avantek AM-1664M, was purchased with thefollowing specifications:
1575.42 MHz
Bandwidth
Gain
Noi se Figure
20 MHz (3 dB)
52 dB
2 dB
26
./
;'
/'
./
/
~/
I I I , \--
'
DO
TT
ED
LaE
SH
OW
S"
p"T
TE
RN
WH
ICH
ME
ET
STH
ES
PE
CF
IC"T
ION
.
-2d
Ole
- -
~-4
dO
le
FIG
.3
-10
.R
EV
ISE
DG
PS
AN
TE
NN
AC
HA
RA
CT
ER
IST
IC
I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
...._1- 1.50 DIA -.
4.63
1-,---.....L.---t---------=:J .....
-1
TYPE TNC CONNECTOR
FIG. 3-11. GPS ANTENNA, OUTLINE DIMENSIONS
28
•
3.4 Dual Channel Receiver Hardware
3.4.1 Architecture and Design Features
The receiver, developed by Stanford Telecommunications, Inc. for LincolnLaboratory, consists of two identical channels fed by a common down-converterand timer/synthesizer as shown in Fig. 3-12. Each channel performs digitalfunctions in Z8000 microprocessors. The functional schematic of Fig. 3-13provides further detail for one of the two channels with analog (hardware) anddigital (software) functions to the left and right of the dashed verticalline. Software switches SWI and SW2 are programmed so that during acquisitionSWI is in the "selected frequency" position and SW2 in "code search". Duringthe navigation mode a steady state condition will occur in which SWI isprogrammed to "AFC" and SW2 to the "DLL" Position. The allocation of receiverfunctions between the Z8000 microcomputers and the position software in theLSI 11/23 is shown in Table 3-4.
The IF noise bandwidth, BIF' is determined by the integrate and dump (Iand D) circuits and by the digital accumulators according to the expression
1
2 T M
where T is the sampling period and M the number of samples accumulated. Theoutput of the low-pass filter is sampled at a 2 KHz rate. Additionalfiltering is provided by the post-detection integrator (PDI).
The code select circuitry can operate in the punctual mode (P) for codesearch and data extraction, or in the early-late dither mode to control thedelay locked loop (DLL). Characteristics of the search detector, and the AFC,code and AGC loops are shown in Table 3-5.
Figure 3-14 shows theAFC lock detector and the carrier and noise ratioestimator embodied in each STI receiver channel. Characteristics of thesecircuits are given in Table 3-6.
3.4.2 Receiver States
When a receiver channel is provided with commands and initialization databy receiver management software in the LSI 11/23 it responds by executing oneof the following Z8000 programs:
• initialization• acquisition• transition• navigation• data demodulation
29
w o
,------------------------,
I
91I
-I
IC
OR
RE
LA
TO
R~
IR
S-2
32
I~
i..
-}J
PR
OC
ES
SO
RI
-.
I----
--I
CO
DE
R~
IP
RE
AM
PI
~
IT
II
I-
DO
WN
~I
I.
CO
NV
ER
TE
RD
UA
LN
CO
LS
I-1
11
23
II
~
iR
EC
EIV
ER
IS
OF
TW
AR
EI
"---
II
TIM
ER
II
SY
ST
EM
IS
YN
TH
ES
IZE
Rr
CLO
CK
-..r-
---
II
!I
IA
-I
ID
UA
LN
CO
r'-r
-I
I~
--p
IRS
-23
2fJ
PR
OC
ES
SO
RI
1---+
I+
28
VP
OW
ER
1---+
~I
D.C
.I
•S
UP
PL
YI--
----.
CO
DE
RI
I1-
--+~
II
~I
Il-.
...II
CO
RR
EL
AT
OR
L--
-I
I-
I'"
IL
.--
J
FIG
.3
-12
.E
XP
ER
IME
NT
AL
DU
AL
CH
AN
NE
LR
EC
EIV
ER
AN
AL
OG
DtO
rrA
L...
•D
AT
AD
EM
OD
1-1
0K
Hz
M(f
l,.
D1:
LO
WP
AS
SI
1
IIt
OI
I I•
A""
I IA
GC
'-1
0K
Hz
QI
M()
2'of
DI
1:L
OW
PA
SS
I1
1-S
TO
RD
ER
FIL
TE
R
8W
1F
IXE
DF
RE
QU
EN
CY
CA
RR
IER
wN
CO
f-'
..A
FC
..E
-P
-L
CO
DE
SE
LE
CT
DO
PP
LE
RT
HR
ES
HO
LD
CO
DE
SE
AR
CH
CE
LL
LO
GIC
LO
GIC
CO
DE
CO
DE
GE
Nt!
RA
TO
RN
CO
aW
21
-ST
OR
DE
RD
ITH
ER
LO
GIC
DLL
er1l
.TE
R..,
1-S
TO
RD
ER
AG
CL
OO
PD
/AF
Il.T
ER
~IG.
3-1
3.
FU
NC
TIO
NA
LS
CH
EM
AT
ICO
FR
EC
EIV
ER
TABLE 3-4
RECEIVER CHANNEL - POSITION SOFTWARE lmNCTIONAL ALLOCATION
REQUIREMENT
Acquistion
SequentialTrack
Z8000RECEIVER CHANNEL SOFTWARE
Sequential detectionPull-in and trackData synchronization
Limited sequential detectionPull-in and trackDwell sequencingInter-dwell prepositioning
LSI 11/23POSITION SOFTWARE
~, selectionAcquisition prepositioningSearch controlAcquisition sequence control
~r constellation revisionNE~w SV prepositioningSE~quence monitoring andcontrol
IPseudo-range I Code state samplingI
PBeudo-range construction&~biguity resolution
IIonispheric and tropospheric
Data Bit demodulation I D:::~::::::::ization (singleI
Demodulation I Word recovery I ehannel)Parity checking Data unpacking
I L:emodulation monitoring andeontrol
-----------------------
32
TABLE 3-5
RECEIVER LOOP CHARACTERISTICS
ISEARCH DETECTOR AFC LOOP CODE LOOP AGC I
Square law power Delayed cross ± 0.5 chips early/ Maintainsdetector product detector late power detector constant power
'" sin 2nllf L
I I I Ilopen loop lIst order loop Icarrier (AFC or PLL) Deals with up
Ioperation loop aided to 10 satellitesignals.
AGC sets noise Variable loop Variable loop Variable loopfloor bandwidth bandwidth bandwidth
Two search rates: Primary loop Primary loop Primary loop50 chips/sec bandwidth 5 Hz bandwidth 5 Hz bandwidth 1 Hz
I83.3 chips/sec and 10 Hz and 10 Hz (for 1st satellite I
acquisition only),
/MUltiPle freq.10 Hz otherwise
Operates withIcell search early/late code Supplies power
reference to lockdetectors
Upon detection doesfalse alarm check
TABLE 3-6
RECEIVER AFC LOCK DETECTOR AND C/No ESTIMATOR CHARACTERISTICS
AFC LockDetector
Detector output'" cos 2nllh
Uses narrow bandIF (BIF=50 Hz) fornarrow frequencydiscrimination
AGC compensated tobe insensitive to noiseand gain variations
Also serves as codesync detector
Detection bandwidthadjustable from 5 Hz up
33
C/No Estimator
Measures AGCnormalized power
Output monotonicwith C/No
Estimate for eachsatellite
Filtered overmultiple intervals
Detection bandwidth I5 Hz
I I•
I~
AN
AL
OG
IDIG
ITA
L
I I I IM
DE
LA
YT
1&
0I
1;
PO
I
IG
AIN
NT
HR
ES
HO
LD
AF
CL
OC
KI
CO
NT
RO
L1;
LO
GIC
DE
CIS
ION
I I ID
EL
AY
TF
RO
MA
GC
w ~I I I I
I\2
L..:.
JP
OI
a1
&0I
MG
AIN
L.P
.C
lNo
1;C
ON
TR
OL
FIL
TE
RE
ST
IMA
TO
RI
()2
FR
OM
AG
C
FIG
.3
-14
.A
FC
AN
DC
/NO
ES
TIM
AT
OR
The initialization routine establishes operating parameters and checksstatus. The acquisition routine is illustrated in Fig. 3-15. Note that theIF bandwidth, BIF , is changed several times during the acquisition sequence,which illustrates the significant advantage of software realizations ofreceiver detection algorithms.
Figures 3-16 and 3-17 show the transition and navigation state diagrams.
Figure 3-18 shows the detailed events that occur during each 220millisecond dwell internal in navigation mode. The prepositioning data willhave a ± 0.16 chip uncertainty for an SV entering the dwell from phasetransition. Should lock fail to occur twice, the receiver will automaticallycommence a ± 1 chip search. If two additional failures occur, a ± 2 chipsearch is initiated. After a total of 6 failures, the receiver managementsoftware (in the LSI 11/23) will command a broader search by relocation of thesearch window. After several subsequent unsuccessful attempts the SV will bedeclared to be potentially faded and less aggressive techniques will beapplied to reacquire it.
During normal operation new SVs will become visible. Figure 3-19 showsthe state diagrams for the acquisition and data demodulation of a new SV.
The receiver hardware was physically assembled in an ATR Chassis as shownin Figs. 3-20 through 3-22. The receiver modules are mounted in amultibus-type chassis, and consist of a mixture of off-the-shelf commercialand custom-built boards. It should be noted that the actual size of acommercial equivalent will be significantly (~8-10 times) smaller because:
a) An extra board to provide the 9K byte ROM memory for each Z8000 wasincluded (the Z8000 boards provided only 8K of local ROM).
b) LSI devices are currently being developed for the CiA Coder and NCOfunctions and,
c) much of the board area is not populated.
35
SEARCH RATE: 50 CHIPS PER SECOND
BIF • 1000 Hz60 ms.
~C· 4 Hz, BLA = 10 Hz
BIF • 500 Hz
140 ms. BLC; = 2 Hz, BLA = 10 Hz
INITIALACQUISITION
SET AGeNOISE FLOOR
WIDE BANDSEARCH
NARROW BANDRE-SEARCH
LAFC, CODE LOOP
PULL-IN ill
AFC, CODE LOOPPULL-IN 1/2
AFCFINE TUNE
BIT SYNCOPERATIONS
FRAME SYNCOPERATIONS
DATA DEMODULATIONCINO ESTIMATION
( EXIT )
TIMEBUDGET
1 SEC(FIRST SATELLITEONLY)
MODE, SATELLITEDpENDENT20.66 sec xilCELLS+ 0.60 - 1.12 sec.
18 ms.
220 ms
2 sec
7.22 - 13.22 sec
10.8 - 22.8 sec.
REMARKS
BL
• 1 Hz (AGC LOOP BANDWIDTH)
AGe LOOP SETTLES
BIF • 1000 HzN CELLS AS REQUIREDPLUS FALSE ALARM CHECK
BIF • 500 Hz (IF BANDWIDTH)3 CELLS, 1 CODE POSITION
BIF • 500 Hz
BLC = 2 Hz, BLA = 10 Hz
BIF = 500 Hz
BLC = 2 Hz, BLA = 10 Hz
BIF • 50 Hz
BtC = 2 Hz, BLA = 10 Hz
B • 50 HzIF
BLC = 2 Hz, BLA = 10 Hz
FIG. 3-15. Receiver Acquisition State DiaRram.
36
ENTER TRANSITION)
I
WIDE BANDSEARCH
,FALSE ALARM
CHECK
~
AFC, CODE LOOPPULL:"'IN III
,AFC, CODE LOOP
PULL-IN 112
~
( EXIT )
MODE, SATELLITEDEPENDENT680-740 ms.
20 ms.
60 ms.
140 ms.
BIF -1000 Hz
+ 15 CHIP SEARCH
BIF • 500 Hz
2 TRIES
BIF • 1000 HzB
LC= 4 Hz, BLA = 10 Hz
BIF • 200 HzB
LC= 2 Hz, BLA = 10 Hz
SEARCH RATE: 50 CHIPS PER SECOND
BLC • CODE LOOP BANDWIDTH
BLA
• AFC LOOP BANDWIDTH
FIG. 3-18. Receiver Transition State Diagram.
37
SLEW CODERS,WAIT FOR BIT EDGE
AFC, CODE LOOPPULL-IN
AFC LOCKDETECTION
TIMEBUDGET<20 mS
80-100 mS
100 mS
BIF • 200 Hz
BLC • BLA • 10 Hz
BIF • 200 Hz
BLC • BLA • 5 Hz
BLC • CODE LOOP BANDWIDTH
BLA • AFC LOOP BANDWIDTH
FIG. 3-17. Receiver Navigation State Diagram.
38
**FG
B+
RGB
=20
ms
+I-
SEA
RC
H
BL
•1
0
-lcu:~
1-AFC
PU
LL
IN~
«9
0%
)f&
4~'
_----II
"'-
--_
......
.._-.-
..,...-
-.....L
.'_
_.L
..-_
....
.L._
_.L
-_..
....
L_
_.L
-__
I..
FGB
'FO
RSE
ARC
H~
1{2
ms
FGB'~
BI
•1
00
0B
I-
50
0B I
-2
00
f.--A
FCPU
LLIN
«9
9%
)~-
AFC
LOCK
DET
ERM
INA
TIO
N-I
20--
1I-
-I
(5SA
MPL
ES)
~~
"."'~_-l'--~'---::,~I-n-_L..-_L..-_L..-_.L...-_..L-_...L.-_"
*B•
200
B=
200
II B
•5
L AFC
LOCK
DET
ERM
INA
TIO
N(4
SAM
PLES
)
±1C
HIP
SEA
RCH
;M
ODE
1
NOSE
ARC
H;
MOD
E0
B-
10
B=
5L
L
~FA
~IAFC
PULLI~
AFC
LOCK
1-r
CHEC
...«
80
%)
-rD
ETER
MIN
ATI
ON
g(3
SAM
PLES
)RG
BI
I.1
I,
II~
±2C
HIP
SEA
RCH
MOD
E2
FGB
'
SEAR
CH
B I=
10
00
B=
500
I B L=
10
BD
200
I BD
10
LS
earc
hR
ate:
50C
hips
per
seco
nd
*BI
IsS
ing
le-S
ided
**FG
BIs
The
Fro
nt
Gua
rdB
and
RGB
IsT
heR
ear
Gua
rdB
and
FIG
.3
-18
.N
AV
IGA
TIO
NS
TA
TE
22
0M
SIN
TE
RV
AL
MO
DE
SA
ND
TIM
ELI
NE
S
18 ms.
BIF • 500 Hz3 CELLS, 1 CODE POSITIONPLUS FALSE A.LARM CHECK
BIF • 1000 Hz60 InS. BLC = 4 Hz, BLA = 10 Hz
BIF • 500 Hz
140 ms. BLC • 2 Hz, BLA 10 Hz
220 ms.B
1F• 500 Hz
ELC = 2 Hz, B = 10 HzLA
BIF = 500 Hz2 sec. BLC = 2 Hz, B
LA= 10 Hz
SATELLITEADDITION STATE )(DUAL CHANNEL)
.-WIDE BAND
SEARCH
•NARROW BAND
SEARCH
•AFC, CODE LOOPPULL-IN III
AFC, CODE LOOPPULL-IN 112
•AFC FINE TUNE
I
BIT SYNCOPERATIONS
•FRAME SYNCOPERATIONS
•DATA
DEMODULATION
•( EXIT J
TIMEBUDGET
6.3 SEC(±150 CHIPS)
7.22 - 13.22 sec
10.8 - 22.8 sec.
BIF • 1000 Hz1 CELL PLUSFALSE ALARM CHECK
BIF = 50 Hz
BLC = 2 Hz, BLA
= 10 Hz
nIF = 50 Hz
BLC
= 2 Hz, BLA = 10 Hz
FIG. 3-19. Satellite Addition State Diagram.
40
FIG
.3
-20
.D
UA
LC
HA
NN
EL
RE
CE
IVE
R
FIG
.3
-21
.D
UA
LC
HA
NN
EL
RE
CE
IVE
R-
DO
WN
CO
NV
ER
TE
R
FIG
.3
-22
.D
UA
LC
HA
NN
EL
RE
CE
IVE
R-
TO
PV
IEW
CH
2
CH
1
3.5 Position Software
3.5.1 Overview
The GPS T and E Equipment software is organized into four functionalareas as shown in Fig. 3-23. These functional areas are described inTable 3-7. Section 3.4 described receiver functions resident in the Z8000microcomputers. The Navigation, Instrumentation and Control and Display Unitsoftware designs are described in Sections 3.6-3.8.
This section describes the position software located in the LSI-11/23Position Processor. The position software, developed by Intermetrics, Inc.,is organized into several functional areas as shown in Fig. 3-24. Thefollowing paragraphs describe each module:
3.5.1.1 System Mode Control
This module maintains control over all other modules.receiver channel modes, prepositions the receiver channelscode phase, and updates the SV list as visibilities changeaircraft altitude.
3.5.1.2 System Time Management
It controls thein frequency andwith time and
The system time management module establishes and maintains system timebased on the 50 Hz clock sent from the dual channel receiver. Aftercorrection to GPS time following acquisition, it maintains time for datatagging and display of GMT to the operator or pilot.
3.5.1.3 Satellite Selection
This module makes use of a stored almanac, stored current position, and abattery-operated calendar clock to select four SVs for initial acquisition.It also will implement a search strategy should less orbital information beavailable or the clock have failed. While in operation it also maintains alist of up to 10 visible SVs.
3.5.1.4 Receiver Management
The functional task breakdown between the position software and the dualchannel receiver was shown in Table 3-4. A high level flow chart whichdescribes the major receiver states from the point of view of the receiverfirmware is provided in Fig. 3-25. The detailed activities of this moduleduring acquisition were shown in Fig. 3-2.
3.5.1.5 GPS Navigation Data Management
This module receives parity-checked navigation message words from thedual channel receiver. It monitors the age of the stored SV ephemeris data,requests updates, monitors data collection and maintains a full data almanacfor all SVs.
44
PO
SIT
ION
DA
TA
PR
OC
ES
SO
RA
CQ
UIS
ITIO
NU
NIT
CO
NT
RO
LA
ND
DIS
PL
AY
~C
HA
NN
EL
-1N
AV
IGA
TIO
NU
NIT
Z8
00
0--
..S
OF
TW
AR
E
PO
SIT
ION
~~
~D
ISP
LA
Y~
SO
FT
WA
RE
SO
FT
WA
RE
..r
Z8
0IN
ST
RU
ME
NT
AT
ION
SO
FT
WA
RE
~C
HA
NN
EL
-2..,
-..Z
80
00
LS
I-1
1/2
3L
SI-
11
12
3
<E
E;.
...-
---
SA
T.
I.D
.<EE~---
MO
DE
<W
AY
PO
INT
----'>
...P
SE
UD
OR
AN
GE
NA
VD
AT
A----=>~
LA
T/L
ON
G.
----i>
;;.
RA
NG
E/B
EA
RIN
G,
ET
C
FIG
.3
-2
3.
SY
ST
EM
SO
FT
WA
RE
TABLE 3-7
SOFTWARE FUNCTIONAL AREAS
• Receiver Firmware (ZSOOO)
Directs detailed acquisition of SVs. maintains receiverfeedback loops (AFC. DLL. etc.). computes pseudo-range.demodulates navigation data and maintains dwell tracking. Isunder the control of Receiver Software.
• Position Software (LSI-ll/23)
Selects satellites. controls Z8000 receiver activities.computes position estimates in latitude-longitude coordinates andmonitors performance.
• Navigation and Instrumentation Software (LSI-ll/23)
Accepts position estimates from the receiver managementsoftware and waypoint definitions from the cockpit display.Computes bearing. distance and time estimates. controls coursedeviation indicator and provides data to the control and displayunit software. Collects. distributes and records data.
• Control and Display Unit Software (ZaO)
Establishes specific formats and protocol for data entry anddisplay on the cockpit display and control unit.
46
UP
CL
OC
KS
YS
TE
MT
IME
TA
GS
IT
IME
MG
MT
I'w
PS
EU
DO
-RA
NG
E-
NA
VS
OL
UT
ION
UP
DA
TE
S
CA
RR
IER
·D
OP
PL
ER
ME
AS
.I
NA
VIG
AT
ION
-P
RO
CE
SS
ING
TR
AC
KE
RI I
,M
EA
S.
SV
PO
SIT
ION
CO
LL
EC
TIO
NI
ST
AT
US
SV
NA
VD
AT
AG
PS
NA
VN
AV
DA
TA
DA
TA
MG
MS
TA
RT
U
ID
AT
A
IS
VP
OS
ITIO
N
IS
VA
ICP
OS
ITIO
N
Ir-
SE
LE
CT
ION
-
IS
VA
SS
IGN
ME
NT
S,
,~
----R
EC
EIV
ER
RE
CE
IVE
RS
YS
TE
MS
TA
RT
UP
DA
TA
DA
U
CO
MM
AN
DS
MG
MT
MO
DE
CO
MM
AN
DS
MO
DE
CT
RL
INT
ER
FA
CE
DA
-,
--
DU
AL
IC
HA
NN
EL
--~~I-_""'--
LS
I1
1/2
3P
OS
ITIO
NP
RO
CE
SS
OR
RE
CE
IVE
R
I
FIG
.3
-24
.P
OS
ITIO
NS
OF
TW
AR
ES
YS
TE
MB
LO
CK
DIA
GR
AM
POWER LIP
OPR RESET "'~I(FROM ANY STATE) .. INITIALIZATION
ON PILOT START-UP CMD
ACQUISITION
_ __ OPR
CONFIGURATION
ON OPR COMMAND
l~
DIAGNOSTICS
4 SUCCESSFULSV ACQ.
(*17 CHIP±750 Hz)
"
,~ ON TRANSITIONSEQUENTIAL
TRACK FAILURE
4 SV SEQUENTIALTRACK
(t.16 CHIP %200 HZ),
TRANSITION
ON LESS THAN3 SV VISIBILITY
NAVIGATION
FIG. 3-25. RECEIVER MODES
48
3.5.1.6 Measurement Processing
This module collects measurements from the receiver channels, resolvesmeasurement ambiguities, compensates for deterministic error effects such astropospheric delay, and computes a position and time fix every 2.2 seconds.
3.5.1.7 Navigation Tracker
This module maintains a current estimate of the aircraft position,smooths position fixes using an alpha-beta tracker and interpolates the datato provide I-second updates to the navigation software.
3.5.1.8 DAU Interface
This module provides linkage with the the Data Acquisition Unit (DAU).It accepts inputs during system start up (position and time if required), andtransfers smoothed navigation estimates (in latitude-longitude coordinates)and aircraft velocity estimates to the navigation software in the DAU. TheDAU interface also accepts receiver channel configuration commands from theDAU, and provides receiver data to the DAU for recording on magnetic tape.The DAU also provides a limited data display capability while in flight. Theinterface to the DAU is a DMA hardware interface since the DAU resides in aseparate LSI-ll/23.
The implementation of the receiver software is under the DigitalEquipment Corporation RSX-IIS operating system. The software was written inRATFOR, a structured FORTRAN language. A top-down design approach was used toachieve developmental flexibility and modular, well documented, maintainablesoftware.
3.5.2 Acquisition Strategy
The basic acquisition strategy was described in Section 3.1. Thissection provides additional detail regarding satellite selection, receiverchannel control and preparation for entry into the transition mode.
While provisions for the cold start mode were incorporated in the design,the typical situation as described here will be a narrow or wide band start asdefined in Table 3-1. Both assume that a battery operated clock provides GMTto ~ 1 minute, and that the present location and an almanac less than 30 daysold are available from a non-volatile memory.
A satellite selection algorithm is used in the acquisition mode in orderto a) select a reasonable set of up to 6 satellites to provide good geometryand shadowing protection during departure if one or two satellites fade duringturns, and b) present an acceptable fix to the CDU within 6 minutes. Later,in the navigation mode, selection is unnecessary as all satellites in view aretracked.
49
Initially, a set of four satellites are selected based on maximizing thesum of the azimuth and elevation differences between all pairs, and providingat least 30° elevation angle to all satellites, if possible. Additionalsatellites, if available. are selected for subsequent acquisition in order toenter the navigation mode with up to 6 in track. Eventually all satellites inview are acquired and placed in the navigation mode.
The time to acquire the first satellite was shown by Intermetrics to havea smaller variance if both channels searched in parallel for the firstsatellite than if each channel searched independently for differentsatellites. Therefore, initial acquisition is done by both channels searchingcooperatively for a single satellite. When the first satellite is acquired,the local clock is corrected to a value within 20 milliseconds of GPS time andthe reference oscillator frequency error is logged to reduce the frequencysearch during the remaining satellite acquisition.
The acquisition strategy includes reacquiring all previously acquiredsatellites following each new acquisition (a detail not shown in Fig. 3-2)because of the need for recent code phase and doppler estimates in order topreposition the receiver channel for the transition mode. If acquisition iscompleted but only three satellites are located, the barometric altitude issubstituted and the initial position fix is computed.
3.5.3 Transition Strategy
When four satellites have been successfully acquired (or if necessary,three plus barometric altitude), the position software commands one of thereceiver channels into the transition mode. In this mode, each satellite isacquired and tracked for 1 second. At the end of each four second schedule afix is attempted. When two successive fixes are successful, the receiversoftware computes prepositioning data in preparation for the navigation mode.
3.5.4 Navigation Strategy
When the position software determines that two successive fixes have beensuccessful, the transition mode is terminated and the receiver channel placedin the navigation mode. Once started the receiver channel Z8000 is able tosustain the 0.22 second per satellite sequential detection of up to tensatellites without further assistance from the position software. Theposition software does, however, have to synchronously request and acceptpseudo-range, code phase, and status data from the navigation mode channelevery 0.22 seconds.
Once one receiver channel is placed in the navigation mode, the secondchannel is used to acquire and track additional satellites in oIder to placethem in a navigation mode slot, and to update ephemeris and almanac data. Asnew satellites rise, a satellite visibility algorithm (run every 3 minutes)enables SV acquisitions by the second channel and allows new SVs to be placedin navigation slots in the first channel.
50
The ten navigation slots are all filled if less than ten satellites arein view t by repeating satellites in the extra slots as necessary. Thissimplifies several algorithms and provides additional margin for satellitesthat are fading but are assigned to more than one slot.
A pseudo-range measurement is declared valid by the receiver channel ifAFC lock occurred during the last dwell interval t and it is within 1000 feetof a predicted value based on a predicted position. When a valid measurementis not received for a particular satellite t the algorithm in Fig. 3-26 isinvoked.
Valid pseudo-range measurements for each 2.2 second scheduling cycle arepropagated to a common solution time using measured doppler and presented tothe position estimation algorithm described in Section 3.5.5.
3.5.5 Position Estimation
A batch-processed linearization (BPL) position fixing algorithm wasselected following a study of alternative techniques. The principaladvantages of the BPL algorithm are that a) it uses all available pseudo-rangemeasurements in every position fix t b) it is a minimum least-squares errorestimate, and c) it can be easily implemented. A description of the positionfixing algorithm is provided in Appendix A.
A separate analysis of tracking filters was conducted which comparedalpha-beta t least-squares second order t and Kalman filters. The functions ofthe filter are to a) accept new position fixes every 2.2 seconds t b) computenew smoothed position estimates at a 1 Hz rate t and c) coast the output for upto 8 seconds* when new position estimates do not occur. Should no estimatesbe available for 8 seconds t the NAV flag is shown to the pilot indicatingunacceptable performance.
The emphasis on low cost and limited aircraft dynamics (200 KTS t 0.5g)led to the selection of a fixed gain alpha-beta filter which had t insimulation, acceptable dynamic performance and the lowest computationalrequirements. The relationship between the position fixing algorithm and thenavigation tracker is illustrated in Fig. 3-27. A more detailed discussion ofthe navigation tracker is provided in Appendix B.
3.5.6 Performance Monitoring
Performance monitoring features are provided in each receiver channel andin the position software. Table 3-8 lists the various parameters that aremonitored in each receiver channel.
* Selected to limit the horizontal position error to 500 feet.
51
ENTER
1DO FOR EACH >SATELLITE
L.....--~
MEAS NOT VALID? )-YES
RECEIVER FAULT?YES
READ CHANNEL STATUS
NO
MEAS INVALIDCOUNT >6
SATELLITEVISIBLE?
NO
YES
INCREMENT MEASINVALID COUNTER
>- YES"'-- J
SET SATELLITE
REPLACEMENT FLAG
OFFSET SEARCHEXHAUSTED?
NO
YES
EXTENDSEARCH
PUT SATELLITE ONACQUISITION LIST
FIG. 3- 26. LOSS OF LOCK ALGORITHM
52
PO
SIT
ION
PR
OC
ES
SO
RD
AU
ME
AS
UR
EM
EN
TI
CO
LL
EC
TIO
NI I I
,I
-I
PO
SIT
ION
RE
AS
ON
AB
LE
NE
SS
-N
AV
CO
OR
DIN
AT
EI
-N
AV
IGA
TIO
N-
CO
MP
UT
AT
ION
CH
EC
K,
TR
AC
KE
R,
TR
AN
SF
OR
MA
TIO
NT
--,.
SO
FT
WA
RE
~I I I I
ME
AS
UR
EM
EN
TI
FIG
.3
-27
.P
OS
ITIO
NS
OL
UT
ION
AN
DM
EA
SU
RE
ME
NT
PR
OC
ES
SIN
G
TABLE 3-8
PERFORMANCE MONITOR PARAMETERS
Receiver Channel Fault Indicators
- AGC gain out of range- CiA coder- Loss of phase lock on synthesizer interrupt- 20-msec system clock not present- NCO frequency word greater than maximum limit- Correlator power not in range- Correlator 2-KHz interrupt not present- CPU fault
Receiver Channel - Position Software Interface Fault Indicators
- Command incorrectly received- Command inconsistent with current mode- Time-out requirement not met
54
The basic performance monitoring algorithm in the position software isshown in Fig. 3-28 and 3-29.
The position software also tests the navigation data for inconsistentvalues such as sudden change in clock bias. Slow drifts are, however, morelikely and may be difficult to detect if operating with a minimum number ofsatellites and high GDOP.
3.5.7 Fail-Soft Techniques
Several fail-soft techniques were developed. If during any mode one ofthe two receiver channels fails, a design was developed in which the remaininghealthy channel provides reduced rate navigation service to the pilotdisplays. The single channel case operates in one of two modes:
a) A mode identical to the high rate navigation mode of the dual channelcase.
b) A mixed mode, illustrated in figure 3-30, which allows datademodulation to occur for one satellite and six satellites to betracked on a 2.4 second cycle.
The latter mode requires a satellitesix satellites for the navigation dwells.implemented and validated in the receiverimplemented in the position software.
selection algorithm to select theThe mixed mode has been fully
channels, and designed but not
As indicated in the description of the position fixing algorithm, Section3.5.5, the position software monitors HDOP and the number of satellites intrack. As a result it takes the actions summarized in Table 3-9.
Another fail-soft technique addressed the problem of rapid recoveryfollowing a momentary power outage. An algorithm was designed which restorednavigation service within one minute of the end of a power outage lasting lessthan 30 seconds. The technique requires that receiver data be remembered andthat a battery backup be provided for the reference oscillator (to sustain theoscillator and system clock for 30 seconds). This technique was notimplemented due to other higher priority work and because it appeared to bestraightforward.
3.6 Navigation Software
The navigation software is organized as shown in Fig. 3-31. It receivessmoothed position estimates from the position software at a once-per-secondrate. The Area Navigation Computation (RNAV) module uses the positionestimates to compute navigation data for display to the pilot. As shown inFig. 3-32, this includes distance-to-waypoint (DIST), bearing to waypoint(BRG) and cross-track deviation (XTK). The cross-track deviation is displayedto the pilot via the course deviation indicator shown in Fig. 3-33; the otherparameters are displayed via the CDU front panel.
55
ENTER
DO FOR EACH SV >-IS PSEUDO RANGE
RESIDUAL>1000 FT ?
YES INCREMENTCOUNTER
IIS COUNTER > 5
YES DECLARE SV BAD
ERASE SV EPHEM~RIS
SET SV REPLACEMENTFLAG
IS SV REPLACEMENTFLAG SET
IS TIME INTERVAL OF LESSTHAN 3 SV TRACK
> 8 SECS?
Iii
TO FIG. 3-29
YES
YES
CHECK VISIBILITYTASK FOR SV'S
(TO REPLACE DELETED SV'S)
RE-ENTERTRANSITION
MODE
ISET NAV QUALITY
TO UNRELIABLE
NOTE: REACQUISITION OF SV REJECTEDFOR BAD PSEUDO RANGE ISATTEMPTED AFTER 800 SECONDDELAY.
FIG. 3-28. POSITION SOFTWARE PERFORMANCE MONITOR
56
FROMFIG. 3-28
IS TIME SINCE LASTFIX > 8 SECONDS
IS SV SET SAME >AS LAST TIME
'-----------'
YES SET NAV QUALITYTO UNRELIABLE
RE-ENTERTRANSITION MODE
IS FIX FLAG SET? >--INCREMENT )-COUNTER
~
IS COUNTER? • >-COMPUTE lit- STEADYSTATE COVARIANCE
OVER LAST K
ENABLE TEST WHENSTEADY STATE WITHCURRENT SV SET
* USING MODEL OF ALPHA-BETATRACKER COVARIANCE
IS POSITION 'f >> 900 FT?
-----."
SET NAV QUALITYTO UNRELIABLE
FIG. 3-29. POSITION SOFTWARE PERFORMANCE MONITOR
57
I2
20
ms
DW
ELL
SV
II-
-FR
ON
T
~b2
NA
VDAT~
BIT
SJ=
:nsv
-1I
IB
AN
DI
123
4I
5I
67
1
I1
.08
SEC
-I·1
.32
SEC
.ID
ATA
AC
QU
ISIT
ION
NA
VIG
AT
ION
VI
00
Ela
psed
Tim
eA
tB
egin
nin
go
fa
2.4
Sec
ond
Cy
cle
0.5
mse
cCO
UNT
LD
ATA
DEI
lOD
BE
GIN
....
....
0
12
34
56
•~s
::t
1IIIIIII-·-til
'11
11
1--1
11
11
1"1
11
11
IFR
ON
TG
UA
RDBA
ND-t
ID
ATA
BIT
II~
.r--
T-~A
T;B
I;;;---J
1-(2
40
mse
c)-l
t
~2
0m
sec
-,
I I.-
tB
EG
INN
ING
OF
A2
.4SE
CO
ND
CY
CLE FI
G.
3-3
0.
Tim
ing
Dia
gram
for
a2
.4S
econ
dC
ycle
Sin
gle
Cha
nnel
Sys
tem
.
TABLE 3-9
FAIL SOFT TECHNIQUES.
I I INumber ofSatellites in Track HOOP Action
Ot 1 t 2 -- Declare Failure;Go to AcquisitionMode.
3 > 10 Use coastedaltitude (Pseudo-baro).
3 < 10 SubstituteBaromentric Altitude.
59
PO
SIT
ION
SO
FT
WA
RE
NA
VIG
AT
ION
SO
FT
WA
RE
PIL
OT
DIS
PL
AY
S
NA
VIG
AT
ION
DA
TA
.I
OP
ER
AT
OR
CD
UI
CO
NT
RO
LA
ND
INT
ER
FA
CE
INT
ER
FA
CE
WA
YP
OIN
TS
EL
EC
TD
ISP
LA
YU
NIT
I(C
DU
I)D
ISP
LA
YC
ON
TR
OL
I\C
OM
MA
ND
SO
AT
AD
ISP
LA
YI
I
ITIO
NE
ST
IMA
TEI
CR
OS
S-T
RA
CK
CO
UR
SE
DE
VIA
TIO
N.
DE
VIA
TIO
NA
ND
IIN
DIC
AT
OR
INE
ER
ING
OA
TAI
-
ST
AR
TU
PD
AT
AI
PO
SIT
ION
PO
SIT
ION
ES
TIM
AT
EA
RE
A..
IP
RO
CE
SS
OR
NA
VIG
AT
ION
INT
ER
FA
CE
CO
MP
UT
An
ON
I(R
NA
V)
DE
SIR
ED
CO
UR
SE
OM
NI-
BE
AR
ING
SE
LE
CT
OR
II
EN
GIN
EE
RIN
GD
AT
A
TO
OA
TA
RE
CO
RD
ING
I
PO
S
'"oE
NG
FIG
.3
-31
.N
AV
IGA
TIO
NS
OF
TW
AR
E.
FU
NC
TIO
NA
LB
LO
CK
DIA
GR
AM
XTK
MA
GN
ET
ICN
OR
TH
WA
YP
OIN
T
~~~~/~-?~2
WA
YP
OIN
TOE
S~EO
'TR
l'ClC
.lCO
UR
SE
l
1 FR
OM
PR
ES
EN
TP
OS
ITIO
N FIG
.3
-32
RN
.A
VG
EO
ME
TR
Y
COl NEEDLE
OBS KNOB
AZIMUTH CARD
NAV WARNINGFLAG
a
TO-FROM FLAG
FIG. 3-33. COURSE DEVIATION INDICATOR/OMNI-BEARING SELECTOR
62
The CDU interface (CDUI) module receives navigation data from the RNAVmodule and passes the data to the CDU for front panel display. CDUI alsoreceives pilot data from the CDU consisting of such data as startup mode, CDIsensitivity, and active waypoint number. The RNAV module passes thecross-track deviation to the CDI and reads the desired course from the OBS.
3.7 Control and Display Unit
The function of the Control and Display Unit (CDU) is to interface thepilot to the test and evaluation equipment. The CDU front panel layout isshown in Fig. 3-34. The front panel consists of an alphanumeric display withtwo lines of 16 characters per line, a 32 position keyboard and three controlswitches. The alphanumeric displays are of the LED lighted-segment type andare 0.25 inches high. These displays are bright, easily read at a distance offive feet, and feature a wide viewing angle (+ 55°). The keypads are compact,off-the-shelf, units with 0.5 inch button spacing; control switches arestandard rotary types.
CDU functions can be divided into two groups: flight plan data entry andnavigation data display. Flight plan entry occurs before takeoff while theGPS receiver is still in the acquisition mode and has not yet obtained aninitial position fix. The pilot uses this time to enter the desired flightplan via the CDU keyboard, with appropriate prompting messages from thealphanumeric display. The flight plan data is passed to the NavigationSoftware via an ARINC 429 interface, as shown in Fig. 3-35. A sample flightplan for a trip from Atlantic City, NJ to Hanscom Field, HA is illustrated inFig. 3-36. The flight plan data is entered in the form of waypoints along theplanned flight path. The data file generated by this data entry process istermed the Stored Waypoint data file. 1 The Stored Waypoint data file for thesample flight plan is shown in Table 3-10.
The waypoints along the flight plan can be of several types: VOR,intersection, latitude/longitude or artificial. VOR and intersectionwaypoints are entered by three-and five-letter labels, respectively.Artificial waypoints are specified in terms of range and bearing from a VOR,intersection or latitude/longitude. As the waypoints are entered, they areaccumulated in the Stored Waypoint data file. The navigation softwaredetermines the latitude and longitude of each waypoint using look-up tablesand calculations as required. The Stored Waypoint data file also includes thecourses to and from each waypoint. These courses can be entered by the pilotor computed by the navigation software.
The other function of the CDU is to display navigation data while theaircraft is in flight. The data required to support the navigation display istermed the Active Waypoint data file. This data file is maintained by thenavigation software and passed to the CDU via the ARINC interface. The ActiveWaypoint data file must be updated at a rate such that the CDU navigationdisplay is always current (i.e., every second).
1. Due to schedule constraints, the flight plan data entry function was notimplemented in the navigation software. Instead, the stored waypoint file waspredefined in the navigation soft\~ .~ ~0 include a set of waypoints for thetest flight area. The effect was che same as if the pilot had entered thewaypoints manually.
63
FIG
.3
-34
.C
ON
TR
OL
AN
DD
ISP
LA
YU
NIT
(CD
U)
IP88
-269
3I
DA
UC
DU
AC
TIV
EW
AY
PO
INT
IK
EY
BO
AR
DD
AT
AF
ILE
I IP
OS
ITIO
NC
DU
AR
INC
42
9I
AR
INC
42
9C
DU
ST
IMA
TE
INT
ER
FA
CE
SO
FT
WA
RE
-D
ISP
LA
Y
I IS
TO
RE
DI
WA
YP
OIN
TS
WIT
CH
ES
DA
TA
FIL
E
E
FIG
.3
-35
.D
AU
-CD
UIN
TE
RF
AC
E
",-,, I-SKR
(MAP)
7
_",.'1.0 -nml
/3S· ~
// 4
·-ODM
ACY
FIG. 3-36. SAMPLE FLIGHT PLAN: ATLANTIC CITY TO HANSCOM FIELD
66
TABLE 3-10
STORED WAYPOINT DATA FILE: ATLANTIC CITY TO HANSCOM FIELD
WPT TYPE1 IDENT TO FROM LATITUDE LONGITUDE OFFSET
0 L/L ACy2 90° 39°27.4'N 74°34.7'W None
1 INT BRIGS 90 59 Lookup Lookup
2 INT MANTA 59 56
3 VOR HTO 56 72
4 INT CHUMR 72 11
5 VOR ORW 11 30 ..
6 VOR PUT 30 25
7 AWP V431 2 25 111 Calculate Calculate GDM 111 °16.0 nmi
8 INT LOBBY 111 113 Lookup Lookup None
9 NOB BE 112 112
10 L/L BED2 112 114 42°28.2'N 7l 0 17.4'W None
11 NOB SKR 114 294 Lookup Lookup None
Note 1: L/L - Latitude/LongitudeINT - IntersectionAWP • Artificial WaypointNDB - Non-directional Beacon
Note 2: IOENT Optional for L/L and AWP types.
67
The contents of the Active Waypoint data file are shown in Table 3-11.This data includes such information as the range and bearing to the activewaypoint, i.e •• the waypoint the aircraft is currently heading to or from.The active waypoints can be selected by the pilot via the CDU keyboard or canbe automatically incremented as the waypoints along the flight plan areoverflown. The specific data displayed is selected by keyboard entries assummarized in Table 3-12. The navigation data appears in the "DISPLAY" fieldof the CDU display.
A typical CDU navigation display is shown in Fig. 3-37. The displayindicates that the CDU mode is "OPERATE", the active waypoint is number 6(Putnam VOR), and that the aircraft is 19.2 nmi away from the waypoint andtraveling toward it. Alternate data can be obtained by pushing one of thekeys in the leftmost three columns. For example, pushing "TTG" would displaythe time-to-go to the active waypoint. The active waypoint could be changedto number 5 by pushing "WPT", "5" and "ENT".
The uppermost rotary switch selects the CDI sensitivity desired. Thethree settings are for Enroute. Terminal and Approach, in order of least tomost sensitive. The sequencing switch controls whether or not the activewaypoint will increment automatically as the waypoints are overflown. Asoftware check is implemented such that the pilot must rotate the OBS knob tothe appropriate course before the active waypoint number is incremented. Thebottom rotary switch controls the initial start-up mode of the GPS receiver.In Auto Start mode. the system assumes that the GPS receiver has not movedsubstantially since the last time the GPS set was turned on. In this case.the system uses the position fix stored in non-volatile memory from the lastflight as the assumed initial location of the receiver. In the Manual Startmode, the operator must enter the initial position of the receiver to thenearest degree in latitude and longitude.
The CDU design is based on the Z80 microprocessor, and employs the STDBUS form factor and bus structure.
3.8 Instrumentation
The GPS T and E equipment was equipped with instrumentation to record:
a) receiver channel and position software activity and outputs,
b) received carrier to noise ratios on each satellite,
c) pilot (CDU. CDI. OBS) activity, and
d) aircraft attitude (roll, pitch. heading).
Receiver channel and position software activity is monitored by recordinginter-task messages. Table 3-13 shows the general content of the variousmessages recorded on magnetic tape during a typical flight. Tables 3-14through 3-21 show the message formats frequently used in assessing status andposition estimation performance. Table 3-22 defines the aircraft attitudedata.
68
TABLE 3-11
ACTIVE WAYPOINT DATA FILE
NAME DESCRIPTION ACCURACY KEY UPDATE INTERVAL
PIAT Present Latitude 0.1' POS 2 sec
PLNG Present Longitude 0.1' POS 2 sec
GS Ground Speed 1.0kt GS 2 sec
'IK Actual Track (Magnetic) 1.0 deg GS 2 sec
WPT Active Waypoint NIA WPT 2 sec
DIST Distance to Active Waypoint 0.1 nm RIB 2 sec
BRG Bearing to Active Waypoint 1.0 deg RIB 2 sec
D'IK Desired Track 1.0 deg CRS 2 sec
N'IK Next Desi red Track 1.0 deg CRS 2 sec
TTG Time-to-go 0.1 min TTG 2 sec
XTK Cross-track Deviation 0.01 nmi XTK 2 sec
'IKE Track Angle Error 0.1 deg XTK 2 sec
VAR Magnetic Variation 1.0 deg GMT 2 sec
GMT Greenwich Mean Time 1 sec GMT 1 sec
OBS Current OBS Setting 1 deg OBS 1 sec
CSE Course Selection Error 0.1 deg OBS 1 sec
• TDrST Total distance flownfrom Waypoi nt O. 1.0 nmi TTG 2 sec
69
TABLE 3-12
DISPLAY FIELD BY KEY ENTRY (OPR MODE ONLY)
KEY DISPlAY (Note 1) EXPLANATION------
RIB WPT RNG 19.2 NMI Range to active waypointRIB WPT BRG 207 MAG Bearing to active waypointTTG WPT TTG 12.1 MIN Time to go to active waypointTTG TOT TIME 2:31 HRS Total time in air from waypoint O.XTK XTK DEV 0.02 NMI Cross-track deviationXTK TRK ERR 14.1 DEG Track Angle Error (DTK-TK)CRS CRS TO 205 MAG Course to active waypoint (Note 2)CRS CRS FROM 241 MAG Course from active waypoint (Note 2)GS GS 135 KTS Ground SpeedGS TK 206 MAG Actual TrackGMT GMT 1453: 11 HRS Greenwich Mean TimeGMT MAG VAR 15 DEG Magnetic VariationPOS LAT 42DG 17.0' Present LatitudePOS LONG 71DG 43.6' Present LongitudeOBS OMNIBRG 170 DEG OBS SettingOBS OBS ERR 2.1 DEG Course Selection Error (DTK - OBS)VOR ***INVALID ENTRY Not used in OPR ModeINT ***INVALID ENTRY Not used in OPR ModeLIL WLAT 41DG 23.6' Active waypoint latitudeLIL WLNG 72DG 42.1' Active waypoint longitudeAWP ART WP: GDM VOR Reference ident and typeAWP 111.OMAG 16.0 NMI Bearing and range from reference
I Note 1:
Note 2:
The first display listed for each key appears the first time the key isstruck, (except CRS key, see Note 2.) Subsequent keystrokes cause the twodisplay lines to alternate.The first strike of CRS Key causes "CRS TO" to be displayed if currentdirection is "TO" active waypoint. Otherwise, "CRS FROM" is displayed.Subsequent keystrokes cause lines to alternate.
70
•
AB
CD
EF
Gt
RIB
GS
VO
RW
PT1
23
HI
JK
LM
N
TTG
GM
TIN
TED
T4
58
PRY
0P
QR
ST
UX
TKPO
SL
/LD
LT7
88
NX
T
VW
XY
ZC
LR0
•EN
TC
RS
OB
SA
WP
DEF
MO
DE
Eil
l:0
WPT [!J
TY
PE
~D
IR
~ID
ENT
~
TER
MEN
RT
IA
PPR
'@j
SEQ
UE
NC
ING
MA
NU
AL@
JA
UTO
AU
TOM
AN
UA
LST
AR
TST
AR
T
'@ /"
OFF
TE
ST
FIG
.3
-37
.G
PS
CD
UF
RO
NT
PA
NE
L
TABLE 3-13
DATA RECORDING MESSAGES
Message l 1 IBlockNumber Source I Content Frequency _I
0 DAU System Reset Startup10 DAU Receiver Configuration Data Startup15 NSW PNP-DAU Time Synchronization Startup20 PSW Position Estimate Sent to NSW l/sec25 DAU Operator Startup Data (if defined) Startup30 PSW Raw GPS Satellite Nav Data (1 word) Variable40 PSW Acquisition Command As Required50 PSW Phased Transition Command As Required60 PSW Navigation Command As Required70 PSW Measurement Command Data 1-5/sec80 PSW Satellite Reassignment Command As Required90 PSW Nav Plus Data Command Single Channel
110 PSW Ephemeric, Clock and Almanac Infrequent120 PSW PSW Error Message130 NSW Receiver Startup Data to PSW Startup140 PSW PNP Status to DAU 3/Minute150 PSW Raw Position Fix 1/2.2 sec Typical160 PSW Tracked Position Fix l/sec170 NSW Pilot Flight Plan Waypoint Data As Required
I180 NSW CDU, CDI, OBS Data l/sec190 NSW Pilot Startup Data From CDU Startup200 NSW Displayed Waypoint Data Type When Changed210 NSW OBS Data l/sec220 NSW Displayed Waypoint Data l/sec230 NSW Clock Status Sent to CDUI Startup400 ADU RAS Data l/sec410 ADU DME Data Variable420 ADU ADU Status Variable430 ADU ADU Error Variable
DAU Data Acquisition UnitNSW Navigation SoftwarePSW Position SoftwareADU Aircraft Data UnitRAS Real Time Aircraft State
72
TABLE 3-14
BLOCK 1110 RECEIVER CHANNEL CONFIGURATION DATA
-'-----'---------------
45-67-89-12
1314-1617-1819-2627-28
~Form~- I Un~~~ =tS~_Scale
constant ~ 1constant I :~;d s 1
I binary, unsigned I N/A I 1 I
I I I I
l!Lte I Deft ni t~on
---l-~k ID (=10)----
2 I Block Length (=14)3 I DuallSingle Channel
I= 0 Dual= 3778 Si ngleStatistics Collect (1 = collect)AGC Loop Corner FrequencysNomi. nal AGC Gai nsSearch ThresholdsRe-search Frequency offsetAFC Loop Noise BandwidthsCode Loop Noise BandwidthsAFC Lock Gai nsBit Sync Histogram Threshold
bytebyteintegerintegerbyte
I byteI integerJ byte
N/A 1N/A 1N/A 1Hz 1Hz 1
_____~~_~ l~-l
TABLE 3-15
BLOCK 1130 RAW GPS NAV MESSAGE DATA WORD
constant bytes 1real second sinteger millisecs. 20 ms.see block 1170bi nary, unsigned NIA 16 LSBsIbinary, 3 LSBs NIA 1binary, 4 LSBs N/A 1binary, LSB N/A 1
binary N/A
binary N/A
binary N/Asee block 1170see block 1170
Definition
Subframes IDWord count (1-10) wi thin subframeParity check flag: 0 = data valid
1 = parity failLeast significiant byte of demodulat
ed pari ty stripped data word2nd byte of demoudlated, parity
stripped data wordMost significant byte of demodulated
parity stripped data wordWide Band PowerNarrow Band Power
20
21
22
171819
23-2425-26
~~I
-r=-1f--=-:Bl:-"O-Ck:--ID· (=30) --------
2 Block Length (13)3-10 GPS Time
11-14 Mission Time15 Channel Status16 SV lD
73
TABLE 3-16
BLOCK #40 ACQUISITION COMMAND DATA
4.064Hz
binary, unsigned CiA 0.14 LSB'S code chipbinary, unsigned CiA code 1
chip
bi nary, unsigned, CiA code 2562 LSB's chipbinary ,unsigned, I CiA 1
I6 LSB's epochs(Note 2) (Note 2) (Note 2)binary, unsigned quanta 1binary, unsigned quanta 2561 LSB
L J
1,~~~;;;t ';;:: fSB ;cale I
real second sbinary, UnSignedl milli- I 20 ms.
secondsbinary, unsigned N/A61SB'sbinary, 2' scomplement
Carrier frequency prepositioningestimate (Note 1)
16 - LSByte17 - MSByte
Sub-chip phase code prepositioningestimate (0-9)Chip phase code prepositioningestimate, (0-1022) leastsignificant byteChip phase code preposi tioningestimate, 2 MSB'sEpoch phase code prepositioningestimate (0-19)Search and Acquisition StrategyBit Count (0-299) (Note 3)
23 - LSByte24 - MSByte
18
19
20
16-17
IByte I Defi ni tion
---ttBlOCk ID (=40)2 Block Length (=12)
3-10 GPS Time111-14 Mission Time
15 SV ID Assignment (1-31)
I 21
I22
23-24
IJ'-----'--------------------
74
•
TABLE 3-16. (CONT'D)
Notes:
(1) All prepositioning estimates are valid at the trailing edge of theparticular 20 msec system clock pulse defining the beginning ofacquisition of an SV.
(2) Search and Acquisition Strategy Byte, contains:
- Bit 7 (MSB)is DATA DEMOD FLAG0: channel to demodulate data1: channel not to demodulate data
- Bit 6 is SYNC FLAG0: channel to perform Bit/Subframe Sync operations1: channel to use PNP given BIT COUNT (Bytes 10 and 11)
to drive Subframe Sync
(DATA DEMOD FLAG - 1 SYNC FLAG - 0 is illegal combination)
- Bit 5-0 is HALF SEARCH APERTURE CODE (HSAC)(format: binary, unsigned) related to half search aperture asfollows
HSAC-0-
12
63
Half Search Aperture (chips)8
1624
512
Half Search Aperture (chips)i.e., HSAC = ---------------------------- - 1
8
(3) Bit Count is valid only if SYNC FLAG = 1
75
TABLE 3-17
BLOCK 1170 MEASURE.MENT DATA
1
1
0.1
256
6.35mHz
binary, unsigned, ICIA4 LSB's code chips
binary, unsigned CiAcode chipsCiA
code chipsCiA
code epochunsigned quanta
binary, unsigned,2 LSB's
Ibinary , 2'comple
mentIbinary ,
bi nary, unsigned,6 LSB's
binary 2'compleImentI
bi nary, unsigned quanta
sub- binary, unsigned, N/A 14 LSB
sub- binary, unsigned, N/A 15 LSB
_______T_Fo~mat - F_ni_ts rBScale II
--~nstant N/A 1
Iconstant I word s 1real second sbinary, unsigned I milli- 20 ms.
secondsN/A
least significantdwell ending (0-1022)2 MSB's
Least significant byte of receiverchannel estimated carrier frequency(Note 2)
Second byte of receiverchannel estimated carrier frequency
Most significant byte of receiverchannel estimated carrier frequency
Sub-chip code phase, valid at dwellending (0-31)
Chip code phase,byte. Valid at
Chip code phase,
Epoch code phase valid at dwellending (-127 to +128) (Note 3)
Wideband Power24 - LSByte2S - MSByte
Narrow Band Power26 - LSByte27 - MSByte
Data word count (0-9) withinframe at dwell endi ng
Data bit count (0-29) withinframe at dwell ending
28
18
21
17
19
22
23
20
29
26-27
24-25
IByte r-----nerrni tion
I 1 Block 10 (=70)
I2 I Block Length (=15)
3-10 GPS Time
1
11-14 I Mission Time15 Channel Status (Note 2)16 SV 10 (1-31)
76
TABLE 3-17. (CONT'D)
Notes:(1) During navigation mode, this data is valid at the leading edge of
the particular 20 msec system clock pulse that defines satellitetransition.
(2) Channel status byte
17 6 5 4 I 3 2 1 CQI*Bit 0Bit 1
**Bit 2Bit 3Bits 4-7
Set when receiver fault sensedInit Mode or Configuration Mode Complete, Carrier Lock,
or Measurement ValidFrame Synch Complete/In Data Demod1 Pass of Search CompletedReceiver State (coded)
o = Idle1 = In Limited Search (± 15 chip)2 = In ± 2 chip Search3 = In ± 1 chip Search4 = In False Alarm Check5 In Re-search6 = In AFC Pull-in #1 (Pre Bit Sync)7 = In AFC Pull-in #2 (Pre Bit Sync)8 = AFC Lock9 = In AFC Pull-in #1 (Post Bit Sync)
10 = In AFC Pull-in #2 (Post Bit Sync)11 = Fine-Time12 = Bit Synch Operations13 = Frame Synch Operations14 = Data Demodulation15 - Statistics Done
(3) Pseudo-range can rollover a bit edge (overflow or underflow),therefore it can be larger than 2Omsec, or it can be negative;referenced to the initial position value.
(4) Wide band (WBP) and narrowband (NBP) power quantities are used inthe following equation to compute the G/No estimate:
C/No = 10 loglO 2000(NBP - WBP)
(40 WBP - NBP)dB Hz
Note: 6 dB must be added to C/No estimate if data is not beingdemodulated.
*Pseudo-Range is valid if Bit 1 is set in Acquisition, Transition orNavigation Phase. Bit 1 is set when mode is complete when inInitialization or Configuration Modes.
**Bit 2 is set when frame synch is achieved and data is beingdemodulated, unless Frame Sync is not required. Then it indicatesdata demodulation.
77
TABL
E3-
18
BLO
CK#1
50RA
WPO
SIT
ION
FIX
DATA
By
teD
efi
nit
ion
For
mat
Un
its
ILSB
Sca
leI
II
1B
lock
ID(=
150)
con
stan
tN
/A1
2B
lock
Len
gth
(=60
)co
nst
ant
wor
dsI
13
-10
GPS
Tim
ed
ou
ble
real
seco
nd
s11
-14
Mis
sio
nT
ime
inte
ger
mil
li-I
20m
s.se
con
ds
15
-22
Po
siti
on
(EC
EF)
Xd
ou
ble
real
feet
I23
-30
Yd
ou
ble
real
feet
31-3
8Z
do
ub
lere
al
feet
39
-42
Vel
oci
ty(E
CE
F)X
-dot
real
feet/
sec
44
-46
Y-d
ot
real
feet/
sec
"-J
47
-50
Z-d
ot
real
feet/
sec
00
51
-58
Clo
ckp
has
ed
ou
ble
real
seco
nd
s5
9-6
6C
lock
freq
uen
cyd
ou
ble
real
sec/s
ec
67B
aro
Alt
imet
erF
lag
(100
0O
OO
O=u
sed)
inte
ger
68S
par
e69
-72
Bar
oA
ltim
eter
Mea
sure
men
tre
al
feet
73-7
6C
ompu
ted
Alt
itu
de
real
feet
77-8
8U
nit
vec
tor
for
Bar
oan
dP
seu
do
-Bar
ore
al
mea
sure
men
ts(3
elem
ents
)8
9-9
0P
seud
o-B
aro
Fla
g(1
000
0000
=u
sed
)in
teg
er
91
-94
Po
siti
on
fix
(EC
EF)
6Xre
al
feet
95-9
86Y
real
feet
99-1
026Z
real
feet
103-
106
6cl
ock
real
seco
nd
s10
7-11
6D
iag
on
alel
emen
tso
ffi
xm
atri
x(4
elem
ents
)re
al
117-
120
Spa
re
TABL
E3-
19
BLO
CK#1
60TR
ACK
EDPO
SIT
ION
FIX
DATA
By
teI
Defi
nit
ion
IF
orm
atI
Un
its
ILSB
Sca
le
1B
lock
ID(=
160)
Ico
nst
ant
-1
2B
lock
len
gth
(=63
)co
nst
ant
wor
ds1
II3
-10
GPS
Tim
eI
do
ub
lere
al
Ise
con
ds
I-
I11
-14
Mis
sio
nT
ime
inte
ger
mil
l1se
c.
-15
-22
Po
siti
on
(EC
EF)
XI
do
ub
lere
al
Ife
et
-23
-30
YI
do
ub
lere
al
Ife
et
-31
-38
Zd
ou
ble
real
feet
-39
-42
Vel
oci
ty(E
CE
F)X
-do
tI
real
feet/
sec
-I
43-4
6Y
-do
tI
real
feet/
sec
-I
47-5
0Z
-do
tre
al
feet/
sec
-51
-58
Po
siti
on
(Geo
det
ic)
lati
tud
eI
do
ub
lere
al
rad
ian
s-
59-6
6lo
ng
itu
de
do
ub
lere
al
rad
ian
s-
67-7
4alt
itu
de
do
ub
lere
al
feet
-75
-78
Vel
oci
ty(L
ocal-
lev
el)
No
rth
real
feet/
sec
-79
-82
IE
ast
Ire
al
feet/
sec
-83
-86
Vert
ical
real
feet/
sec
-8
7-9
4cl
ock
phas
eo
ffse
td
ou
ble
real
seco
nd
s-
95-1
02cl
ock
freq
uen
cyo
ffse
td
ou
ble
real
sec/s
ec
-10
3-11
0ti
me
oftr
ack
er
up
dat
ed
ou
ble
real
seco
nd
s-
111-
114
Tra
cker
up
dat
e(E
CE
F)AX
real
feet
-11
5-11
8AY
real
feet
-11
9-12
2AZ
real
feet
-I12
3-12
61A_
C_l_O
_Ck
L--_
_r_e
_a_l
----L
_s_e
_c_o
_n_d
_s_L
I_
TABLE 3-20
ACTIVE WAYPOINT DATA BLOCK #200 (FROM CDU)
II constant
I constant words
Ipositive integer Imilliseconds II
WORD DEFINITION
II 1 IBlock ID (= 200)
I 2 IBIOCk Length (=14)
I 3-4 IMission TimeI
\Active5(bits 8-15) Waypoint number(0 to 19 )
5(bits 2.1.0) CDI Sensitivity
I 0 = Enroute1 = Terminal2 = Approach
6-9 Active Waypoint Latitude
10-13 Active Waypoint Longitude
14(bits 0-5) RNAV Parameter Number(11 through 32)
11 = Range-to-Waypoint12 = Bearing-to-Waypoint13 Time-to-go14 Not used15 Cross-track deviation16 Track angle error17 Course to waypoint18 Course from waypoint19 Groundspeed20 Track angle21 Greenwich Mean-Time22 Not used23 Present latitude24 Present Longitude25 OBS setting26 Not used27 = Artificial Waypoint
reference type28 = Artificial waypoint
reference ident.29 = Artificial waypoint
offset range30 = Artificial waypoint
offset bearing31 = Waypoint latitude32 = Waypoint longitude
80
FORMAT
integer
integer
double real
double real
integer
UNITS
radians
radians
LSB SCALE
1
1
20 ros.
1
1
1
TABLE 3-21
AREA NAVIGATION DATA BLOCK #220 (TO CDU)
WORD DEFINITION FORMAT UNITS LSB SCALE
o = okay1 = not OK
constant1
2
3-4
5(bit 8)
5(bits 1,0)
IBlock In (=220)
IBlOCk length (=20)
IMission Time
INAV flag
To/From flag 0 = from1 = to2 = neither
I II constant I16-bit wordS/
Ipositive integer/milliSeCondsl
I integer I
integer
1
1
20 ms.
6(bits10,9,8) Solution Quality 0 OPR1 = UNR2 = LCH3 BAD4 STR
6 (bits 0+5) RNAV Parameter No. (see block#200)
7-8 RNAV Parameter
9-10 lOBS Setting
11-12 CDI Deflection
13-16 Present Latitude
17-20 /present Longitude
I
81
integer
integer
real
real radians
real dots
double real radians
double real radians
1
1
TABLE 3-22
REAL-TIME AIRCRAFT STATE DATA
Value of RangeParameter LSB Min Max
Heatiing (360°/512) 0 360° (binary 512)
Roll 1° -31° (binary 0) +32° (binary 63)
Pitch 1° -31° (binary 63) +32° (binary 0)
Outside airTemperature 0.625°C -40°C (binary 0 +39.375°C (binary 127)
IndicatedAir Speed 1 rom H2O 0 (binary 0) 1023 (binary 1023)
Rate ofClimb 24 mm H2O/ min -768 (binary 0) +744 (binary 63)
Note: Pitch reading is not linear for pitch angles greater than 20 degrees.
82
The receiver carrier-to-noise-density ratio (C/No ) is estimated using thewide-band energy detected by the AGC loop and the narrow-band detector shownin Fig. 3-14. Scaled estimates are recorded for each measurement in blocks 30and 70. C/No is then calculated using the formula in NOte (4), Table 3-17.
Aircraft altitude is measured using standard aircraft instruments whichhave been connected via a digital interface to a Z80-based Aircraft Data Unitassembly (ADU). The ADU in turn communicates via a serial RS232 interface tothe DAU, nominally once per second. The altitude parameters are defined inTable 3-19.
Pilot activity is monitored by recording request and response messagesbetween the DAU and CDU using the formats shown in Tables 3-20 and 3-21.
83
4.0 MEASURED PERFORMANCE
4.1 Test Facilities
4.1.1 Ground Test Laboratory
During integration of the GPS T and E system a tower-mounted antenna wasemployed. Initial evaluation of the hardware and software functions werecompleted in the laboratory before flight tests commenced.
4.1.2 Aircraft Installtion
The GPS T and E equipment was installed in a Rockwell Aerocommander SOOA,as shown in Figs. 4-1 through 4-4. The T and E equipment was configured asshown in Figs. 4-5 and 4-6. The CDU and CDr are installed as shown in Fig.4-7. The traditional difficulty of pushing buttons on a vertically mountedcockpit device was relieved for the CDU because the engine controls provide aconvenient stabilizing hand rest for most power settings. The performanceenvelope for the Aerocommander is shown in Table 4-1.
4.1.3 Mode S Experimental Facility
Position truth was established using the Mode S Experimental Facility(MODSEF) operated at Lincoln Laboratory. It includes a monopulse surveillancesystem capable of interrogating ATCRBS and Mode S Transponders. Its principalcharacteristics are provided in Table 4-2.
During GPS tests Mode S Experimental Facility surveillance accuracy wasverified using a calibration transponder located 6 miles from the Mode SExperimental Facility. The aircrafts altimeter calibration was also verified.Since most of the Hanscom runways are visible to the Mode S ExperimentalFacility sensor, a test was conducted in which the aircraft parked at a taxiway reference point, to verify surveillance range/altitude.
The Mode S Experimental Facility surveillance data for the GPS testaircraft was recorded and identified based on either an ATCRBS discrete codeor Mode S ID (both transponders were available on the Aerocommander).
84
00
V1
•
ex>
0\
1C
P13
0-95
6I
OP
SA
NT
EN
NA
FIG
.4
-2.
GP
SA
NT
EN
NA
INS
TA
LL
AT
ION
.'
00 ........
PA
SS
EN
GE
RC
OM
PA
RT
ME
NT
r-----
BA
GG
AG
EC
OM
PA
RT
ME
NT
PIL
OT
CO
MP
AR
TM
EN
T
FIG
.4
-3.
GP
ST
ES
TA
IRC
RA
FT
INT
ER
NA
LLA
YO
UT
----
WIN
DO
W
FIG
.4
-4.
GP
SF
LIG
HT
SY
ST
EM
INP
AS
SE
NG
ER
CO
MP
AR
TM
EN
T
OF
AE
RO
CO
MM
AN
DE
RA
IRC
RA
FT
DA
UT
ER
MIN
AL
FIG
.4
-5.
INS
TR
UM
EN
TA
TlO
NR
AC
K
\0 o
FIG
.4
-6.
RE
CE
IVE
RR
AC
K
•
P130
-168
3
FIG
.4
-7.
AE
RO
CO
MM
AN
DE
RN
33
3B
E.
CO
CK
PIT
VIE
W
TABLE 4-1
AEROCOMMANDER FLIGHT CHARACTERISTICS
Crui se Speed
Ceiling
Time at 10 Kft.
Assumptions
2 Pilots1 Operator
160 KIAS
10 Kft.
1.5 Hours
92
160 KIAS
10 Kft.
2.5 Hours
..
TABLE 4-2
MODE S EXPERIMENTAL FACILITY SURVEILLANCE CHARACTERISTICS
Antenna
Transmit
Receiver
Nominal Range
Surveillance AccuracyRangeAzimuth
Calibration
I Display
4° monopulse
150 watts with SLS
Phase monopulse10 dB noise figureRSLS
75 nmiles
40 feet 100.05 deg 10
Test transponder at surveyed site6 miles north of The Mode SExperimental Facility
TPX 42 display
93
4.1.4 Analysis Software
The data recorded on the aircraft, and at the Mode S ExperimentalFacility were processed as shown in Fig. 4-8 to produce a basic surveillancedata base and error statistics.
The difference between the GPS and Mode S Experimental Facility positionestimates was determined by using every ungarbled Mode S Experimental Facilityreport (unsmoothed, occurring at 4 second rate), and the two adjacent GPSposition estimates as shown in Fig. 4-9. Conversion from the Mode SExperimental Facility R, a, H position estimate to the aircraft's XYZ in earthcentered earth fixed ECEF coordinates was done using matrix transformations*which account for the WGS-72 ellipsoid of revolution earth model.
Plots of XY versus time, GPS position error with respect to the Mode SExperimental Facility versus time, and error statistics were developed.
GDOP was independently estimated using Z STARS, a Lincoln program whichprovides visibility to current satellites. Using Z STARS, GDOP was calculatedbased on all satellites actually in track.
4.2 Laboratory Tests
4.2.1. Antenna
The Chu Associates, Inc., model CA-3224 volute antenna was tested whilemounted on a 5-foot by 5-foot aluminum plate shaped to simulate the local AeroCommander fuselage as shown in Fig. 4-10. Anechoic chamber azimuth gainmeasurements were made at various elevation angles. The results, shown inFigs. 4-11, -12 and -13, show that the specification was met or exceeded overmost elevation angles except within the range 5-10° where the gain was 2-dBlower than specified. Measurements above 50° elevation were not possible dueto mechanical conflicts between the fuselage plate and the anechoic chambermount. It is assumed that a gain of 0 ± 1 dBi in the region SOo-90° has beenprovided.
The elevation gain profile, Fig. 4-14, was developed by computing themean azimuth gain at each elevation angle from samples taken every 10° inazimuth. The standard deviations in azimuth gain were also computed, as shownby the dotted lines.
* See Appendix C for details.
94
~IU
RV
INTERPOL~TE
MO
DE
ID
AT
AA
Z,
R.
AL
T.
_A
ND
LO
CA
LX
,Y
.Z
E.F
.fI
LT
ER
CO
OR
D
1C
ON
Y
PO
IIT
ION
ER
RO
RE
RR
OR
DA
TAT
IUE
AN
AL
VII
I.
LO
CA
LX
.Y
.ZT
EC
E'
-C
OO
RD
~IL
OC
KC
0t4V
'.0
LA
TL
ON
G.
-.
PI
fLIG
HT~
ILO
OK
AIR
CR
An
AT
TIT
UD
ET
APE
40
0--.
PL
OT
SO
FT
WA
RE
ILO
CK
CD
U-
CD
I-O
BI
11
0-...
CIN
o
ILO
CK
-...L
..--
t7O
TR
AC
KE
D_
GD
OP
GD
OP
(AL
LIN
TR
K)
lAT
.L
D:.
IEL
ICT
CA
LC
VII
IIL
I
TIM
E--Ii
GP
IIA
T:I
(LO
OA
LA
Z.
EL
)E
CE
'lA
T.
LD
:.~
ZIT
AR
I-
CO
OR
DG
DO
PG
DO
P(A
LL
INV
IEW
)-
DA
BIE
'C
OO
RD
--Ii
CO
NV
CA
LC
G
FIG
.4
-8.
PO
ST
FLI
GH
TA
NA
LYS
ISS
OF
TW
AR
E
oG
PS
PO
SIT
ION
ES
TIM
AT
ES
@M
OD
ES
PO
SIT
ION
ES
TIM
AT
ES
TR
UT
H
oo
MO
DE
SR
EP
OR
TA
T'I
o
LIN
EA
RIN
TE
RP
OLA
TIO
NT
OE
ST
IMA
TE
GP
SP
OS
ITIO
NA
T'J
1......-'..-
..-...-"1
1-1
".,,5"
".,
..."
,..
".,
\ \ @P
OS
ITIO
NE
RR
OR
FIG
.4
-9.
GP
SP
OS
ITIO
NE
ST
IMA
TIO
NE
RR
OR
AL
GO
RIT
HM
GPS ANTENNA
110·L 10'
+ /
I1-.-28,·-_1
f------.. --'11'·---
ELEVATION VIEW
ff-~-~8"-'1
2~·
I IIIIIII I
II
O· II / GPS
I _~ANTENNA
I I
II
25' III I
I
II I
ICURVED IREGION
,
(TYP)
PLAN VIEW
FIG. 4-10. CURVED GROUNDPLANE
97
LEFT WING
LEFT WING
NOSE
O· EL
o dBi
I----- + -----------t---<>\--+----!-<---+--
I
TAL
NOSE
OdBI
- SOEL
------+\+--+---f---+--+----- RIGHT WING
TAL
FIG. 4-11. Gain VB. Azimuth for Elevation Cuts at -5° and 0°;Chu Model CA-3224 Volute Antenna.
98
NOSE
LEFT WtlG
17' EL
OdBI
..---It----------- + --------+-+-+-+-f-;-- RIGHT WtlG
TAL
NOSE
o dBi
8' EL
....LEFT WtlG --1----,1----------+--------t-t--i--\t-+-+- RIGHT WtlG
I
TAL
FIG. 4-12. Gain vs. Azimuth for Elevation Cuts at +6° and +17°;Chu Model CA-3224 Volute Antenna.
99
NOSE
LEFT WING ---11---------
SO· EL
o dBl
-e -4 -2 .2 +.------+-4-+-+t-+-+-- RIGHT WING
TAL
NOSE
o dBi
30· EL
LEFT WING --t-+----------+ -------f--+--+-Hr-t-- RIGHT WING
I
TAL
FIG. 4-13. Gain VB. Azimuth for Elevation Cuts at +30 0 and +50 0;
Chu Model CA-3224 Volute Antenna.
100
20'
GAIN (MEASUREMENTS AT 38 AZIMUTHS AVERAGED)
SPREAD (2 0') OF MEASURED GAINS
GAIN REQUIREMENTS (SEE GPS QTL .. -5)
ELEVAnON ANGLES AT WHICH 38 GAINMEASUREMENTS WERE MADE
IL- __ ---!4!!.. _
;t-- _./ ......'" r-- ...-/,. -......... .... ..........
~ ~/ .... -........ --- ....I , MIN - ........... _~ ......... -...
I :----l - -- ... ------I '
/ "I I
/ !I I
I /I I
/ I
I "/ II
I/,
.....MA~.··
+2
0
i~ -2<C!J<z -4zw~z<
-8
-8
A
-10 -5 0 5 10 20 30 40 50 80 80 80
ELEVATION ANGLE (DEGS)
FIG. 4-14. VOLUTE ANTENNA GAIN, ELEVATION PROFILE
101
4.2.2 Receiver Channel Performance
The receiver factory acceptance tests, conducted using the equipmentconfiguration shown in Fig. 4-15, verified compliance with all receiverspecifications including:
a) Receiver Hardware - PNP Software Interface.b) Acquisition time versus doppler and C/No •c) Data error rate versus C/No'd) Transition performance versus C/No 'e) Navigation performance versus e/No'f) Nav plus Data performance versus e/No 'g) AFe error performance.h) Pseudo-range accuracy.i) CINo measurement accuracy.
During the acceptance tests, acquisition of all satellite codes now inuse with doppler offsets up to > 1000 Hz (> 750 Hz requirement) wasdemonstrated. Tests of the Transition, Navigation and Navigation plus Datamodes were accomplished by phase-locking the receiver oscillator to thesatellite simulator oscillator and manually entering prepositioning code anddoppler data into the PNP simulator. CPU activity monitoring verified thatthe receiver Z8000 microcomputer task execution times did not exceed theirallotted times in any mode. Finally, all visible satellites were successfullyacquired using a roof antenna. All tests were conducted on both channels.
Following shipment of the receiver to Lincoln Laboratory, the acceptancetests were repeated in a laboratory using the same test configuration as inFig. 4-15 except that a volute antenna identical to that installed on theAerocommander was used.
Table 4-3 summarizes the major requirements and performance measuredduring the factory acceptance tests and during the post-acceptance evaluation.Results of specific measurements are shown in Figs. 4-16 through 4-24, and theresults of a test in which a GPS satellite was tracked for 5 hours are shownin Fig. 4-25.
The measured receiver performance was also evaluated with respect totheoretical predictions. The results, shown in Figs. 4-26, -27, and inTable 4-4, show good agreement.
To assess the 5-hour tracking test data, the expected receive C/NO wascalculated as shown in Table 4-5. Worst-case RF power (CiA code only) at theuser antenna is seen to be -160 dBw. Note that the receiver preamp is assumedto be located at the antenna for minimum cabling loss.
102
PN
PS
IMU
LA
TO
R
I I I I I I L--
l
I II
ST
ID
UA
LC
HA
NN
EL
RE
CE
IVE
RI
1--.1
,---
----
--,
II
II
CH
AN
NE
L1
~I
I~
II
IR
S2
32
I6
80
04
I-
I
WI
J.lP
RO
CE
SS
OR
I.....
...I
CH
AN
NE
L2
II
SIN
GL
EC
HA
NN
EL
IS
AT
EL
LIT
EI---
II
SIM
UL
AT
OR
II
II
II
I5
0H
zC
LO
CK
II
"----
CL
OC
KI
TE
RM
INA
L
I
RO
OF
TO
PA
NT
EN
NA
FIG
.4
-15
.D
UA
LC
HA
NN
EL
RE
CE
IVE
RA
CC
EP
TA
NC
ET
ES
TC
ON
FIG
UR
AT
ION
TABLE 4-3
RECEIVER HARDWARE PERFORMANCE
Parameter
Acqui si tion
ITime to acquire
Requirement
25 sec(based on 6 min TTFF)
Measured Performance
25 sec. P >.9
AFC errors
Bi t Error Rate
~f < 10 Hz for datademodulation
for GINo) 36 dBHZI
C/No Error
Phased Transition
AFC Error
Pseudo Range Error
C/No Error
Navigation
± 1 dB
Not specified
Not specified
± 1 dB
<1C/No .. 0.7 dB
< 5 Hz
< 45 ns for C/No) 36 dBHz
<1C/No < .85 dB
AFC Error
Pseudo Range Error
Not specified
± 1 dB <1C /No < 1.0 dB
for C/No I) 36 dBHz I
Nav Plus Data
Same as Navigation Same as Navigation Same as Navigationand Acquisition and Acquisition and Acquisition
Acqui si tion 4.5 dB at 50 elevation.Link Margin* At least 4 dB 10 dB at 400 elevation,
for the volute antenna.
*Assuming 4 dB noise figure. See Table 4-5.
104
•
12
0
\N
OD
AT
A, \
\
10
0\ \ \
80
() w en
.......
wC
H2
0~
\Jl
i=C
H1
06
0() <
40
20
34
36
38
40
42
44
46
RE
CE
IVE
RIN
PU
TC
/No
dB
Hz
FIG
.4
-16
.A
CQ
UIS
ITIO
NT
IME
SO
FA
SIN
GLE
SA
TE
LL
ITE
48
50
C·+
2-
21
8,q
52
1.4
1.2
1.0
N :r:f-
'z
0 0'\
-b 0.8
0.6
0.4
52
C-+
2-2
18
0
38
40
42
44
46
48
RE
CE
IVE
RIN
PU
TC
/No
dB
Hz
FIG
.4
-17
.R
EC
EIV
ER
AF
CP
ER
FO
RM
AN
CE
INA
CQ
UIS
ITIO
NP
HA
SE
36
34
1.0 0.8
m "C
t-'
0o
Z-..
..I..
..
:;0
.6 0.4
0.2
r42
-21
90
oL--...J3L4----.l--3..l..6--L---....l...------l..--,J----L-----J4L2------.l.--4-:1.4:---...l.--~-----J--__::1=_----'--____:~----I....-----J S2
RE
CE
IVE
RIN
PU
TC
lNo
dB
Hz
FIG
.4
-18
.R
EC
EIV
ER
ClN
oE
ST
IMA
TO
RP
ER
FO
RM
AN
CE
INA
CQ
UIS
ITIO
NP
HA
SE
70
60
50
40
30
20 3
43
6I
II
II
I
38
40
42
44
46
48
50
RE
CE
IVE
RIN
PU
TC
/No
dB
Hz
FIG
.4
-19
.R
EC
EIV
ER
PS
EU
DO
RA
NG
EP
ER
FO
RM
AN
CE
INP
HA
SE
DT
RA
NS
ITIO
N
52
6 5 4
N :I:
~~
0-
\Db
3-
2 1
34
36
38
40
42
44
46
48
RE
CE
IVE
RIN
PU
TC
INo
dBH
zFI
G.
4-2
0.
RE
CE
IVE
RA
FC
PE
RF
OR
MA
NC
EIN
PH
AS
ED
TR
AN
SIT
ION
50
52
C4
2-2
13
6
1.2
1.0 0.8
m "0
~0
oZ 0
0.6
b
0.4
0.2 3
43
63
84
04
24
44
64
85
0R
EC
EIV
ER
INP
UT
CIN
od
BH
z
FIG
.4
-21
.R
EC
EIV
ER
CIN
oE
ST
IMA
TO
RP
ER
FO
RM
AN
CE
INP
HA
SE
DT
RA
NS
ITIO
N
52
70
60 50
CH
2
0 q)
CD c: ..4
0b
~ ~ ~
30
20
•
34
36
38
40
42
44
46
48
50
52
RE
CE
IVE
RIN
PU
TC
/No
dB
Hz
FIG
.4
-22
.R
EC
EIV
ER
PS
EU
DO
RA
NG
EP
ER
FO
RM
AN
CE
INT
HE
NA
VIG
AT
ION
ST
AT
E
C4
2-
2H
:5
6 5 4
N :I:
Z -b 3
'""' '""'N
2 1
34
36
38
40
42
44
46
48
RE
CE
IVE
RIN
ClN
odB
Hz
FIG
.4
-23
.R
EC
EIV
ER
AF
CP
ER
FO
RM
AN
CE
INT
HE
NA
VIG
AT
ION
ST
AT
E
50
52
C:~
2-
21R
:i
1.2 1.0
0.8
m "0 0 Z ..... 0 b0
.6
t-'
t-'
w
0.4 0.2
•
II
II
I
34
36
38
40
42
44
46
48
50
RE
CE
IVE
RIN
PU
TC
/No
dB
Hz
F~G.
4-2
4.
RE
CE
IVE
RC
/No
ES
TIM
AT
OR
PE
RF
OR
MA
NC
EIN
NA
VIG
AT
ION
ST
AT
E
52
C4
2-2
18
4
N :E:
Z o z .... o 52
50
48
46
44
42
40
38
--.
..........
....-......
. ........,
........
..
'""-
"-"-
" "" "
"
C/N
o}
AV
ER
AG
EO
F6
00
ME
AS
UR
ED
--
--
DO
PP
LE
RS
AM
PL
ES
EV
ER
Y6
MIN
UT
ES
AV
ER
AG
EC
/No
DU
RIN
G1
0:1
6-
2:2
0=
47
.8d
BH
z
VA
RIA
TIO
NIN
C/N
oD
UR
ING
10
:15
-2
:20
-1.8
dB
(1u
)
A=
EX
PE
CT
ED
C/N
oS
EE
TA
BL
E4
-e.
40
00
30
00
20
00
N1
00
0:E
:
II:
W
o~ Q
.. o-1
00
00
-20
00
-30
00
-40
00
36
20
.00
33
.20
46
.50
58
.10
62
.90
56
.70
44
.\0
30
.00
16
.00
3.1
0
EL
EV
·I
I,
I1
0:3
01
1:0
01
2:0
01
3:0
0\4
:00
15
:00
ES
T1
7N
OV
19
81
SA
T#e
(11
06
4)
10
:00
34
L-
....
L..
....
.I-
....
....
......
...L
-..
....
....
....
..--I.
---J
L--_
_~
7.7
0
FIG
.4
-25
.R
EC
EIV
ER
PE
RF
OR
MA
NC
ED
UR
ING
SA
TE
LL
ITE
PA
SS
•
8 5 4 3 2 1
36
38
40
42
NA
VIG
AT
ION
ST
AT
E
44
..
BIF
-2
00
Hz
SA
FC
-1
0H
z
SY
ST
EM
LO
SS
-2
dB
46
50
ClN
o-
dB
Hz
FIG
.4
-28
.C
OM
PA
RIS
ON
OF
TH
EO
RE
TIC
AL
AN
DM
EA
SU
RE
DR
EC
EIV
ER
FR
EQ
UE
NC
YT
RA
CK
ING
PE
RF
OR
MA
NC
E
50 40
()
30
w CI) c::: ... ~
..... ..... 0\
20
10
NA
VIG
AT
ION
ST
AT
EB
IF•
20
0H
z
BC
l=
10
Hz
SY
ST
EM
lOS
SE
S=
2d
B
36
38
40
42
C/N
o-
dB
Hz
44
46
48
50
FIG
.4
-27
.C
OM
PA
RIS
ON
OF
TH
EO
RE
TIC
AL
AN
DM
EA
SU
RE
DR
EC
EIV
ER
CO
DE
LO
OP
PE
RF
OR
MA
NC
E
....
TABLE 4-4
COMPARISON OF THEORETICAL AND MEASURED RECEIVER PERFORMANCE
Characteristic Predicted Measured
•
I =-=- - I IMinimum Acquisition C/No 35.7 dB Hz 35 dB HzBIF - 1000 Hz
BL - 10 HzL - 2 dB
Nav Loss of Lock 31.5 dB Hz 33 dB HzBIF - 200 Hz
BL - 10 HzL - 2 dB
TABLE 4-5
USER RECEIVED POWER FOR CIA CODE ONLY
Satellite I Satellite IOverhead (at at Elevation
Zenith) Angle of 5° -Satellite TransmitterPower (dBw) 14.25 14.25
RF Loss (dB) 1.0 1.0Antenna Polarization
L08s (dB) 0.25 0.25Antenna Gain 15.0 12.0
-Satellite ERP (dBw) 28.0 25.0Path Loss (dB) 182.5 184.2Atmospheric Absorption (dB) - 0.8
Power atUser Antenna (dBw) -154.5 -160.0
Receiver noise (dBw(4 dB noise Figure) -200 -200
Receiver C/No (dBHz) 45.5 40.0
117
It is evident from Fig. 4-25 that the measured CINo values were greaterthan expected. The mean CINo for elevational angles greater than 15° was47.8 dB-Hz. 2.3 dB greater than the 45.5 dB-Hz prediction for a zenith aspectIt was postulated that the additional power was due to excess satellitetransmit power. excess receiver antenna gain, and excess sensitivity in thereceiver preamplifier.
To verify this a calculation of the expected received CINo and linkmargin referred to the preamplifier input was made. as shown in Table 4-6, andplotted in Fig. 4-25. It is apparent that for satellite number 6 the link hasabout 2 dB additional margin, due probably to excess satellite radiated power.Also, the variations in CINo were examined and found to be generallyconsistent with the ripple in the volute antenna patterns.
4.2.3 Static Acquisition Performance
Initial laboratory tests using the tower mounted volute antenna indicatedthat the T and E Equipment meets the Time to First Fix Requirements of6 minutes maximum when the receiver is initialized with current almanac data.
The sequence of events in a representative static acquisition is shown inTable 4-7.
4.2.4 Static Position Accuracy and Stability
Tests were conducted to characterize the static accuracy and stabilityperformance. Test data, obtained using the tower-mounted antenna, wascollected and analyzed. Results from a representative 26-minute testconducted on 19 November 1982 using five GPS satellites were:
a. RMS horizontal error: 93 feet
b. RMS vertical error: 104 feet
A scatter-plot of the static test results is shown in Fig. 4-28.
A static test was also conducted on 9 May 1983 using a later version ofthe GPS position software. This version of the position software(Version 3.0) included the capability of navigating with three GPS satellitesand the baro-altimeter. The RMS horizontal error for this run. lasting63 minutes, was 95 feet, nearly the same as the five-satellite case. Althoughthis version of the software was not flight tested, the basic capability forthree-satellite navigation was demonstrated.
118
•
TABLE 4-6
EXPECTED LINK MARGIN
Correction Correction I Link Margin Ifor Actual for Measured Net C/No for 35 dBz
Elevation Minimum user Atmospheric EDCR Volute (3 dB Noise I Acquisition IAngle Recei ve Power1 Loss2• Antenna Gain3 Figure)4 Threshold
(degrees) (dBw) (dB) _(dB) (db Hz) (db) ~_=
5 -160.0 +1.50 -2.0 40.5 5.5
10 -159.7 +1.74 0 43.0 8.0
20 -159.0 +1.86 +1.9 45.8 10.8
40 -158.0 +1.92 +1.2 46.0 11.0
60 -158.8 +1.94 +0.4 (est) 44.5 9.5
90 -160.0 +1.95 -0.4 (est) 42.5 7.5
1. From Rockwell ICD-GPS-200, May 81. (Assumes 2 dB atmospheric loss).2. Digital Communications by Satellite, Spilker Page 171.3. GpS-QTL-4-8, page 10.4. The nominal noise figure for the EDCR is 4 dB. The actual noise figure present
during the characterization measurement was 3 dB.
119
TABLE 4-7
STATIC ACQUISITION EXAMPLE
TimeEvent (Minutes : Seconds)
•Start 0
Satellite 4 acquired 0:32
Satellite 6 acquired 1:55
Satellite 5 acquired 2:49
Satellite 8 acquired 3:56
First Internal Fix 4:00
Transition Mode Started 5:00
Fix Sent to Navigation Software 5:05
Navigation Mode Started 5: 11
120
..
GP
SS
TA
TIC
TE
ST
19
NO
V8
2
17
47
-1
80
3G
MT
FIV
ES
AT
EL
LIT
ES
VIS
IBL
E
ER
RO
RM
EA
NS
TD
DE
VR
MS
18
0H
OR
IZ
VE
RT
89
FT
-50
FT
27
FT
92
FT
93
FT
10
4F
T
.. ..
.'. .
'. . #
. .. .
'. .
.. .
...
..
..
t'. . J
'
.'.
. ..
.....
....:
....
.."."
....
.."-
....
'".
....
..".
.t·
•...
:..~:
..~.
.~
..""
.:.....
"....
....
...
••~
-.••
Y••
:..
.."
.."..
...
···.·.
:)·.It
·.·.
.."".~.".....
.."
.".'..:
..'..
...
....
•..
..c::.
.:-...
..•.
......
~.
_...
....
....
..'
..."..
..4
0
12
0
f-'
N.....
f-'
.... C (J 0 .... ~ l-8
0w w u..
-
o-4
00
.04
08
01
20
FE
ET
(LO
CA
L)
FIG
.4
-28
.G
PS
ST
AT
ICT
ES
T,
19
NO
V8
2
4.2.5 Interference Effects
Static tests conducted in the Aerocommander indicate that the UHF radios,VOR receivers, ATC transponder and DME did not appear to interfere with theoperation of the preamp and receiver channels. There was no observed effecton the operation of the aircraft avionics when the GPS receiver wasoperated.
In addition, a simulation of interference effects in the RockwellAerocommander test aircraft was carried out for the FAA by the DoDElectromagnetic Compatibility Analysis Center (ECAC). The ECAC study showedthat the Aerocommander avionics would not interfere with the operation of theGPS receiver for the particular antenna configuration used in the tests[Ref. 12].
4.2.6 Performance Monitoring
The performance monitoring features described in Section 3.5.6 wereverified during static tests. Features verified include the receiver faultmonitoring via receiver firmwave diagnostics and receiver-position processorinterface fault detection via the Position Software.
4.3 Flight Tests
During the course of the GPS test and evaluation project, a total of 25flights were made during which a total of 33 hours of flight data wascollected. A variety of flights were conducted at a total of seven differentairports as listed in Table 4-8. These tests measured system performanceduring takeoffs, landings, approaches and turns. The tests conducted forairports in Massachusetts and southern New Hampshire were within the range ofMODSEF surveillance.
The flight test program was divided into engineering and operationalphases. The engineering tests were conducted primarily to determine theeffect of aircraft dynamics and link margin on the accuracy and reliability ofthe GPS navigation system. The operational tests were conducted to determinethe performance of the system under a variety of conditions typical of generalaviation use and to verify the compatibility of the system with conventionalair navigation systems.
4.3.1 Engineering Flight Tests
A total of 14 engineering test flights were made during which a total of16.5 hours of data was collected. These flights were made over the six-monthperiod lasting from June to December of 1982. Position software developmentcontinued over this period, so that most of the effort was directed atassessing the accuracy of the system in various dynamic situations, such aslevel flight, turns, climbing and descent. At the beginning of December,1982, the position software was frozen at version 2.1, which was used for allsubsequent engineering and operational flight tests.
122
•
TABLE 4-8
AIRPORTS FOR GPS FLIGHT TESTS
Test MODSEFAirport Location Takeoff Landing Low-Approach Surveillance
Hanscom Field Lexington, MA X X X Yes
Gardner Municipal Gardner, MA X Yes•
Logan International Boston, MA X X X Yes
Boire Field Nashua, NH X Yes
Manchester Airport Manchester, NH X X X Yes
Burlington Burlington, VT X X NoInternational
Dulles Washington, DC X X NoInternational
123
4.3.1.1 Engineering Tests with Preliminary Position Software
An engineering test using the preliminary position software was conductedon 28 October 82. The results of this test are included here to illustratethe benefit of tracking all satellites in view and to quantify the fade marginduring steep turns.
In this test, the aircraft was initially in a holding pattern at theLOBBY intersection as shown in Fig. 4-29. After emerging from the holdingpattern, the aircraft performed two 360° turns at a bank angle of 30°. Theaircraft next was flown to intercept the Shaker Hills waypoint (SKR NOB) andthen landed on runway 11.
Figure 4-30 shows the satellite visibilities during the 28 October 1982flight. All five operational satellites were at least 10° above the horizonand SV5 had the lowest elevation angle, ranging from 18° to 25° during therun. Figure 4-31 shows the GDOP, HDOP and VDOP values during the flight. TheHDOP and VDOP values are calculated relative to the local MODSEF coordinatesystem. The GDOP value ranged from 5.1 to 5.4.
During the flight, the aircraft was under continuous MODSEF surveillance,allowing the system navigation accuracy statistics to be computed bypost-flight processing. The position accuracy statistics are calculated intwo ways based on: 1) the position fixes generated every 2.2 seconds, and 2)the a-B tracker estimates generated once per second. The position fixstatistics give the accuracy of the independent position fixes in the normalnavigation mode. The tracker estimate statistics give the system accuracy asdisplayed to the pilot using linear extrapolation between position fixes.
The position accuracy statistics for the 28 October flight are shown inTable 4-9. For this perliminary version of the position software, thehorizontal error from the a-B tracker position estimate was 364 feet (95%).This horizontal error value easily meets the AC90-45A requirement fornon-precision approach and is better than the 500 ft (95%) goal for straightand level flight. The rms horizontal error was 211 feet, which also meets theFRP requirements for non-precision approach.
The portion of the 28 October 1982 flight containing the 30° bank angleturns are shown in Figs. 4-32a and 4-32b. As shown in Table 4-10 the 95%horizontal position estimate error was 429 feet, a factor of two better thanthe 1000 ft 95% accuracy goal during 30° bank turns. Figures 4-33a and 4-33bshow the horizontal and vertical errors versus time during the 30° bank turnsfor the position fixes and the trackes estimates. Figure 4-34 shows themeasured carrier-to-noise-density (C/No ) ratios during the 30° bank turns forthe five visible satellites. Note that SV5 drops below the 33 dB-Hzloss-of-lock threshold at 390 sec and 515 sec. These two occasions correspondto the times when the aircraft was banked away from the satellite. FromFig. 4-30, the satellite elevation was about 20° so that the satellite was at-10° elevation angle with respect to the GPS antenna.
124
+5
...
x-v
plo
t:G
PS
alc
inM
OD
SE
Flo
ca
lsyste
m
28
OC
T8
2,
RU
N1a
~o x as o o LL
. w en c o ~ .. -Q) a> LL.
o
,,-.
--
+-
Sta
rto
fte
st
La
nd
ing
~----
-
-10
-5
Fe
et,
MO
DS
EF
loca
lX
10
4
FIG
.4
-29
.T
ES
TF
LIG
HT
RU
N1
a,2
8O
CT
82
o
__
__
__
__
__
__
__
__
_-------------8
---
----
-----
9
~~o:..:.
..:....:
.o__o
_0
_0
0
-,.~-.
-----
.---
~-.---
--.---
~~~
0--o
~~~-.-
---~~
-:--.
~--.--
-:-:~_~-..
-~-o=::'_-.=_=-~~ 6
~--.--
_0
-·---'-
-'-
'0
--·_
0.._
0-
15
00
15
00
00
.-.-
_._
0-_
0-
5
____
__..
_..
-·_
----
--··
4
__
=__"
=---
..-:
-----
---
9-------~_.
6 8
10
00
TIM
:S
EC10
00
TIM
ES
EC
50
0
50
0
._
_..
_4
---
---~-
-_._
--
_.-~'
_"
__
-0
-
----
-
-----
_.
60
50
z Q4
0... 4
( > w ~3
0w
20 10
01-
"N
35
00
\
30
0
J:2
50
... :::l
:IE2
00
N "!C
15
0
10
0
50 0
0
FIG
.4
-30
.S
AT
EL
LIT
EV
ISIB
ILIT
IES
FO
RR
UN
1a
,2
8O
CT
'82
..
4.3
4.
I
4.0
4.2
--------
--------
------
--------
---3.9
L______
_____
3.8
3.7
.I
I1
I•
2.0
11
.91 I
I'sl
1.7
I----~
I.6
..I.'_
__
_i_
____l~___J
__
_IL
__
L._
_"_
_._
..l.
._
_...L
_._
_'_
__
_l_
_..
..L
..._
_L
.__
L_
__
_L
__
__
i__
__
lc_
_''_
__
''_
__
_'_
'_-----',
---'.
Q. o o J:
Q. o o >
5.4
Q. o o CJ
5.3
5.2
5.
I
5.0
..L
_-'-
--
........
...'--
....
..I.
._--
..J..
.._
-'-_
-'--
_.L
.---
--I_
--I.
_--
"-_
......
..._~_"'_______'
___
'__
__
I...
.___'__~
__
'__
__
L_
...._
o5
00
10
00
TIM
E.
SE
C.
FIG
.4
-31
.D
ILU
TIO
N-O
F-P
RE
CIS
ION
,2
8O
CT
82
FL
IGH
T
TABLE 4-9
POSITION ACCURACY STATISTICS FOR RUN la, 28 OCTOBER 1982
•
Position Fixes:
StandardError Mean Deviation RMS 95%
Horizontal 130 111 171 276 ft
Vertical 185 100 210 375 ft
Tracker Estimates:
StandardError Mean Deviation RMS 95%
Horizontal 128 168 211 364 ft
Vertical 194 126 231 423 ft
128
..
....~
(~
l- I [, \.
~ ~-
<'
\,' \-
~-,
-,
,---
...
)-z...
GP
SP
osi
tio
nF
ix
!
/-:
----
----
--""
:
5 o10
('t) 0 ~ >< ca 0 0
......
U.
N 1.0
W (J') C 0 ~ - -Q) Q) u.
-65
-60
-55
-50
Fe
et,
MO
DS
EF
loca
lX
10
3
FIG
.4
-32
a.
PO
SIT
ION
FIX
PE
RF
OR
MA
NC
E,
30
0B
AN
KTU
RN
rt>0 Y
'" X -' c( 0 0 -' LL W
t-'
C/)
WC
00 :E t- W W LL
15
10 5 o
~"'.
./
.
~~~DSEF
TR
AC
K
~,.,
..--
----
,'''-
GP
SP
OS
ITIO
NE
ST
IMA
TE
(a-
J3T
RA
CK
ER
)
....
....
-65
,,
It,
t'
t,
•I
!'1ttl
-60
-55
FE
ET
,M
OD
SE
FL
OC
AL
X1
03
.I
'-5
0
..
FIG
.4
-32
b.
PO
SIT
ION
ES
TIM
AT
EP
ER
FO
RM
AN
CE
,3
00
BA
NK
TU
RN
•
TABLE 4-10
POSITION ACCURACY STATISTICS FOR 30 0 BANK ANGLE TURNS
Position Fixes:
StandardError Mean Deviation RMS 95%
Horizontal 181 54 189 245 ft
Vertical 166 104 196 322 ft
Tracker Estimates:
StandardError Mean Deviation RMS 95%
Horizontal 182 201 271 428 ft
Vertical 191 135 234 423 ft
131
27
5
+1
00
a:: o ~+
30
0w ..... a: w >
t-'
w N
a: o a: a: w ~ a: o :I: ..... u.
22
5
17
5
12
5
+5
00
I \:\fi
I\\}/
\I
II
4/
,~I
30
0
30
0
40
0
TIM
ES
EC
40
0
TIM
E.
SE
C.
IV\ I\
!
50
0
50
0
..
FIG
.4
-33
a.
PO
SIT
ION
FIX
ER
RO
RD
UR
ING
30
°T
UR
NS
•
60
0.0
50
0.0
l lL
30
0.0
I~
I~
200.
0~'v'V
.vt
Io:::L
.~.~o~.......
.....
a:4
00
.0
o a: a: w N a: o :J:
20
0.0
30
0
80
0.0
1 i600.0~
I-I
lL:
a:4
00
.0o a: a: w I a: w >
0.0
40
0
TIM
E.
SE
C.
50
0
"""-
1....
......
......
_..-1..~
I•.•
1,,·,.1•.
I.!
,••.
,.l.
It
.,t
•.•.
I.
,•.
1.,
••',.,,1
••••
'-2
00
.03
00
40
0
TIM
E.
SE
C.
50
0
FIG
.4
-33
b.
PO
SIT
ION
ES
TIM
AT
EE
RR
OR
DU
RIN
G3
0°
TU
RN
S
MEA
NC
INO
.(d
B-H
z)
60
(»
50~~~~~rvWVN'~
46
.7,.:
40
C 0)
30
20
60
50
CID4
6.7
...4
0C (I
)3
0
20
60
CI
50
46
.9
I-'
,.:4
0w
C.p
o(I
)3
0
20
60
105
0
1\~
44
.7...
40
C (I)
30
·~V
20
60
..5
0
~~
46
.5...
40
C 0)
30
20
30
04
00
50
0
TIM
ES
EC
FIG
.4
-34
.S
AT
EL
LIT
EC
/No
DU
RIN
G3
00
TU
RN
S
•
According to the GPS antenna gain characteristic of Fig. 4-14, changing theelevation angle from +20 0 to -10 0 causes an 8 dB drop in antenna gain. Thisloss in antenna gain would drop the satellite CINo from the mean value of 44.7dB-Hz to 36.7 dB-Hz. Since the CINo value for SV5 dropped to less than33 dB-Hz during portions of the turn, it seems likely that shielding from thewings, engine nacelle and propellers occurred in addition to the antennaloss.
A significant conclusion from the 30 0 bank turn example is the importanceof tracking all visible satellites rather than a subset of four SVs. Despitethe momentary drop of SV5, there were always at least four satellites in trackduring the turn. As a result, the position solution was maintainedcontinuously throughout the turn.
The results from the flight tests using the preliminary position softwarecan be summarized. First, the accuracy of the GPS navigator exceeded therequirements of AC90-45A for two-dimensional area navigation in the enroute,terminal and non-precision approach modes. Second, the system maintained theposition solution during 30 0 bank turns. Finally, it is seen that the systemcan maintain continuous position updates during momentary outages due to asatellite fade if adequate satellite coverage is provided.
4.3.1.2 Engineering Tests with Final Position Software
The position software was frozen at version 2.1 at the beginning ofDecember, 1982. The remaining engineering flight tests were conducted withthis software version. The aim of these tests were to: 1) measure theaccuracy improvement in the version 2.1 software, 2) verify performance duringtakeoffs and landings, 3) examine the behavior of the a-~ tracker and 4)determine the possible effect of multipath on system performance.
4.3.1.2.1 Accuracy Tests
To illustrate the accuracy performance achieved using the final versionof the position software, the Hanscom flight shown in Fig. 4-35 was selected.The flight was one of several touch-and-go landings performed on12 December 1982 at runway 29 and the results are summarized in Table 4-11.For this flight segment lasting about ten minutes, the rms horizontal errorwas 104 feet and the 95% horizontal error was 175 feet. It should also benoted that the 95% vertical error in this case was 215 feet, which marginallymeets the vertical navigation requirements of AC90-45A shown in Table 2-3.The horizontal and vertical errors for the 12 December 1982 flight segment arealso shown in Figs. 4-36a and 4-36b for the position fixes and the trackerestimates, respectively. The two large horizontal errors at about 150 secinto the run are due to misses in the MODSEF surveillance rather than GPSerrors; these artifacts do not significantly affect the error statistics.
135
.-
.'.
,..._
-------------------------------_
._---.,
HA
NS
CO
MF
IEL
D
---~9
~_
_T
AK
EO
FF
LA
ND
ING-------~
'~SKR
NO
B •
1,.
Jt
",
tt
i.,
tit
25
20
I",!
,.
t'
II,
15
\05
J.
f
o-5
-10
f2
51 j 1 1 1 i 1 t
20
j,., 0 X
15
...J
i
4(
I0
'.0
1 <...
J u..1
0, \
UJ
\I-
'en
wa
0\
0"-
~'- -
~5
UJ
UJ u..
0 -sL-1
5
FE
ET
.M
OD
SE
FL
OC
AL
X1
03
FIG
.4
-35
.F
LIG
HT
TE
ST
,2
1D
EC
82
TABLE 4-11
POSITION ACCURACY SIATISTICS, 21 DECEMBER 1982 TEST
•Position Fixes:
StandardError Mean Deviation RMS 95%
Horizontal 34 57 66 112 ft
Vertical 65 49 81 162 ft
Tracker Estimates:
StandardError Mean Deviation RMS 95%
Horizontal 34 98 104 175 ft
Vertical 87 65 108 215 ft
137
20
0M
OD
SE
FT
RA
CK
ING
ER
RO
R
15
0..
J < I Z o N10
0,a: o :::
t
50 a
.l.-
_--'
L..
.1..
..:.
....
----
-'--
~_
_..
.L_
_..
..l.
__
_.L
__
__
_I.
.__
...L
_'"
__
__
__
_'_
__
__
_L
_I_
__
__
_I
30
0
..J < o I a: w >
20
0
10
0 a
-10
0
60
05
00
40
03
00
TIM
E,
SE
C.
20
01
00
-2
00
..L..._
_..
L.-_
_L
-_
----1_
_..-
-J..._
_--
--'-
-..
.1..
-__
-'-
--_
-----.J
L--_
_--
--l._
_..-
-J..._
_--
--L_
_-1
.-_
----'
a
FIG
.4
-36
a.
PO
SIT
ION
FIX
ER
RO
RS
,2
1D
EC
82
FL
IGH
TT
ES
T
60
05
00
••
•I
••
I•
I•
••
•I
•,
,t
,I
•••
40
0
I.
30
0
TIM
E,
SE
C.
I•
,••
I•
,
20
0
,•
It
•I
10
0
30
0.0
0:
25
0.0
0 0:
0:
20
0.0
w -oJ~ I-
15
0.0
z 0 N
M,0
:1
00
.00 :I
:
50
.0
0.0
..... V.l
\.D
40
0.0i
30
0.0
0:
20
0.0
40 0
:0
:W
10
0.0
-oJ~
0.0
0 I- 0:
-1
00
.0w >
-20
0.0
-30
0.0
0
FIG
.4
-36
b.
PO
SIT
ION
ES
TIM
AT
EE
RR
OR
S,
21
DE
C8
2F
LIG
HT
TE
ST
Figure 4-37 shows the pseudo-range residuals for the 21 December 1982flight. The pseudo-range residuals denote the difference between the measuredpseudo-range and the calculated range based on the ~-a tracker positionestimate. Note that the magnitudes of the residuals increase for roughly 100seconds starting at about 400 seconds into the run. This time intervalcorresponds to the sharp turn during the flight near the Shaker Hills waypoint(SKR NOB).
The pseudo-range residuals increase during turns because of the behaviorof the ~-a tracker. The tracker projects the last position estimate forwardto the next position fix time, assuming that the aircraft is moving in astraight line at constant velocity. Because acceleration is not modelled, thetracker position estimate error increases during a turn. As a result, thepseudo-range residuals also increase. forcing the position solution algorithmto apply larger corrections to the position estimate in order to produce theposition fix.
4.3.1.2.2 Tracker Performance
As a further illustration of the effect of the ~-a tracker on thepseudo-range residuals during turns consider the flight segment of Fig. 4-38.In this test, a series of five 360° turns was made at 15° to 20° bank angle.Figure 4-39 shows the pseudo-range residuals for this test. It is readilyapparent from the plot that the five-turn series causes a variation in thepseudo-range residuals of about + 100 feet. It may be further noted that theresiduals for SV8 vary in the opposite sense from the other satellites; thisbehavior stems from the fact that SV8 is west of the aircraft while the restof the satellites lie to the east.
4.3.1.2.3 Multipath Tests
Tests were also conducted to determine if multipath effects could beobserved. In one multipath test the aircraft was flown over the ocean on agenerally south-westerly course while descending from 1500' to 500' inaltitude and then climbing back to 2500'. The satellite elevation anglesduring this test ranged from 25° to 55°. The flight profile was selected inorder to maximize multipath effects, which should vary according to altitude.The sea state was calm, also promoting maximum multipath magnitude. Theground path and altitude profile for the test are shown in Figs. 4-40 and4-41. respectively. The satellite visibilities are shown in Fig. 4-42a.
The pseudo-range residuals for the test are shown in Fig. 4-42b. It canbe seen that the residuals do not change in character throughout the test.There is a bias in the SV9 residuals of about 50 feet. but this bias does notchange with altitude.
It was concluded from this test that multipath effects were notsignificant for satellites of 25° elevation angle or greater. There were noeffects discovered in any of the flight tests that could be attributed tomultipath.
140
••
20
0.0
(I) ....
0.0
< en
-20
0.0
20
0.0
CD ....0
.0< en
-20
0.0
20
0.0
CD
I-'
...:0
.0+:
-<
I-'
en
-20
0.0
20
0.0
10 ....
0.0
< en
-20
0.0
20
0.0
~ ....0
.0< en
-20
0.0
01
00
20
03
00
40
05
00
60
0
TIM
E,
SE
C.
FIG
.4
-37
PS
EU
DO
-RA
NG
ER
ES
IDU
AL
S,
21
DE
C8
2F
LIG
HT
TE
ST
22
0
....
....
.
•1
.."..
"...
...
r<> 0 ... X ..J < 0
f-'
0po
...
J2
10
N
.... W W II..
....
....
....
..
....
....
....
"..
...
...
..
....
....
....
.....
... .. '.
....
....
..
20
0L
........
........
........
........
L-o
-...
.....
..L...
.....
.....
...L
....
....
....
....L
-o-..
......
.....L
....
..........J~..
....
....
....
....
.....
o..L
....
....
....
..L...
.....
....
.....L
......
......
........
a.....
......
......
......
......
..o..
L...
......
......
.J
90
10
0
FE
ET
,L
OC
AL
X1
03
I1
0
FIG
.4
-38
.F
IVE
36
00
TU
RN
SE
RIE
S,
25
JAN
83
•
80
07
00
60
05
00
40
03
00
tt
•,
••
••
••
••
•Itt
••
I••••
I•
,•
!.,
t.Itt
•,
I,
•,
t,
,t
t•
I•
,,
,I
•I
20
01
00
20
0.0
CJ) I-
0.0
-0< U
)
-20
0.0
_
I-eo
20
0.0
1
LLI-
0.01
w0<
..J
W-< :l
-20
0.0
Cl w
200.0
1w
(Qf-
"IX
:~ w
WI-
0.0
C'
0<Z
U)
-20
0.0
J
0< IX:
02
00
.01
Cl
:l
10
Jw W
I-a.
0< U)
1
-2
00
.0J
20
0.0
~ I-0
.00< U
)
-20
0.0
0
TIM
E,
SE
C.
FIG
.4
-39
.P
SE
UD
OR
AN
GE
RE
SID
UA
LS
FO
RF
IVE
36
00
TU
RN
S
24
.0
.-------------------------------------.,
¢
o yo
-
~20
.0
() o ..J
U.
W en a o :::i:
I W W U.
15
.0
5.0
/'
I' I/
/" , I ," , , , / I I \ I , I ~
10
.0
FE
ET
,M
OD
SE
FL
OC
AL
10
4
15
.0
FIG
.4
-40
.G
RO
UN
DT
RA
CK
FO
RM
UL
TIP
AT
HT
ES
T,
25
JAN
83
•
•
30
00
I-2
00
0I&
-
W 0 :::> l- I- ..J
10
00
-e:
I-'
-i"-
V1
30
04
00
20
01
00
o""&"O'~""&"O'
""&
"O'
~.o..J...."~
~...
......
JL..o
....o
.'
'J..
..!
J..
..I
'.1
.. ''..
.L'
0-1'
'.LJ
'o
TIM
E,
SE
C.
FIG
.4
-41
.A
LT
ITU
DE
PR
OF
ILE
FO
RM
UL
TIP
AT
HT
ES
T,
25
JAN
83
_._
._._
.-
--------------
---
---
------
--------
9
35i---------
======-:-==:===-==-===~----==5 8
60
·
55
50
z 04
5l- .e: >
40
w .oJ
W
30
....
....
....
...
..
62
5~
L..J
....o
~.~.~
',
..J.'
..1..1~~
'.L
..''-'
'.'-
.L..i
'~
''
,'
''
.&..'~
,.0
-',,
-,.,
35
0
J: l => ::i: N .e:
30
0
25
0
20
0
__.
..
.--
--._
.--_
.._-
-.--
----
----
----
----
---.
----
----
----
----
--.-
--4
_._
_._
._
._
_._
._
_._
_._
._
.-._
._
._
.-
5-------------------------.---------
9
......
......
......
......
......
......
..6
15
0
10
0
50
±:::::;
:;::::;:
::::L::
:::=:t::
::::::=
:::::L:
:;=::::
:::i::::
:;:;::::
::::L::
::::=::
:::L=::
::iL::::
=::::i,
L::::.=
.::::,::
::i:,:::
:.=.=
.=.::L
.=.=.
=.:::
:L:.:.=
:=..:::
:JL:::=
:::i,::,
::.::,:.
::L,:.:
:::,:.:;:
,:::.:.
:.,..,8
48
94
90
49
14
92
TIM
E(S
EC
.G
.M.T
.)X
10
2
FIG
.4
-42
a.
SA
TE
LL
ITE
VIS
IBIL
ITIE
SF
OR
MU
LT
IPA
TH
TE
ST
,2
5JA
N8
3
•
••
30
02
00
.TIM
E,
SE
C.
10
0
0.0
-10
0.0
-20
0.0
20
0.0
10
0.0
0.0
-10
0.0
-20
0.0
20
0.0
10
0.0
0.0
-10
0.0
-20
0.0
20
0.0
10
0.0
~ ~0
.0
en-1
00
.0
-20
0.0
.Lo
.~
o...J
...o~
.......a.
.c..
........
L..o..o.~L....o
~~t.....L..&
...
o
co t e(
U)
20
0.0
CJ)
10
0.0
~0
.0
en-1
00
.0
-20
0.0
20
0.0
co1
00
.0
.,.:
U- en ...J
e( :> a en w a: w CJ z e( a: o a :> w en Q.
FIG
.4
-42
b.
PS
EU
DO
RA
NG
ER
ES
IDU
AL
SF
OR
MU
LT
IPA
TH
TE
ST
,2
5JA
N8
3
4.3.1.3 Summary of Engineering Test Results
Based on the engineering tests, several conclusions were reached. First,the accuracy of the GPS system exceeds the requirements of AC90-45A fortwo-dimensional area navigation in the enroute, terminal and non-precisionapproach modes. As expected, the system accuracy was not sufficient to meetall of the requirements of AC90-45A for precision (three-dimensional) ILS-typelandings, although the performance was surprisingly close to meeting some ofthose requirements. The GPS accuracy also met the projected requirements fornon-precision approach as stated in the Federal Radio Navigation Plan.Finally, the system performance exceeded the project goals for level and 30°bank turns.
The second conclusion from the engineering tests was that the system isnot unacceptably affected by aircraft dynamics. The ability of the system tomaintain position updates throughout 30° bank angle turns was demonstrated.It was also shown that the system maintained continuous position updatesduring momentary satellite outages during a turn.
A third conclusion was that the a-S tracker performance is adequate eventhough it does not attempt to model vehicle acceleration. Given the lowairspeed and limited dynamics of general aviation aircraft, the use of thelinear tracker appears justified. The test results show that the systemaccuracy remains well within the horizontal accuracy requirements ofAC90-45A.
A final conclusion from the engineering tests was that multipathinterference did not appear to have a significant effect on system performanceDespite specific tests designed to produce multipath effects, no pseudo-rangeerrors occurred which could be attributed to multipath.
4.3.2 Operational Flight Tests
Upon completion of the engineering flight tests, a series of operationalflights were conducted to test the performance of the system as an areanavigator. These operational tests were conducted in three different areasrepresenting various airport types likely to be encountered by generalaviation aircraft. The areas were:
• Burlington International Airport - Burlington, Vt.
• Logan International Airport - Boston, MA.
• Hanscom Field - Lexington - Manchester Airport - Manchester, NH.
The tests at the Burlington, Vt. airport were designed to determine if a GPSnavigator can operate in an area surrounded by mountainous terrain. The testsat Logan International were made to verify that GPS can be operated at anurban airport adjacent to large buildings. The flights in the HanscomField/Manchester Airport area were designed to verify the correct performanceof the GPS navigation system in typical general aviation operations.
148
•
4.3.2.1 Burlington, Vt. Tests
The approach patterns to Burlington International airport are shown inFig. 4-43. There are two non-precision approaches to the airport: an NOBapproach to Runway 15 and a VOR approach to Runway 1. The airport is locatedbetween two mountain ranges, the Green Mountains at about 20 miles to the eastand the Adirondack Mountains at about 25 miles to the southwest.
For the Burlington operational tests, the GPS test aircraft was flownfrom Hanscom Field in Massachusetts to Burlington International on27 January 1983. A series of non-precision approaches were made using boththe VOR and NOB approach procedures. Both the BTV VOR and the BT NOB wereentered as waypoints in the GPS navigation software for this purpose.
4.3.2.1.1 Enroute Performance
During the first portion of the test flight, the aircraft was enroute tothe Burlington VOR from Hanscom Field as shown in Fig. 4-44. At the beginningof run 2a the aircraft was 70 nautical miles away from the VOR on the 151 0
radial. At the altitude the aircraft was flying (4250 feet), the VOR-DME wasunusable for ranges greater than 30 nm on this radial.
Figure 4-45 shows the altitude and ground speed for run 2a as estimatedby the position software. Fig. 4-46 shows the GPS satellite azimuth andelevation during the run. Note that only four satellites were visible duringthis period. The satellite carrier-to-noise-density values are shown inFig. 4-47. Note at the beginning of the run that SV9 has a GINo about 5 dBlower than the other SVs. At this time, SV9 was at 80 elevation angle, butstill had a substantial loss-of-lock margin of about 12 dB.
The pseudo-range residuals for 2a are shown in Fig. 4-48. The residualsare seen to remain small, typically less than + 50 feet, over most of the run.There are, however, three instances where the residuals became large: oncefor SV9 and twice for SV4. Analysis of the data showed that theseperturbations occured when the position software was unable to collect a validpseudo-range measurement from the receiver. Due to an apparent software bug,the failure to collect a valid measurement caused the next measurement to beimproperly time-tagged. The improper time-tagging, in turn, caused the rangeto the satellite to be calculated incorrectly. The incorrect rangecalculation then caused the pseudo-range residual to be in a error by about500 feet.
It is significant that a pseudo-range error of this magnitude has only asmall effect on the navigation accuracy; this is due to two reasons. First,the error is short-term, lasting for a maximum of two navigation cycles.Second, the batch least-squares method minimizes the effect of large residualson the position fix. In this case, the 500 foot residual error caused aposition fix error of only about 100 feet. A position estimate error of thismagnitude is not discernable on the pilot course deviation display. It shouldalso be noted that pseudo-range reasonableness check in the position softwarerejects any residual greater than 1000 feet, preventing a bad position fixfrom being generated.
149
BTNOB
BURLINGTONINTERNATIONAL"i-
fI~BTVVOR
036 0
FIG. 4-43. BURLINGTON, VT APPROACHES: NOB RWY 15 AND VOR RWY 1
150
o
v-1
00 ... X ..
J~ 0
f-I
0V
I..
Jf-
I-
I- w-2
0w LI
.
-3
0
BT
V•
VO
R
o1
02
0
FE
ET
,L
OC
AL
X1
04
30
40
FIG
.4
-44
.B
UR
LIN
GT
ON
,V
TO
PE
RA
TIO
NA
LT
ES
T,
RU
N2
a2
7JA
N8
3
50
00
l- ll-
45
00
ui 04
00
0::> l- I-
35
00
-l -e
30
00
en2
50
0a.. lI
-3
50
f-'
03
00
V1
WN
w2
50
a.. en 02
00
z ::>1
50
0 a::1
00
C)
50
40
54
10
41
54
20
TIM
ES
EC
.G
MT
X1
03
FIG
.4
-45
.A
LT
ITU
DE
AN
DG
RO
UN
DS
PE
ED
,R
UN
2a
,2
7JA
N8
3
80
······
······
······
······
······
·6
42
04
15
41
04
05
--.._-
_..-
-..-._
--_
..-.-
.--_
.._--
_.-..
__
_----
_..-_
..-..
-_._
__.__
.__.-_
.__-
-~ 5
_----------------------9
8
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
..6
----
--_.
_---
----
----
----
-_..
_._-
----
--_.
_-_..
..._
---_
.._.
_-_.
_-_.
_...
_-_.
_._-
----
--_..
._--
--_..
__._
--_.
_---
-.-.~
--~
~..........-=:
::::-=
-'-=--
==-=-
====
===-=
==_.:
..~.
60
z 0 I- 0< >4
0w ..
J W
20 0
r-'
VI
\..oJ
37
5
35
0
:I:
32
5I- ~ :::E
30
0N 0<
27
5
25
0
22
5
TIM
E(S
EC
.G
.M.T
.)X
10
2
FIG
.~-~6.
GP
SS
VP
OS
ITIO
NS
60
CD5
0l- e
( en4
0
30
60
CD5
0l-
I-'
e(
'.nen
40
..,...
30
60
CD5
0l- e
(4
0en
30
60
~5
0l- e
( en4
0
30
40
54
10
41
54
20
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-47
.S
AT
EL
LIT
EC
/NO
.V
AL
UE
S,
RU
N2
a,
27
JAN
83
50
0
CJI ...
0~ en
-50
05
00
co ...0
~ en
-50
0I-
'5
00
\.11
V1
co ...0
~ en
-50
05
00
't ...0
--t'lllll
..~ en
-50
0
t,
$
40
54
10
TIM
ES
EC
.G
.M.T
.X
10
2
41
54
20
FIG
.4
-48
.P
SE
UD
OR
AN
GE
RE
SID
UA
LS
,R
UN
2a,
27
JAN
83
The results of this test show the ability of the GPS system to performenroute navigation in mountainous terrain with a minimum set of foursatellites. It is of particular interest that the GPS was able tosignificantly outperform VOR in this mountainous area. GPS was able toprovide equivalent navigation service 70 nmi away from the VOR while the VORitself is unusable at ranges greater than 30 nmi. This performance advantageof GPS results from the fact that the GPS satellites are above the horizon,while VOR is ground-based and much more susceptible to terrain blockage.
4.3.2.1.2 VOR Approach
Runs 2b and 3b consist of non-precision VOR approaches to the Burlingtonairport. Figure 4-49 shows the ground track for run 2b. The aircraft firstintercepted the VOR on the 151° radial, then proceeded outbound on the 216°radial and performed a procedure turn. After intercepting the VOR on the 36°radial, the aircraft was then flown to the field for a low approach at runway33. Some gaps will be noted in the ground track; the gaps are due to datarecording problems rather than interruptions in the position solution.
Figure 4-50 shows the altitude and ground speed profiles for run 2b.Figure 4-51 shows the satellite visibilities for the run. Although fivesatellites are shown, this run was still navigating with four satellites (SVs4,6,8 and 9). The satellite C/No values are shown in Fig. 4-52. It is seenthat the satellite signal strength drops by as much as 10 dB in some turns,but always remains above the 33 dB-Hz loss-of-lock threshold. Figure 4-53shows the pseudo-range residuals; note the increases in the residualscorresponding to the turns. There are also some large residuals at the end ofthe run for SV4; these jumps result from the software bug previouslydescribed.
The ground track for run 3b is shown in Fig. 4-54. For this run, theaircraft intercepted the VOR then flew the 216° outbound radial and made theprocedure turn as before. After intercepting the VOR, an OBS setting of 36°from the waypoint was flown. It was found, however, that this resulted in a42° course from the waypoint, taking the aircraft to the east of the end ofrunway 33. Because of air traffic control considerations, a right downwindapproach was then flown to runway 15 and a landing made for refueling.
Post-flight analysis revealed an error in the navigation software whichhad the result of biasing the OBS setting by 6°, explaining the incorrectcourse of 42° commanded by the navigation software. This error was correctedfor subsequent operational flight tests.
Figure 4-55 shows the altitude and ground speed profiles for run 3b.Note that during the last portion of the run, the aircraft is on the groundand taxiing around the airport. The satellite visibilities are shown inFig. 4-56. All five satellites were in track during this run, including SV5which was initially at 10° elevation angle. The satellite C/No values areshown in Fig. 4-57. Note that SV5, the low elevation angle satellite, wasdropped on two occasions during the run. These occasions correspond to theprocedure turn and the base leg turn, at which times the aircraft was banked
156
30
...,.....-------------------------------------,
...... '"
'.'. '..
'.
'.. ,
", '.
..'......
... '. "
........~
'.-"..
. '.. .......
..~~-~ \ J I I / , I I !
';=
.../--
~b~
{.'
,, l
; !,1 J•
./!
.,.~
~
,,/
/"
.'PROCEDURE~'
•.•'
TU
RN
\.",
.•••
'---
_#
'
10 o
20
-10
-30
-20
I') 0 ~ X ..
J~ 0 0
t-'
..J
V1
-....J
... W W U.
~-4
0..
L.
~......:L~
......
JL.o
...o.
....
....
-30
-20
-10
o1
02
030
40
50
FE
ET
,L
OC
AL
X1
03
FIG
.4
-49
.B
UR
LIN
GT
ON
,V
TV
OR
AP
PR
OA
CH
,2
7JA
N8
3
......
V1
OJ
40
00
I- U. w
30
00
0 ::l
I-2
00
0I- ..
Jc(
10
00 0
(J)
30
0Il
.U
.
02
50
w w Il.
(J)
02
00
z ::l
0 a: C)
15
0 42
04
21
42
24
23
42
44
25
42
64
27
42
8
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-50
.A
LT
ITU
DE
AN
DG
RO
UN
DS
PE
ED
,R
UN
2b,
27
JAN
83
z o .... ~ > W ..J W
80
60
40
20
r-----------------------8
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
....·.
...····
·····6
._
._
._
._
_._
._
._
_._
._
._
.-5
o...
b.e:
''''
''''
''':
::i:
:IC:
.l.:
:I:l
I:::
z.;.
-::a
::::
;;::
:o,~
O::'
;'::
;:::
I::L
.:O:
::;:
::;:
:I;.
.;:~
..:I
:;;:
~...
a...
.....
L.o.
...L.
o.o..
L...o
.~
40
0
20
0-
30
0}
-------------
-------------------9
.........................................................................................
u_
6=~
=~:.:.:.:.~=
~:.:.:.:.-::
=':'='
".:':'
:':'~
:':'
:':'
".:'
:";
~:.:.:
.:.-:=
:.:.
:.:.
:.-::
':':':
'-.-.
••-::
=.:.=
.-::.:
.:.:.-
::.:.:
.:.-::
.:.:.:
.-::,:
,,:':'.
:::':'
:':'-:
:.:.:.
:':::"
:':':"~
10
0 o.L
.......
........
........
.........
L..o
......
......
o..L.
..o...
......
......
......
......
....J
......
......
......L
.......L
..........
...-.
......
.,,_.
...-I
....
.__
--=...
..,,"'
""'"':
t:::cI
l::lC:
ll::t:
l:z:::
cet:l:
::J::I
::I:I:
:::li:
::a:::
::a:::
:J::%J
a~.o.J
42
04
2I
42
24
23
42
44
25
42
64
27
42
8
TIM
E(S
EC
.G
.M.T
.)X
10
2
FIG
.4
-51
.S
AT
EL
ITE
VIS
IBIL
ITIE
S,
RU
N2
b,
27
JAN
83
60
0)
50
I- oe(
40
(/)
30
60
CIO
50
I- oe(
40
(/)
30
60
t--'
CQ0
\5
0<::
>I- oe
(4
0(/
)
30
60
~5
0I- oe
(4
0(/
)
30 4
20
42
14
22
42
34
24
42
54
26
42
74
28
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-52
.S
AT
EL
ITE
C/N
O.
VA
LU
ES
,R
UN
2b
,2
7JA
N8
3
42
84
27
42
64
25
42
44
23
42
24
2/
-25
0
-50
02
50 0
:"-2
50
-50
02
50 0
-25
0
-50
02
50 0
-25
0
-50
0 42
0
CD I .:(
Cf)
~ I -< Cf)<0
I .:(
Cf)
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-53
.P
SE
UD
O-R
AN
GE
RE
SID
UA
LS
,R
UN
2b,
27
JAN
83
(\J 0 .. X ..
JI-
'4
(0
\0
N0 ..
J .... W W U.
30
20
10 o
-10
-20
~r,
(I
,.~
_~B
UR
LIN
GT
ON
!.~
AIR
PO
RT
".'.. .....
...~ }) ,
til .. ... .'•
II
IJ .': .., .',
•I
i!
./i
/i--
BT
V:
.'V
OR
i..
1; 4
PR
OC
ED
UR
E~/!
TU
RN
,~./
('
,~./
....--
",....
__
•..
rr
-25
-20
-15
-10
-5o
10
20
30
40
50
FE
ET
,L
OC
AL
X1
02
FIG
.4
-54
.B
UR
LIN
GT
ON
,V
TV
OR
AP
PR
OA
CH
,R
UN
3b
,2
7JA
N8
3
40
00
30
00
w c2
00
0;:
) ... ...1
00
0.... <
0
-10
00 3°°1
c w Wf-
la..
20
00
\(J
)W
C z
10:J
;:) 0 a: C'
43
64
37
43
84
39
44
04
41
44
24
43
44
44
45
44
64
47
44
8
TIM
ES
EC
.X
10
2
FIG
.4
-55
.A
LT
ITU
DE
AN
DG
RO
UN
DS
PE
ED
,R
UN
3b,
27
JAN
83
70
60
""p
"".
··T
·..
·•
..··
·•··
·"
6 8z o l • > W ..J
W
50
40
30
4._.
._.__.
._...
__._._
_~-
:-::::
:.::-
:-:=.
::.:-:
-::::
.:::::
-::::.
::.:::
.-:.:.:
.:::.:.-
:.-:::
::::.-
:-:.::
.:::.:.-
::9
------
20 10
_._
._
._
._
5_
._-_
._._
.-.
_._
.-._
._.-
_._
.-.-
35
0
------------------------------
93
00
42
50
~-.~
.=:.
:.~:
.:.:
;,.-
::;,
;.;,
.~-~.
~~.~-
.~-~~
~~~~
~.~.
~~::~
~.~~.~
~~.~~:
:~~~
~~:~~
~~.~:
~~~.~
~~~.~...
~~.~.~
'.~:~.-
.~.~~
~~~~~~.~~~.~~~.~~.~~~~~~~~~~:
~:I
:I-
20
0::
I ::E N1
50
•1
00
50
__
__
__
__
__
__
__
__
__
__
__
__
__
8
44
74
46
44
54
44
44
34
42
44
\4
40
43
94
38
43
7
O-L
.--J
~..
.o..
..JL
_...
t...
...L....
.o..
...L
....
....
.J...
..L.
........J~
...-
J
43
6
TIM
E(S
EC
.G
.M.T
.)X
10
2
FIG
.4
-56
.S
AT
EL
LIT
EV
ISIB
ILIT
IES
,R
UN
3b
,2
7J
AN
83
60
0>5
0
...:4
04
( en3
0
20
60
co5
0
...4
04
( en3
0
20
60
co5
0
...4
0r-
'4
(
'"en
V1
30 20
60
10
50
...4
04
( en3
0
20
60
~5
0
...4
04
( en3
0
20
43
84
39
44
04
41
44
24
43
44
44
45
44
64
47
2T
IME
SE
C.
G.M
.T.
X1
0
FIG
.4
-57
.S
AT
EL
LIT
EC
/NO
.V
AL
UE
S,
RU
N3
b,
27
JAN
83
away from SV5. The position solution is maintained continuously in thesecases, however, since the other four satellites remained in track. Thisexample again illustrates the importance of tracking all visible satellitesrather than the minimum set of four.
Figure 4-58 shows the pseudo-range residuals for run 3b. It can be seenfrom the figure that there are a number of jumps in the pseudo-rangeresiduals. Except for the two cases in which SV5 suffered a loss of lock, thejumps are all due to the position software problem discussed earlier. In thisrun, the position software appears to occasionally fail to collect a validpseudo-range measurement before it is overwritten by the receiver with thenext measurement. This problem is symptomatic of a lack of available CPU timein the position processor. The lack of CPU time is attributable in part tothe large amount of engineering data being recorded by the system. The datarecording function introduces an overhead which would not be present in aproduction system.
It can also be noted from Fig. 4-58 that there is a substantial bias(-75 feet) on the SV4 residuals and a smaller bias in the opposite directionon SV6. The source of this bias is not known, but it may be due to agingephemeric data from SV4. Unlike the other satellites, SV4 was operating on acrystal oscillator because its atomic clocks had failed. Consequently, SV4'sclock drift is not well modelled by the ephemeric data from the satellite.The bias is clearly not due to multipath because it does not vary withaltitude, remaining constant throughout the run.
The results of runs 2b and 3b show that GPS can be used to simulate a VORapproach in a mountainous area. Furthermore, despite a position software bugthat introduced pseudo-range jumps, the system maintained reliable operationthroughout standard-rate turns even during the temporary loss of a low-anglesatellite. Also, the system kept a low (10 0
) elevation angle satellite intrack during level flight.
4.3.2.1.3 NOB Approach
The ground track for two NOB approaches to runway 15 (run 4) is shown inFig. 4-59. For this run, the aircraft took off from runway 15 and proceededoutbound on the 326 0 radial from the NOB. After a procedure turn, theaircraft made a non-precision approach on the 146 0 radial. A missed approachwas executed on advice from air-traffic control and a second approachinitiated. At this point, the run terminated due to a momentary aircraftelectrical power interruption.
The altitude and ground speed profiles for run 4 are shown in Fig. 4-60.The satellite visibilities are shown in Fig. 4-61. Figure 4-62 shows the,measured satellite CINo values for the turn. Note the large number of losseslof lock for SV6 during the run. Examining the satellite elevation, aircraft!altitude and topography of the region, it is clear that these dropouts are notIdue to terrain blockage. Examination of the data shows that SV6 was droppedlonce temporarily during the takeoff due an aircraft power surge. Lock wasIreestablished for the satellite in the first SV6 dwell slot in the navigation\Cycle but not in the second. As a result, the satellite CINo estimates forthe second dwell slot are not reliable, but were inadvertenly plotted.
166
50
0
Cl)
I-0
c( en
-50
05
00
GO I-0
c( en
-50
05
00
CD I-0
c(
-I-
'en
(j\ -...J
-50
05
00
10 I-
00< en
-50
05
00
~ I-0
c( en
-50
04
37
43
84
39
44
04
41
44
24
43
44
44
45
44
64
47
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-58
.P
SE
UD
O-R
AN
GE
RE
SID
UA
LS
,R
UN
3b,
27
JAN
83
""0 ... X ..... ...J -< 0
I-'
00
\...
J0
0.... ~ w w u.
30
20 10 o
-10
-20
40
30
20
10
o-1
0-2
0-3
0-3
0...
&.o
..L
....
....
""""
""....
.&..o
.....-
lI...l
...l.
~....L.o.
.....J...
..........
..........
...L..
....
FE
ET
(LO
CA
L)
X1
03
FIG
.4
-59
.B
UR
LIN
GT
ON
,V
TN
OB
AP
PR
OA
CH
,R
UN
4,
27
JAN
83
40
00
30
00
w 02
00
0:::> ..... .....
10
00
..J <
0
-10
00
30
0
o2.
00w W Q
.W o
10
0Z :::> o a: CJ
0
•
44
74
48
44
94
50
45
14
52
45
34
54
45
54
56
45
74
58
45
9
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-60
.A
LT
ITU
DE
AN
DG
RO
UN
DS
PE
ED
,R
UN
4,
27
JAN
83
z o i= -e > w -I
W
60
......
......
......
......
......
......
......
......
•:
.O
S"
30
20
_._
._
_._
._
_._
._
_._
._
_._
._
_._
._
._
._
._
._
.-
5
I0.L
......
........
.....L
......
..a...
....IL
.....o
....
.....
L..o
......
......
......
......
....r
.J...
....
....
.....
L..r.
......
......
...a....
......
.....r
.J...
......
......
..~...
......
......
......
.......
...a.
......
...a.
......
Il...o
...I..
......
......
......
......
.....o
...L
....
....
o...o
...I.
....
...~
35
0
30
0
25
0
20
0
15
0
10
0
50
-------------------------------9
~~~~~~~~~~~~~~~~~~~~~~=~
......
......
......
......
......
......
......
......
......
......
......
.__
_6
__
__
__
__
__
__
__
__
__
__
__
_8
o.L...~....J...,,...
......
......
&...
......
......
......
L....
......
......
J'_o
....o
......
....J
L...
o...a
.....
....
....
....
....
....
....
....
....
....
....
.....J.
..,,...
......
......
L...
......
...__
o...I
.....
....
....
....
.J..
....
....~
.L..
o...
.....
......
......
......
......
.......a
........
......
......
......
......
......
44
74
48
44
94
50
45
\4
52
45
34
54
45
54
56
45
74
58
45
94
60
46
14
62
46
3
TIM
E(S
EC
.G
.M.T
.)X
10
2
FIG
.4
-61
.S
AT
EL
LIT
EV
ISIB
ILIT
Y,
RU
N4
,2
7JA
N8
3
,
l e(
Cf)
l e(
Cf)
(Q l- e(
Cf)
I-'
'-J
I-'
10 l- e(
Cf)
•
60
40
20 o 60
40
20 o
60
40
20 o
60
40~"",,~'"
.,""
'-""#
20
... l e(
Cf)
o6
0
40 20 0..
.&..
.-_
-'-
_-L
._---L
_---'_
_L
.-_
..L
...-
_-'-
_..
....
L._
---L
_---'_
_1
...-
-_..
L..
.-_
....
L..
._..
...L
._--
-L_
---'
_--
--I
44
74
48
44
94
50
45
14
52
45
34
54
45
54
56
45
74
58
45
94
60
46
14
62
46
34
64
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-62
.S
AT
EL
LIT
EC
/NO
.V
AL
UE
S,
RU
N4
,2
7JA
N8
3
International airportThe number of flights
clearances into thecoincided with time ofwas gained at Logan tolarge, busy urban
It should be noted that the verison 2.1 position software does notcontain a provision for replacing failed satellites in the navigation modedwell cycle. This provision was built into version 3.0 software which was notcompleted in time to support flight testing. Had the satellite replacementfeature been installed. the position software would have been able toreestablish lock in the second SV6 dwell slot. It should be noted that thesatellite replacement feature. while not flight tested. was validated forversion 3.0 in laboratory tests.
The pseudo-range residuals for run 4 are shown in Fig. 4-63. Theresiduals are similar to those for run 3. Again note that SV4 had a negativebias and SV6 had a positive bias. The SV6 positive bias is probably due tothe batch least-squares position fixing algorithm attempting to minimize thenegative bias on SV4 due to its less accurate clock.
The results for run 4 are seen to be similar for the VOR approaches. Itis seen that the GPS system functioned well on the NDB approach in thismountainous region.
4.3.2.2 Logan International Tests
Operational GPS test flights were made into Loganat Boston. MA on 25 January 1983 and 28 January 1983.into Logan was limited by difficulty in obtaining ATCairport during the satellite visibility period, whichhigh air traffic density. However, enough experienceverify successful operation of the GPS system at thisairport.
Figure 4-64 shows the ground track recorded by the GPS system during run2 on 25 January 1983. The test aircraft approached the LYNDY waypoint (anon-directional beacon site) from the northeast. After intercepting thewaypoint, the aircraft began a non-precison approach to runway 22. Uponlanding, the aircraft taxiied to the end of runway 22R and took off. Afterintercepting the MILTT waypoint (also an NOB site), the aircraft turned to thenorthwest and departed the Logan area. Throughout the run, the GPS systemmaintained all five satellites in track and supplied continuous positionupdates. (Note: the gaps in the ground track shown in Fig. 4-64 are due todata recording dropouts).
Figure 4-65 shows the altitude and ground speed recorded by the GPSsystem during the Logan run. As seen in the figure. the middle part of therun was spent taxiing around the airport. During the time the aircraft wastaxiing. there were numerous airport buildings and large aircraft in theimmediate vicinity. Figure 4-66 shows the satellite visibilities for the run.Note that four of the satellites are to the west of the airport, towards thedowntown Boston area, and the remaining satellite is to the east, towards theocean. Figure 4-67 shows the dilution-of-precision values for the run ascalculated in post-processing. As seen in the figure, the GDOP varied fromabout 6.3 to 6.7 during the run.
172
•
..
50
0
.... c(
Cf) .... c(
Cf)
-50
05
00
co ...=f-
-'c(
0'-
JC
f)
w
-50
05
00
10 .... c(
Cf)
o+-~4,H',",~
..""""",..,...
..'....~"ff-,.~
-50
05
00
.... ~0~4~'_'i.l'tr_~
-5
00
:J..
---_
--L
.__
L--_
...L
....
_--'-_
_.L
....
._-L
._--1_
_.L
-_
--L
._---JI.
.-..
_..
.L..
.._
---L_
_"'
--_
...L
._--
-'_
----
J
44
74
48
44
94
50
45
14
52
45
34
54
45
54
56
45
74
58
45
94
60
46
14
62
46
3
TIM
ES
EC
.G
,M.T
.X
10
2
FIG
.4
-63
.P
SE
UD
O-R
AN
GE
RE
SID
UA
LS
,R
UN
4,2
7JA
N8
3
9.,
----
----
----
----
----
----
----
----
-...
~o ... )( ... w W LL
oLY
ND
YW
PT
o5
FE
ET
X1
04
10
15
FIG
.4
-64
.L
OG
AN
INT
ER
NA
TIO
NA
LT
ES
T,
25
JAN
83
,R
UN
2
•
..
30
00
t- IL W2
00
0e ::
;) t-1
00
0t- ..
Jc(
0
-10
00
f/)
3°°1
Q..
IL e.....
20
0-...
..IW
VI
W Q..
f/) e
10
0z :::
:l 0 a: CJ0 4
95
50
05
05
51
0
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-65
.A
LT
ITU
DE
AN
DG
RO
UN
DS
PE
ED
,2
5JA
N8
3,
RU
N2
70
60
...
...-
---.
•-.-
---.
-•..
..•
--~
----~-~--_..
.---
----
-_._
-_.--
----
--
z o I < > W ..J W
50
40
30
__
__
_._
._
._
._
._
.-
5-~.===-=-'~'---"'---'---'===:-
----.----
-------
..9
20
"-
.··"0
••••
_•••••••
'.'
,.
8 6
35
0
-------------
----
----
----
----
----
----
--9
•••••••••••••••••••
,•••••.•••••••••••••••••••••••
,-
•n
••••,.
6
_._
._
._
._
._
_._
._
._
._
._
._
._
._
-_
._
.--,-,-,·5
30
0
25
0J: I- :J
20
0::E N <
15
0
10
0
50 4
95
50
0
TIM
E(S
EC
.G
.M.T
.)X
10
2
50
5
8
51
0
FIG
.~-66.
SA
TE
LL
ITE
VIS
IBIL
ITY
,2
5JA
N8
3,
RU
N2
..
•..
51
0
t•
,
•f
50
5••
I••
5.4
0
5.3
5
Q.
5.3
00 0
~>
5.2
5
5.2
0
5.1
5-
3.5
3.0
Q.
02
.50 :I
:
2.0
-1
.5
6.7
6.6
Q.
6.5
0 0 CJ6
.4
6.3
6.2
II
49
55
00
TIM
E,
SE
C.
X1
02
G.M
.T.
FIG
.4
-67
.D
ILU
TIO
N-O
F-P
RE
CIS
ION
,2
5JA
N8
3,
RU
N2
The satellite CINo values are shown in Fig. 4-68. All of the satelliteCINo values remained well above the 33 dB-Hz loss-of-Iock threshold, althoughSV6 dropped substantially (about 10 dB) during the turn after takeoff fromrunway 22 R. The drop in CINo is not surprising since the aircraft bankedaway from SV6 during the turn.
It is significant that the GPS system kept the satellites in trackcontinuously while taxiing despite the buildings and large aircraft near tothe GPS aircraft. Although blockage of the satellite signal is probable whenthe aircraft is immediately adjacent to a building, significant blockage didnot occur on the taxiways or runways.
The pseudo-range residuals for the 25 Janaury 1983 run are shown inFig. 4-69. The residuals showed no jumps of the kind seen earlier, indicatingthat the position software was always able to collect pseudo-rangemeasurements from the receiver. Note the positive bias (+75 feet on SV9 andnegative bias (-50 feet) on SV5. As explained earlier, these biases appear tobe due to old ephermeric data transmitted by the satellites.
A second test flight was made at Logan International on 28 January 1983.Figure 4-70 shows the ground track for a low-approach and departure fromrunway 4L. The altitude and ground speed profiles for the flight are shown inFig. 4-71. After departing the runway, the aircraft made a turn to thenorthwest prior to intercepting the LYNDY waypoint on instructions from airtraffic control. The aircraft then departed the Logan area for HanscomField.
The satellite CINo values for the 28 January 1983 Logan run are shown inFig. 4-72. The run was made late in the satellite visibility window, so thatonly four satellites were available for navigation. Nonetheless, the GPSsystem was able to provide continuous navigation updates throughout the run.The pseudo-range residuals for the run are shown in Fig. 4-73. Note thepositive change in the SV8 residual and negative changes in the SV5 and SV9residuals corresponding to the initial turn in the run. Again, this behavioris explained by the satellite positions: SV8 was to the east while SV5 andSV9 were to the west. SV4 was more to the north and the aircraft velocity inthat direction changed little.
The Logan runs demonstrated that GPS can be used for non-precisionapproach into a large urban airport. It was also found that the GPS ~ystem
was able to maintain navigation service while on the taxiways and runwaysunder conditions that generally render VOR/DME systems inoperable. It shouldbe cautioned that these observations apply to the GPS system while innavigation mode; the lock threshold is 2 dB higher (less sensitive) for theacquisition mode. Also, some satellite signals are likely to be blocked whenthe aircraft is parked next to a large building such as a passenger terminal.Nonetheless, the satellite C/No margins appear quite acceptable once theaircraft is out on the taxiways.
178
•
•
•
51
05
05
50
0
55
50
45
40
35
55
50
45
40
35
55
50
45
40
35
55
50
45
40
35
55
50
45
40
35
..L...
o''-'
-'''-'
-'.&
....I.
..-
..L
~...
L.o
.--
L....
i...
..L..J
......
....J
-...J
49
5
CD I < UJ
I < UJ
CD I < UJ
I < UJ
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-68
.S
AT
EL
LIT
EC
/NO
.,2
5JA
N8
3,
RU
N2
20
0
(l)
0.... ee
-20
0C
J)
-40
02
00
' IQ
)
-20:1
.... ee en
-40
02
00
<0
0.... ee
-20
0I-
'C
J)
00 :..::>
-40
0_
20
0
II)
0.... ee
-20
0C
J)
-40
0.J
20
0
•0
.... ee CJ)
-20
0
-40
04
95
50
05
05
51
0
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-69
.P
SE
UD
O-R
AN
GE
RE
SID
UA
LS
,2
5JA
N8
3,
RU
N2
•
•
10
......
---
.---
-....
........
-....
......
......
. .....
0....
........
........
... .
..... ...
,..
-10
..-.,
..'O'
'.. '.r<
)'..
0. ...
..... '.
X". ..
.....
.....
-J....
~4
(-2
00
0....
..I-
'0 0
"'-\1
-J .......
....-"
,w
...w u..
-30
• ! , ! i , , I i-4
0I
RU
NW
AY
4L
1/-5
00
102
03
04
05
06
07
0
FE
ET
(LO
CA
L)
X1
03
FIG
.4
-70
.L
OG
AN
INT
ER
NA
TIO
NA
L,
28
JAN
83
20
00
t- LI-
15
00
ui 01
00
0;:
)t- t-
50
0..
.I0
(
0
-50
03
00
i-'
f/)
ex>
0-
25
0N
LI-
02
00
w W 0-
f/)
15
00 z ;:
) 01
00
[[ 0
50 0
48
94
90
49
14
92
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-71
.A
LT
ITU
DE
AN
DG
RO
UN
DS
PE
ED
,2
8JA
N8
3,
RU
N2
a
•
•
60
CD5
0l- e en
40
30
60
co5
0l- e
40
en
I-'
30
00
60
VJ
lO5
0l- e en
40
30
60
'lI'
50
l- e4
0en
30
48
94
90
491
49
24
93
49
4
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-72
.S
AT
EL
LIT
EC
/NO
.,2
8JA
N8
3,
RU
N2
a
25
0 o
-25
0
CD I- oe fI)
f-'
CO~
10 I- oe fI)
-50
0
25
0j
OJ
'A~
1\.
-../
-~V
~vi
IiV
'"~V
-25
0
-50
0J
25
0 o
-25
0
-50
02
50 o
-25
0
-50
04
89
49
0 TIM
ES
EC
.G
.M.T
.X
10
2
49
\4
92
FIG
.4
-73
.P
SE
UD
O-R
AN
GE
RE
SID
UA
LS
.2
8JA
N8
3.
RU
N2
a
•
4.3.2.3 Hanscom Field/Manchester Airport Tests
4.3.2.3.1 25 January 1983 Test
The final set of operational tests were made in the vicinity of HanscomField at Lexington, MA and Manchester Airport at Manchester, NH. The purposeof these tests was to verify the use of GPS as a navigation system in typicalgeneral aviation operations. Figure 4-74 shows the GPS-generated ground trackof such a flight into Manchester Airport on 25 January 1983. In this case theaircraft was flown north until the 337° radial to the MHT VOR waypoint wasintercepted. The aircraft was then flown on the radial, intercepting thewaypoint and descending for a landing at runway 35.
The altitude and ground speed profiles for the flight are shown inFig. 4-75. As seen in the figure, a substantial disturbance occured in thenavigation updates just prior to landing. Analysis of the test data showedthat this disturbance was caused by an incorrect position fix calculation.The calculation was incorrect because the position software failed to collectvalid measurements from the receiver in both SV4 dwells and a pseudo-rangefrom the previous navigation cycle was included in the position fixcalculation due to a position software error. The system corrected theposition error on the next cycle, 2.2 seconds later.
It should also be noted that the magnitude of the change in the positionfix was greater than a thousand feet, so that the fix should have beendeclared invalid by the position software. However, the limits on positionfix changes were unrealistically high in the Version 2.1 software used for thetest, so that the fix was passed on to the a-~ tracker. Despite this fact,the disturbance was sufficiently damped out such that it was not operationallysignificant.
The satellite visibilities for the 25 January 1983 run are shown inFig. 4-76. Note that SV6 was successfully tracked down to 4° elevation angle,below the nominal 5° elevation angle mask. Even more surprising, the aircraftwas on the ground at the time. Figure 4-77 shows the satellite G/No valuesfor run. Note that SV6 is dropped for a short time, just before landing, butis brought back into track. Shortly afterward, SVs 4,5,8 and 9 are alsodropped but tracking is immediately recovered.
Figure 4-78 shows the pseudo-range residuals.residual was caused by the incorrect position fix,The other two jumps occur in SV8 and are caused bytracking for that satellite.
The jump in the SV4as previously discussed.the temporary loss of
4.3.2.3.2 4 February 1983 Test
For subsequent operational tests in the Hanscom Field/Manchester Airportarea, a standard test scenario was developed as shown in Fig. 4-79. For thisscenario, the aircraft departs from Hanscom Field, runway 11, and is flown tothe Shaker Hills NDB waypoint (SKR 4). After intercepting the waypoint, the
185
18
....
..--
----
----
----
----
----
----
----
----
---.
...,
RW
Y3
6
MHT
VO
R
16 /4 12
,..
-I -< 0
I-'
00
0-I
10
0'
.... to- W W U-
8 6 4
-18
-16
-14
-/2
-/0
-8
33
70
RA
DIA
L
. I I ( \ \ \ , I -6-4
-20
2
FE
ET
(LO
CA
L)
FIG
.4
-74
.M
AN
CH
ES
TE
RA
IRP
OR
T.
25
JAN
83
.R
UN
3
..
30
00
...2
00
0LI
.
W1
00
00 ~ !::
0... ..
J4
(-1
00
0
-20
00
t/)
40
0Q
.t-
'LI
.O
Jci
30
0-.
.,J
w W Q.
20
0t/
)
0 z1
00
~ 0 a: CJ0
I
51
55
20
TIM
ES
EC
G.M
.T.
X1
02
FIG
.4
-75
.A
LTIT
UD
EA
ND
GR
OU
ND
SP
EE
D,
25
JAN
83
,R
UN
3
--
-_
._---
-_.__
.__
._--_
._--_
._-----
----
70
60
50
z 0 j::4
04
( > w3
0..
J W
20 10 0
......
<Xl
0::>
40
0
30
0
-_._
--_
._
._---
-_.-
_.__
._
-----
----
----
-_.__
.__._
-
---
----
-------------------------
____
____
____
____
____
___.
.._
._-.
.._----.--_._----.--------._-.---.--~4
--
-y
o----
4 5 9 8 6•
I
5
---
----
----
--------
--------
----
-----
-~
-9
20
0
10
0
".~.
'.
, , , , ·, · , ·,
6 8
..
51
5
TIM
E(S
EC
.G
.M.T
.)X
10
2
FIG
.4
-76
.S
AT
EL
LIT
EV
ISIB
ILIT
IES
.2
5JA
N8
3.
RU
N3
52
0
(Q ~ ~ enf-
>(X
l
1.0
10
~ ~ en
..
60 ::~
30
20
60 20
.b••••
I.
I,.,
I.
0tit
•••
I•••
1••••
••••
,I••••
I••
Il••••
I.
,••••
I••••
I,
•1
.,••
I••••
"t
•,
,
51
55
17
TIM
ES
EC
.G
.M.T
.X
10
2
52
05
21
FIG
.4
-77
.S
AT
EL
LIT
EC
/NO
.,2
5JA
N8
3,
RU
N3
•If.
!•
1•
,••
4t
,t
,It.
•A!'
I..
•I
,t
I••
.,
••
,,
d'
!.
I.
••
I••
!,.!
..•..
•.
l••
,I
II•••
If
••
•I
52
05
15
....
..01
•.,.
"..
W'W
--\0
1\'
~.••_~~"
•..~..
."~
-,
in
w.'
.yo
..~
"j".
-,
T'9
,..
.
50
0
Ol
I-0
< CI)
-50
05
00
co I-0
< CI)
-50
05
00
CIO
I-0
< CI)
f-'
1.0
0-5
00
50
0
." I-0
< CI)
-50
05
00
• ~0
< CI)
-50
0
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-78
.P
SE
UD
OR
AN
GE
RE
SID
UA
LS,
25
JAN
83
,R
UN
3
MANCHESTER
AIRPORT
PROCEDURE TURN
1570
MISSED APPROACH
O 26.1 nm356
HANSCOM
1120 FIELD
RWY11
1140
19.2 nm
LOBBY'~ "30
9.9 nm
~BE1 "-4.0
nm"t'.0 1140
nm~ ij4.3 n~ SKR 4
FIG. 4-79. TEST SCENARIO FOR HANSCOM FIELD/MANCHESTER
AIRPORT OPERATIONAL FLIGHTS
191
aircraft then intercepts the 356 0 radial to the Manchester VOR waypoint(MHT 6). After overflying the Manchester waypoint, the aircraft is turned to157 0 outbound, performs a procedure turn and turns inbound on the 337 0 radial.Upon intercepting the waypoint again, the aircraft begins the descent torunway 35. A missed approach is then performed with the Manchester waypointas the Missed Approach Point (MAP).
After overflying the Manchester waypoint again, the aircraft departs onthe 221 0 radial for the LOBBY intersection waypoint (LOBBY 2). The aircraftenters a holding pattern at LOBBY, then departs on the 113 0 radial for theBedford NDB waypoint (BE 1). The descent to runway 11 at Hanscom Field beginsafter the Bedford waypoint is overflown. The aircraft then lands at Hanscomor performs a low-approach to start another circuit of the course.
The ground track for Run la on 4 February 1983 is shown in Fig. 4-80.The test aircraft was under continuous MODSEF surveillance commencing justbefore the aircraft reached the SKR waypoint and ending as the aircraft leftthe LOBBY waypoint. As can be seen from the figure, the aircraft wassuccessfully navigated to the selected waypoints using the GPS system. On theShaker Hills to Manchester leg, the pilot maintained the course indicated bythe system fairly closely, but departed from the GPS course on the Manchesterto LOBBY leg due to ATC traffic advisories. The altitude and ground speedprofiles for the run are shown in Fig. 4-81.
The satellite visibilities for Run la are shown in Fig. 4-82 and thedilution-of-precision calculations are given in Fig. 4-83. Figure 4-84 showsthe satellite CINo values during the run and Fig. 4-85 shows the pseudo-rangeresiduals.
Since the test aircraft was under MODSEF surveillance during the run thesystem accuracy could be determined by post-flight processing. Figure 4-86ashows the horizontal and vertical position fixing error as a function of time.There are several large horizontal errors shown in the figure; analysis of thedata shows that the large values are instances in which the MODSEF facilityfailed to track the aircraft. These spurious errors do not substantiallyaffect the error statistics. The horizontal and vertical error in the GPSposition estimate (tracker output) is shown in Fig. 4-86b.
Table 4-12 shows the position accuracy statistics for Run la. Thehorizontal accuracy of the tracker estimates is seen to be well within thesystem performance requirements and goals.
In order to examine the behavior of the system during a turn, the holdingpattern section of Run la was expanded as shown in Figs. 4-87 and 4-88. TheMODSEF radar track is plotted with the GPS position fixes in Fig. 4-87 andwith the tracker estimates in Fig. 4-88. The horizontal and vertical errorsfor the position fixes and tracker estimates are shown in Figs. 4-89 and 4-90,respectively. As seen in Fig. 4-87 and 4-89 the horizontal position fix errordoes not build up during the turns, and the only major discrepancies betweenthe GPS position fixes and the radar track appear to be MODSEF trackingerrors. Note in particular that during Turn 3 the horizontal position fixerror remains less than 150 feet throughout the turn.
192
PR
OC
ED
UR
ET
UR
N
MIS
SE
DA
PP
RO
AC
H
--.---.
--.-
~
SK
R
I,
,I
.,
I.
-6-4
-20
24
68
FE
ET
,L
OC
AL
X1
04
MA
NC
HE
ST
ER
RW
Y3
51
6
18...-----------------------------------------...,
14
12
v 01
0.... X -l
~«
\00
8w
0 -l ... w
6w u.
4 2 0
-2
-1
6-1
4-1
2-1
0-8
FIG
.4
-80
.O
PE
RA
TIO
NA
LT
ES
T,
4F
EB
83
,R
UN
1a
l- LL ui 0 => !:: I- ~ .c
t-'
en1.
0.p
.Q
.LL 0 W W Q
. en 0 z => 0 a: CJ
25
00
20
00
15
00
10
00
50
03
00
25
0
20
0
15
0
100
-o..
...I
'--
-o...
..I~
...L
o....
...JI
....o
....
~...Lo.~.......JI....o....
......
L..
..J...
....
....
....
JL.-
....
J
42
84
29
43
04
31
43
24
33
43
44
35
43
64
37
43
84
39
44
04
4I
44
2
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-81
.A
LT
ITU
DE
AN
DG
RO
UN
DS
PE
ED
,4
FE
B8
3,
RU
N1a
,(P
AG
E1
OF
2)
••
30
00
45
54
56
45
44
53
45
24
50
45
14
49
44
84
47
44
64
45
44
44
43
44
2
15
0--
~
44
1
25
0
20
0
50
03
50
30
0o w W Q
.fI
) o z :::> o 0: CJfI)
Q.
Ii.
w2
00
0o :::> .....
15
00
..... ....I
4(
10
00
.....2
50
0Ii
.
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-81
.A
LT
ITU
DE
AN
DG
RO
UN
DS
PE
ED
,4
FEB
83
,R
UN
1a,
(PA
GE
2O
F2
)
z o ... « > w -J W
60
50
40
30
20
----
----
.....-
----
----
---
----
----
----
-------
-----------
-~'-'
'.~~~~~~~:::~-=
-_..-
-----
---.--
'..
'".
-...
..--
----
----
-----
-"..
...-.
...
----
----
--.-
-..--
--.-
--.-
-,--
----
'-'-'
-_._
._._._
._._.
'---'-
'--I0
...L.
.......
........
........
..L...
......
......
......
...L.
........
........
........
....L
......
......
......
......
..L...
......
......
....0
-1--
'--'
......
......
......
..J<
---J
,....
....
....
.--'
--'.
....
...-
....
....
....
....
....
....
....
....
....
....
...--'-
-''--'-
....
....
....
....
....
....
....
....
..--'-
-''
....
....
....
....
....
....
....
....
....
....
....
....
.-'-'.
....
....
.L-I
35
0
---
----
-----------------
---
---
7\0
...-
=-=-
=-=-
=..:.:.
-:.-:
....--
----
----
---
-----------------------------------------~~~.---_.---_.----
----
----
----
----
------
-_
.__.--
----
---_
.---
---
----
--.--
-_.-
_.-
_.-
--_.
----
30
0
25
0
20
0
15
0
10
0
.........
-.
.. .
....
....
....
....
....
-.
.",-
'"-
.
50
j------------------
o..l..
..l..
.L..
a.-'
-'L
....
.....L
-'-'
....
L.
J<--
-J,.
...J
....
.oJ.
.....
L..i
......
.o...
J..a
.....J
._
_~
43
04
40
TIM
E(S
EC
.G
.M.T
.)X
10
2
FIG
.4
-82
.S
AT
EL
LIT
EV
ISIB
ILIT
IES
.4
FEB
83
.R
UN
1a
45
0
••
••
5.0
4.5
45
54
50
44
54
40
3.0
6.0
5.0
4.0
3.0
2.0 1.0
6.0
5.8
5.6
5.4
5.2
5.04
25
43
04
35
3.5
Q. o o Cl
Q. o o J:Q. o
4.0
o >
1-0~
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-83
.D
ILU
TIO
NO
FP
RE
CIS
ION
,4
FEB
83
,R
UN
1a
GO
l
50!M
II'4
0,'
"
30J
20
J :~l.~-
tfrlIf
"I\1."
""''''
'n'''l
'''''~
40rtT I
30~ I
20
JGO~
50
-i I
4°1 3°1 20
JG
Oi I ::'j
~2
0 ~~
..,
OJ ... ..:: CI.l
CD ..: ..:: CI.l co ... ..:: CI.l ...,. ... ..:: CI.l... ..:: CI.l
43
04
40
45
0
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-84
.S
AT
EL
LIT
EC
/NO
.,4
FEB
83
,R
UN
1a ••
•
co ...: 0< en
50
0
-50
05
00
1 I o~..,~
.....~~
,...~"..
..~...
...,.._~
I
-50
0J
50
0-
.... 0< en .... 0< en
-50
0_
50
0l I O-~••'
.~W•
••"".~""","""""""t~jrS""~f
".'ft...
.#fIf~.~
I I-5
00
J5
00
~~~~~~1~.~,~.~~1~.~.~.~~~~,~1~.~.~.~••••••
,.
I.
t••
I••
,!
II'.
tI,.
'"
"~
-50
0
~ ~°1"iL}J~."'_••"
".~~w-f1f",,"."lIIIIlIl~"''Wii...,
....."
-.,WJt~filllI~fIfI'"
en
43
04
40
45
0
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-85
.P
SE
UD
OR
AN
GE
RE
SID
UA
LS
,4
FEB
83
,R
UN
1a
N o o
70
0-,
60
0
l- ll.
50
0
a: o a: a: w ~ a: o J:
20
0l I
100~I
tI
II:
0,o a:
,a:
-1
00
-iw ...: ffi
-20
0:>
-30
0
•
-40
0....
....L-.
.J...
....
.!
•I.
,!,
+...
....1..•••
I•••
1•
..L
,..•
•,.
1•
....
I••
!I
I~
•••I.·
!I
!••~
43
04
40
45
0
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-86
a.
PO
SIT
ION
FIX
ER
RO
R,
4F
EB
83
,R
UN
1a
•
60
0
50
0.... IL
.
II:
40
00 II
:II
:3
00
w ci II:
20
00 :I
:
10
0 0tv 0 I-
'
40
0
20
0.... IL
.
II:
00 II
:II
:W ....
-20
0II
: w >-4
00
-60
04
30
44
04
50
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-86
b.
PO
SIT
ION
ES
TIM
AT
EE
RR
OR
,4
FE
B8
3,
RU
N1
a
TABLE 4-12.
POSITION ACCURACY STATISTICS,
4 FEBRUARY 1983, RUN 1A
(42806 - 45511)
Position Fixes:
•
Error
Horizontal
Vertical
Tracker Estimates:
Error
Horizontal
Vertical
Mean
78
104
Mean
76
114
202
StandardDeviation
121
68
StandardDeviation
134
82
RMS
144
124
RMS
154
140
95%
257 ft
221 ft
95%
268 ft
260 ft
•
35
.-------------------------------------------.
30
r<>
t:N4
Lo
om
aN
FI~0
MO
DS
EF
TR
AC
K-
25
)( ..J -e 0
N0
0..
J
WII
.. W2
0en 0
LO
BB
YWPT~
TUR
N~
0 ~ ..: w w II..
15
0RN'
10 -1
15
-1\0
-10
5-1
00
-95
-90
-85
-80
FE
ET
,M
OD
SE
FL
OC
AL
X1
03
FIG
.4
-87
.H
OLD
ING
PA
TT
ER
N,
4F
EB
83
,R
UN
1a:
PO
SIT
ION
FIX
ES
I3
5,-
--------------------------------------.
10
30
I') 0 ... X ...I
0(
25
0 0 ...I
LL W enN
00
0+-
-:E
20
...LOBBVWPT~
w w LL
15 -1
15
-11
0-1
05
-10
0-9
5-9
0-8
5-8
0
FE
ET
,M
OD
SE
FL
OC
AL
X1
03
FIG
.4
-88
.H
OLD
ING
PA
TT
ER
N,
4FE
B8
3,
RU
N1a
:T
RA
CK
ER
ES
TIM
AT
ES
•
30
0TU
RN
1TU
RN
2TU
RN
3
•
TUR
N4
N o VI
.... u.
25
0
20
0
15
0
10
0
.... u. .... D:
W >
o
-10
0
-20
0
-30
0
45
04
51
45
24
53
45
44
55
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-89
.P
OS
ITIO
NF
IXE
RR
OR
SF
OR
HO
LD
ING
PA
TT
ER
N,
4F
EB
83
,R
UN
1a
tv o 0\
l LL a: o a: a: w N ir o J:
35
0
30
0
25
0
20
0
20
0
10
0
TUR
N1
TUR
N2
TUR
N3
TUR
N4
l LL a: o a: a: w .-: a: w >
o
-10
0
-20
0
-30
0
-40
0
-50
0
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-90
.T
RA
CK
ER
ES
TIM
AT
EE
RR
OR
FOR
HO
LD
ING
PA
TT
ER
N,
4F
EB
83
,R
UN
1a
..
Figures 4-88 and 4-90 show the performance of the tracker in the holdingpattern. It is seen that a considerable bias error occurs in the trackerestimates during some of the turn segments. This bias error is especiallynoticeable in Turn 3 where the horizontal error increases to as much as300 feet. A possible explanation for the increased bias error in Turn 3 canbe obtained from Fig. 4-91 which shows the altitude and ground speed profilefor the holding pattern. It is seen that the ground speed was about 300 feetper second (180 kts) at the beginning of Turn 3 but declined to 200 fps by theend of the turn 60 seconds later. Because this deceleration is unmodelled,the tracker tended to overestimate the distance traveled between fixes at thebeginning of the turn. As the ground speed declined, the errors in Turn 3also declined. By contrast, the ground speed was low at the start of Turn 4but increased during the turn; consequently, the horizontal errors tended toincrease during the turn.
The position error statistics for the holding pattern are shown inTable 4-13. The horizontal error of the tracker estimates is clearly wellwithin the system requirements and goals.
Run 1b on 4 February 1982 was similar to Run la, except that a lowapproach was made at Manchester airport and the aircraft was flown back to theBedford NOB (BE 1) waypoint instead of the LOBBY waypoint. The accuracystatistics were similar to Run 1a.
4.3.2.3.3 9 February 1983 Test
A second operational test of the type conducted on 4 February 1983 wasmade on 9 February 1983. Figure 4-92 shows the ground track for Run 1a. Forthis test, the aircraft took off from Hanscom Field runway 29 (instead ofrunway 11) due to prevailing wind conditions. After departing the runway, theaircraft was turned 1800 toward the Shaker Hills NOB (SKR). The remainder ofthe run was made according to the test scenario of Fig. 4-79. The gap in theground track after the turn at the SKR NOB is due to an inadvertentinterruption in the data recording; the GPS navigation updates were continuousthroughout the run. The altitude and ground speed profile of the run is shownin Fig. 4-93.
The satellite visibilities are shown in Fig. 4-94. Although thevisibilities for all five satellites are shown, only four satellites (SVs4,5,6 and 8) were used for navigation during the run. Despite this the systemmaintained continuous navigation updates for a period of nearly two hours.The dilution-of-precision calculations for Run 1a are shown in Fig. 4-95.Note that the GDOP for the run is higher than that for the test of 4 February1983 due to the use of only four satellites.
The horizontal and vertical position fix error versus time is shown inFig. 4-96. The corresponding data for the tracker estimates is shown inFig. 4-97. It should be noted that the MODSEF radar tracking data did notstart until time = 43085 sec due to difficulties with the radar system; thestart of the radar tracking is indicated in Fig. 4-92. The accuracystatistics for the run are summarized in Table 4-14. These statistics comparefavorably to the 4 February data.
207
27
50
TUR
NTU
RN
TU
RN
TUR
N1
23
4t-
25
00
u. ui2
25
00 ;:
) t:2
00
0t- -J -e:
17
50
15
00
35
0(J
')
Na..
0u.
00
30
00 w w a..
25
0(J
')
0 z2
00
;:)
0 0:
C)
15
04
50
45
14
52
45
34
54
45
5
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-91
.A
LT
ITU
DE
AN
DG
RO
UN
DS
PE
ED
FO
RH
OL
DIN
GP
AT
TE
RN
,4
FE
B8
3,
RU
N1a
TABLE 4-13.
POSITION ACCURACY STATISTICS FOR HOLDING PATTERN
4 FEBRUARY 1983, RUN lA
Position Fixes:
StandardError Mean Deviation RMS
Horizontal 55 75 93
Vertical 151 73 198
95%
155 ft
277 ft
Tracker Estimates:
StandardError Mean Deviation RMS 95%----
Horizontal 49 49 70 253 ft
Vertical 157 98 185 331 ft
209
18 16 14 12
....1
0...
IC 0 0 ...
I8
..... ...tv
Wf-
'W
0U
-6 4 2 0
LO
BB
YW
PT
PR
OC
ED
UR
ET
UR
N
86
42
o-2
-4-6
-8-1
0-1
2-1
4
-2_
..L
._
__
_...
L...r
...
L.
....
...
-16
FE
ET
(LO
CA
L)
FIG
.4
-92
.O
PE
RA
TIO
NA
LF
LIG
HT
TE
ST
,9
FE
B8
3,
RU
N1a
•
,
44
04
35
•!
•I
•,
I•••
I,
•,
•1
••••
I••
,•
I•
t••
I••
,••
,•
,,
•
30
00
l- lL W2
00
00 :) t:
10
00
I- .... <l(
0
N I-'
40
0I-
'C
/) no lL3
00
0 w w no2
00
C/) 0 Z
10
0:) 0 a: CJ
,.l!,.,
I.
,0
.•
I.
,I,
•I••
•I
.,
tI,
42
54
30
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-93
.A
LT
ITU
DE
AN
DG
RO
UN
DS
PE
ED
.9
FE
B8
3.
RU
N1
a(1
OF
2)
45
5
•••
I.
••
I••••
I"
•I••••
I••
,•
I••
I••
45
04
45
30
00
25
00
... u.2
00
0U
J c :;)
15
00
t: .../0
00
...J -e
50
0 0
N ~ N
30
0
C/)
25
00.
. u. ti2
00
UJ
UJ
0..
C/)
15
0c z
10
0:;
)
0 a: (!)
50 0
44
0
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-93
.A
LT
ITU
DE
AN
DG
RO
UN
DS
PE
ED
,9
FEB
83
,R
UN
1(2
OF
2)
•
60
., ~
50~
.....-..
..-
____
_.,
--•
----
.4
.....~
-
---_.
....-
6
-._
--
---
--
--
--(9
)
__
__.-5
''--:::-._
.8
---.~.---
.~._----.--
------.
---.-
.-'-
-_.
-_.
---
-
.-_.--
---
--
-~
..~.
--'-
----
---3
0
20
40
z o .... 4( > W ..I
W
N ......
W
35
0'
------
30
0
25
0
____
____
____
____
___
4
---_=~_:-
...~-~-----.&..-:..-
:::..
:.::.:
..:.-
---.
..---
...-----_--_-_-=~.--.--.--
-5
-----~-=.=-~.-_.
--'-
--._.
-.--
--(9
)
20
0
150
·..·
·..
6
10
0
50
_------------------8
t•
I~~.r......~..,
",'..II~.~o-
"--'.
...J.
..~
I_.~.~.o...Ll
.JIL
...o.
,..1
.'..1
'o 4
20
II
42
5
,,
I
43
0
I.
«
43
54
40
44
54
50
45
54
60
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-94
.S
AT
EL
LIT
EV
ISIB
ILIT
IES
,9
FE
B8
3,
RU
N1a
46
04
55
45
04
45
44
04
35
43
04
25
5.2
5
5.0
0
4.7
5n. 0
4.5
00 >
4.2
5
4.0
0
3.7
5
4.0
3.5
n.3
.00 0 :J
:2
.5
N2
.0t-
'~
1.5
6.5
6.4
n.6
.30 0 CJ
6.2 6.1
6.0
42
0
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-95
.D
ILU
TIO
NO
FP
RE
CIS
ION
,9
FEB
83
,R
UN
1a
..
50
0
t- Il.
40
0ri 0 IE
:3
00
IE:
W -' < ...2
00
z 0 ~ IE:
01
00
X
N0
I-'
V1
40
0
30
0t- Il
.
IE:
20
00 IE
:1
00
IE:
W -'0
< 0 t--1
00
IE: w >
-20
0
-30
0 43
0
1I
•t
••
I,
•,
d
44
0
...
.•
t«
,I
•,
,t
•I
".,.
I,
•,
II
•,
•,
1•
•..
....
....
....
,
45
0
TIM
ES
EC
.G
.M.T
.
FIG
.4-9
6.
PO
SIT
ION
FIX
ER
RO
R,
9F
EB
83
,R
UN
1a
50
0"'
~I
ILI I
a:4
00
,
0 a: a:3
00
w ..J 0< ~
20
0z 0 N a: 0 :I
:
N6
00
I-'
0"
~ IL4
00
a: 0 a:2
00
a: w ..J 0<
00 ~ a: w
-20
0>
-40
0 43
0
••
,I
tI
«•
I.,
,.
,
44
0
TIM
ES
EC
.G
.M.T
.X
10
2
tI••••
I•
,t
•I
•,
«•Itt.
tit
•t
••
45
0
FIG
.4
-97
.T
RA
CK
ER
ES
TIM
AT
EE
RR
OR
,9
FE
B8
3,
RU
N1
a
TABLE 4-14.
POSITION ACCURACY STATISTICS FOR 9 FEBRUARY 1983
RUN lA (43085 - 45770)
Position Fixes:
Error
Horizontal
Vertical
Tracker Estimates:
Error
Horizontal
Vertical
Mean
73
68
Mean
73
84
217
StandardDeviation
112
48
StandardDeviation
135
62
RMS
134
83
RMS
154
105
95%
260 ft
158 ft
95%
299 ft
198 ft
At the end of run la, the aircraft had completed one complete circuitaround the test course so a second circuit as started (run Ib) as shown inFig. 4-98. The aircraft made a low-approach at 500 feet over Hanscom Fieldand proceeded to the Shaker Hills waypoint. The run terminated near the LOBBYwaypoint when SV6 dropped below the 5° elevation angle mask. The altitude andground speed profiles for run 1b are shown in Fig. 4-99.
The satellite visibilities for run 1b are shown in Fig. 4-100. Againnote that SV9, although included in the figure, was not used for navigation.Note also at the end of the run that SV6 had dropped below the nominal 5°elevation angle mask. The dilution-of-precision values for the run are shownin Fig. 4-101.
Figure 4-102 shows the satellite GINo values for the latter portion ofrun lb. Note that the SV6 GINo value at the end of the run is above the33 dB-Hz loss-of-lock threshold. The satellite GINo does drop below thethreshold several times during the initial turn of the holding pattern but thesatellite is quickly re-established in track on each occasion.
The horizontal and vertical position fix error vs. time is shown inFig. 4-103. The corresponding tracker estimate error is shown in Fig. 4-104.The position accuracy statistics are summarized in Table 4-15. In general,the position accuracy is seen to be less for run 1b than Run 1a; however, theGnop has also become larger in the latter run as seen from Fig. 4-101.
218
..•
•
22
.5,.
....
------------------------------------..
...,
15
.0
"1" 0 ... X
NU
.1
0.0
t--'
W\0
C/) 0 0 ~
5.0
0.0
SK
R
10
5o
-5
MO
DS
EF
X1
04
-1
0-1
5-
5.0
..L.
...&
J.,.
..o.
...o...J~~
o.....L.~
..L...
oJ.
,...
o.L.
.o..~
loo.
l
-2
0
FIG
.4
-98
.O
PE
RA
TIO
NA
LT
ES
T.
9F
EB
83
.R
UN
16
N N o
t LL w o :;:) t- t ~ <I(
f/)
Il.
LL o w W Il.
f/) o z :;:) o cr CJ
46
5
TIM
ES
EC
.G
.M.T
.X
10
2
47
04
75
FIG
.4
-99
.A
LT
ITU
DE
AN
DG
RO
UN
DS
PE
ED
.9
FE
B8
3.
RU
N1
b(1
OF
2)
••
•
t IL w o :l t- t ... <
".
50
00
35g,
.L,-----------~
~L
.__
....
J
I.
.,
I•
••
••
1•
I•
+••
I•••
,•••••
t'••
,••••
I••••
I
en D.
ILtv tv t-
'0 W w
20
0D
. en 01
50
z :l
10
00 a: "
50 0 4
70
•I
••
".,
••
47
54
80
48
5
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-99
.A
LT
ITU
DE
AN
DG
RO
UN
DS
PE
ED
,9
FE
B8
3,
RU
N1
b(2
OF
2)
__
.__._
...-
.,---
..5
--._
._.--
.--..--,--._
.z o I < > W ..
JW
70
60
50
40
30
____
..._
..
.._.
....
.._
..__
....
.-.-
----
-..-
.4
_...--
_....
---_
.-_.....
.----
---_
........-
-...--_
........
.--_..
-_._
--.,=
-=,,:
._,_.~-
---.---
------
- ---
----
-.----
- ---.
(9)
20 10
....
-.
..
8
......
......
......
.. ......
......
.6
N4
00
N N
30
0
J: I- ::>2
00
:E N <1
00
----
---_
...__
._--
----
---_
.._.....
._--_
..._-
----
----
----
----
-:
----
.---
•.•.
.---
•.--
----
.-.-
.-.-
----
---.
-•.-
.:
5_
._._
.-._
_._
._._
.-._
.--'-
'-'-
'--'-
'--r'-
'-'
·---------------------+
--
(9)
· ·...
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
......
.....•
......
......
......
....
6
----------------------r---
8
46
04
65
47
04
75
48
04
85
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-10
0.
SA
TE
LL
ITE
VIS
IBIL
ITIE
S,
9F
EB
83
,R
UN
1b
•
...
tI
-
II
I•
!,
I••••
I•
'..L
---J
5.5
01
5.4
5
5.4
0Q
.0 0
5.3
5>
5.3
0
5.2
5
3.5
3.0
tvQ
.tv
0U
J0 J:
2.5
2.0
7.2
1 I7
.0
Q.
06
.80 CJ
6.6 6.4
--~
46
0
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-10
1.
DIL
UT
ION
-OF
-PR
EC
ISIO
NV
AL
UE
S,
9F
EB
83
,R
UN
1b
CIO t o< CIJ
t o< CIJ
GO~ 40~"I""~"'~
20 o
jG
Ol
::j~~~~~
GO
N N ~
t o< CIJ t o< co
40
20 o
GO
40
20 o
.L...
....L
..A..
....L......~.'~'..L'
,,.
L '~'~.~•
.L'.~.~.~,L
'.~
,L
',
L.o
'IL
..-o.
,..
.J.'l.
.o.
''
,~
,'
,~••.
..1.'
,..
.L'~
',..
.L1,...
L',
......
.J.I
47
04
75
48
04
85
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-10
2.
SA
TE
LL
ITE
C/N
O.
VA
LU
ES
,9
FE
B8
3,
RU
N1
b
••
..•
•
0: o 0:
0:
W ..J
oe( ... z o N 0: o :I:
0: o 0:
0:
W ..J
oe( o ... 0:
W >
30
0
20
0
10
0 o
-10
0
-20
0
-30
0
-40
0
-50
04
60
47
0
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-10
3.
PO
SIT
ION
FIX
ER
RO
R,
9F
EB
83
,R
UN
1b
48
0
l LL a: o a: a: w ..J~ I Z o N a: o :r:
48
04
70
o
75
0
25
0
50
0
-5
00
..L..o
-.....
.-....
........
........
........
........
........
....o
...J
L.-
-...
....
...J
.....
....
....
...-.L
......
....~~~~~~~--L.
......
......
......
......
...L
....o
-.....
.....
.....
.L.o
......
......
......
......
......
......
....o
..JL
...-..
......
.....
....
J
46
0
-25
0
l LL a: o a: a: w ..J~ o I a: W >
N N 0'
TIM
ES
EC
.G
.M.T
.X
10
2
FIG
.4
-10
4.
TR
AC
KE
RE
ST
IMA
TE
ER
RO
R,
9F
EB
83
,R
UN
1b
••
•
•
TABLE 4-15.
POSITION ACCURACY STATISTICS FOR 9 FEBRUARY 1983
RUN 1B (46030 - 48580)
Position Fixes:
•
•
Error
Horizontal
Vertical
Tracker Estimates:
Error
Horizontal
Vertical
Mean
185
153
Mean
191
159
227
StandardDeviation
117
83
StandardDeviation
142
97
RMS
219
174
RMS
238
186
95%
322 ft
293 ft
95%
378 ft
329 ft
4.3.2.4 Summary of Operational Flight Test Results
The operational flight tests demonstrated the use of the GPS CiA codenavigator in mountainous terrain, at a large urban airport and in typicalgeneral aviation operations. The position accuracy for the operational flighttests in the Hanscom Field/Manchester Airport is summarized in Table 4-16.These statistics combine the results from 4 February 1983 (Run 1a),9 February 1983 (Runs 1a and 1b) and 10 February 1983 (Run 2) into a singlerun lasting 2.7 hours. As seen from the table, the 95% horizontal accuracy ofthe tracker estimates was 333 feet, which easily meets the 0.3 nm accuracyrequirement of AC90-45A for two-dimensional area navigation. Thisperformance also essentially meets the proposed 328 foot (95%) accuracyrequirements of the Federal Navigation Plan for non-precision approach.
The operational flight tests also demonstrated the compatibility of theGPS system with current air navigation practices and procedures. The testpilots responded favorably to the operation of the system, noting inparticular that the GPS navigator gave a more stable cross-track deviationindication for non-precision approach than that provided by a VOR receiver.It was also noted that the GPS could obtain the course and distance to aselected waypoint while on the ground, a feature not always available with aconventional RNAV unit due to VOR coverage.
It was found that the GPS system was able to maintain navigation in themountainous area near the Burlington, VT Airport and in the vicinity of largebuildings at Logan International Airport. The system was able to maintainsatellites in track to elevation angles as low as 5°, even when on theground.
The system was generally able to maintain satellites in track duringturns with only temporary loss of lock. The system was able to recover fromthe temporary loss of satellite without a substantial effect on the systemposition accuracy. It was seen that the horizontal position accuracy wasmaintained when the system was forced to navigate with three satellites duringthe temporary loss of one satellite in a turn.
228
•
•
•
TABLE 4-16.
POSITION ACCURACY STATISTICS FOR OPERATIONAL FLIGHT TESTS
4-10 FEBRUARY 1983
..
Position Fixes:
StandardError Mean Deviation RMS 95%
Horizontal 98 143 174 305 ft
Vertical 103 78 129 242 ft
Tracker Estimates:
StandardError Mean Deviation RMS 95%
Horizontal 98 161 189 333 ft
Vertical 115 93 148 277 it
•
229
5.0 FUNCTIONAL REQUIREMENTS OF A GENERAL AVIATION RECEIVER
The results of testing the GPS Tests and Evaluation System evaluation ofthe GPS T and E equipment indicate that a general aviation GPS navigatorshould satisfy the following functional requirements:
5.1 General Requirements
5.1.1 Reliability
The GPS navigation system encompassing the user equipment, satellitevehicles and ground support equipment, must have a combined reliability equalto or surpassing that of alternative navigation systems. The GPS userequipment must therefore provide navigation data without operationallysignificant outages due to fades or multipath, assuming that the NAVSTARconstellation provides acceptable coverage.
5.1.2 Integrity
The GPS navigation system must not provide misleading information underany conditions of operational significance. It is therefore necessary thatthe GPS receiver continually monitor its own performance and indicate to thepilot when the navigation information displayed is no longer in compliancewith the accuracy requirements.
Further, provisions must be incorporated to allow the pilot to verifythat the navigation information is accurate using either built-in testequipment, an auxiliary test system, or a procedural check.
5.1.3 Compatibility
The GPS navigation system must provide a pilot interface which iscompatible with existing air navigation systems.
5.2 Technical Requirements
5.2.1 Link Margin
The GPS receiver link margin, for all satellites above 50 elevation angleand with nominal effective radiated power, should be at least 3 dB duringacquisition and at least 6 dB once a satellite is in track. These marginsassume the aircraft is in level flight.
5.2.2 Startup
The receiver should automatically acquire and track all visiblesatellites.
230
•
•
5.2.3 Time-to-First-Fix
The receiver should automatically provide a first fix to the navigationdisplay within 6 minutes of power on, unless the set has not been operated forlonger than one month.
5.2.4 Continuous Navigation Solutions
The GPS receiver should provide position estimates or related navigationdata derived from all satellites in view, to the navigation display at a 1 Hzrate.
5.2.5 Position Estimation
The position estimation algorithm should substitute a synthetic altitude(coast) if the HOOP, computed each fix interval, becomes unacceptably high, orif the number of satellites drops temporarily to three.
5.2.6 Performance Monitor
The performance monitor should perform the following tests:
Test
Test satellite health data.
Test pseudo-range forconsistency withprevious measurementor present prediction.
Test position estimationvariance.
Maintain coast timer.
231
Action
Delete unhealthy satellites.
Do not use inconsistent values inthe position estimator.
Set NAV FLAG to notify pilot ofbad data if variance exceedsthreshold.
Set NAV FLAG to notify pilot of baddata if coast time > T where T isselected so that an aircraftmaneuver will not cause an errorgreater than the applicable FAAaccuracy requirement.
References
1. HQ Space Division (AFSC), CPS Status Meeting, 4-6 June 1980.
2. J.C. Hansen etal, "NAVSTAR CPS Civil Applications Study", Final Report,INTRADYN (June 1979), SAMSO TR-79-63.
3. "Design Study of a Low Cost Civil Aviation CPS Receiver System", Magnavox(December 1979), NASA Contractor Report 159176.
4. FAA Advisory Circular 90 - 45A, February 21, 1975.
5. "Federal Radionavigation Plan", DOD - No. 4650.4 p. 11, Department ofTransportation and Department of Defense, (July 1980).
6. Rockwell International, "NAVSTAR CPS Space Segment/Navigation UserInterfaces," ICD-CPS-200, May 27, 1981.
7. B.D. Elrod and F.D. Natali, Stanford Telecommunications, Inc.,"Investigation of a Preliminary CPS Receiver Design for General Aviation",STI/E-TR-8022, July 1978.
8. Aerospace Corporation, "Effects of Multipath on Coherent and Non CoherentPRN Ranging Receiver", TOR-0073 (3020-03)-3, May 1973.
9. General Dynamic Electronic Division, "Multipath Analysis", DRB D9000540B,Rev. 1 (February 1974).
10. General Dynamics, "Proposal for NAVSTAR GPS, Vol. 2, User System Segment,Technical Discussion, P3631007, June 1974.
11. "Cost Development of the Dual Channel CPS Navigator for General AviationApplications", FAA Report EM-82-27, August 1982.
12. "EMC Analysis of the Civil-Use CPS Receiver on Four Specific AircraftConfigurations," DoD Electromagnetic Compatibility Analysis Center,ECAC-CR-82-048, April 1982 (Draft Report).
232
•
APPENDIX A
POSITION ESTIMATION ALGORITHM
The receiver software uses a least-squares fix which operates by batchprocessing a set of up to ten GPS pseudo-range measurements that aresequentially obtained by the receiver. The algorithm produces a new fix every2.2 seconds. The algorithm is outlined in Figs. A-I and A-2.
The algorithm begins by referencing all 10 measurements to a common timeby utilizing the pseudo-range rate estimates derived from the receivers AFCloop, i.e,
•where
t r common reference time
tk is the measurement time of the k-th satellite and
PRR and PRR are the pseudo-range and pseudo-range rate for the k-thsatellite.
The common reference time is selected to be the mid point of the2.2 second measurement batch interval in order to minimize prediction errors.Since the AFC loop has a tracking error standard deviation of 7 Hz (equivalentto a range rate of 4.1 feet per second), the measurement error contributed bythe translation process is = 4 feet, one-sigma.
The pseudo-range measurements are then corrected for propagation delayeffects using the models shown in Table A-I.
Referring to Figure A-I, let r = (x,y,z) denote the user position vectorin ECEF (earth-centered, earth-fixed) coordinates and let Ei = (xi,Yi,zi)denote the position vector of the i-th GPS satellite. The pseudo-rangemeasurements are scalars given by:
mi = I~ - Eil + 1 (1)
where 1 is the receiver clock bias, normalized by the speed of light to unitsof distance. Equation (1) is expanded results in a Taylor series about the
estimated user position, r, and estimated receiver clock time t r = t + T.
The equation is then linearized by neglecting all but first-order terms:
a~ a~~r +
ar at~b
A-I
(2)
START
COMPUTE FIX TIME ASMIDPOINT OF FIRST &LAST MEASUREMENTS
EXTRAPOLATE USERPOSITION TO FIX TIME
DO FOR EACH PRMEASUREMENT
ARE THERE MORE YESTHAN 2 VALID MEAS
IS MEASUREMENTVALID A
..
IS HOOP GREATERTHAN 10 &
NOT FIRST FIX
YES ADD PSEUDO ALTITUDE MEAS,COMPUTE MEAS GRADIENT
ARE THERE ONLY 3 YESVALID MEAS & >-----------,NOT FIRST FIX
SET POSFIXCOMPLETION FLAG
COMPUTE POSITIONFIX CORRECTION
IS POSITION FIXCORRECTIONLESS THAN O{3
READ BARO ALTIMETER I
COMPUTE PREDICTEDALTITUDE, RESIDUAL AND
MEAS GRADIENT
SET POSITION FIXAV AILABLE FLAG
•
•
FIG. A-1. POSITION FIX ALGORITHM
A-2
•
COMPUTE SV POSITIONAT TIME OF FIX
COMPUTE PREDICTED RANGEAND MEAS GRADIENT H
IS THIS FIRSTMEASUREMENT
THIS SV ELSE
YES
COMPUTE GPS TIMEOF WEEK OF SIGNAL
TRANSMISSION
COMPUTEPSEUDO RANGE
COMPENSATE FORTROPO DELAY
IS CHANGE IN PR SINCELAST MEAS :> 1000 FT
& NOT FIRST FIX
COMPUTE TIME OF WEEKOF SIGNAL TRANSMISSION
TO RESOLVE AMBIGUITY
NEXT P.R.
• PROPAGATE MEAS TOTIME OF FIX USINGMEASURED DOPPLER
COMPUTE PR RESIDUALAT FIX TIME
PR • PSEUDO RANGE
NEXT P.R.
FIG. A-1. POSrnON FIX ALGORITHM (CONT.)
A-3
..............
................. d4......
......
\
" \" \ d3
\\ -USER ~ _..--"--
.................
SV.c......
"- ....~-------I------------~ ~4
FIG. A-2. GPS POSITION FIXING GEOMETRY
A-4
r
•
•
•
TABLE A-I. SIGNAL PROPAGATION DELAY COMPENSATION MODELS
IONOSPHERIC [MODEL NOT IMPLEMENTED]
t-I4cIONO = e (E) [dI + d2 cos (2n (----)] (feet)
24
e = 1, E = 90°, e ~ 3, E = 5°
E = elevation angle to the satellite
dI and d 2 are data base parameters
t is local time in hours
TROPOSPHERIC
CTROPO = (feet)SinE
N = Sea level refractivityo
h = estimated altitude
hs Exponential scale height (22500 ft above meanearth radius)
h = estimated altitude
E satellite elevation angle
A-5
where ~r = r - r and ~b = t r - t r = T - T. The ~mi are the residualpseudo-range errors due to the user position estimate error, ~~, and receiverclock bias estimate error, ~b. Rewriting equation (1), we have:
thus
(x - xi) (y - Yi) (z - zi)~mi = -------- 6.x + -------- ~y + -------- 6.z - ~b
I~ - bl IE.. - bl IE. - bl~r
[hi! ' hi2 , hO ' -1 ] • [--]6.b
where the hij are the direction cosines from satellite i to the userposition.
When N measurements are available (N ) 4), ~m in matrix form becomes:
~r
~m = H • [--]~r
HO M
where H is an N x 4 matrix and ~m is an N-dimensional vector. Solving for ~S:
~r
6S =~b
(3)
In the special case of N=4, then (HTH)-l HT = H- 1 •
The resulting values of 6.r and ~b are used to correct the previous userposition and clock bias estimates to produce new estimates for the nextmeasurement cycle.
The batch-processing method allows the data for all satellites in view tobe utilized to produce a position fix, and obviates the need for a satelliteselection algorithm, since the position fix utilizing all N satellites willhave better CDOP than any subset of four of the satellites. Furthermore, itcan be shown that the residual error is minimized in the least-squares senseby equation (3), under the linearity assumption of equation (2).
A-6
•
•
APPENDIX B
POSITION ESTIMATION FILTER
The position estimates developed by the algorithm described in Appendix Aare processed by a filter (tracker) before they are released to the navigationsoftware.
An alpha-beta fixed gain filter was selected following a study of tracker. .
alternatives. The filter operates on X, Y, Z and t to produce X, Y, Z, X, Y,
i, t and t. The velocity gain is related to the position gain by S = 02/(2-0).Figure B-1 shows the general flow for the tracker •
B-1
START
IS THERE A POSITIONFIX AVAILABLE
CONVERT ECEFTO GEODETIC
YES
NO
PROPAGATE TRACKER STATETO CURRENT TIME
COMPUTE TRACKER UPDATEFROM POSITION FIX CHANGE
PROPAGATE TRACKER UPDATEFROM FIX TIME
TO CURRENT TIME
UPDATE TRACKER STATE
•
TRANSMIT TONAVIGATION SOFTWARE
FIG. B-1. POSITION ESTIMAoriON TRACKER
B-2
.,
•
•
,APPENDIX C
COORDINATE CONVERSION ALGORITHM
In order to compare the actual aircraft position, as estimated by theMode S Facility, with recorded GPS T&E position estimates it is necessary toto convert the GPS estimate coordinates to Mode S Facility local coordinates.The latter is described in the following paragraphs.
The three coordinate systems of concern are: (1) earth centered earthfixed (ECEF), (2) local, and (3) geodetic. The ECEF system is a Cartesiancoordinate system fixed to the earth and defined such that the z-axis pointstoward the North pole, the x-axis points out along the intersection of theGreenwich meridian and the equator. with the y-axis chosen to complete aright-handed coordinate system. A local system is constructed so that thex-axis points east. the y-axis points north and the z-axis points "up". Thegeodetic coordinates consist of longitude. latitude, and altitude above sealevel.
Conversions between these coordinate systems depend on the model used forthe surface of the earth. For this effort the WGS-72 model was selected.WGS-72 is an ellipsoid of revolution whose semi-major (equatorial) radius is6378135.0 meters, and whose semi-minor (polar) radius is 6356750.5 meters.
Conversion from local coordinates to ECEF (and inversely) requires arotation and a translation. (The rotation, in effect, aligns the coordinateaxes and the translation displaces the origin). If P is a point whosecoordinates in the local system are expressed as the three-vector quantityv = (x,y.z), then its coordinates in ECEF. V = (X,Y,Z) are given. in general,by:
V Rv + D [ 1 ]
where R is a 3x3 rotation matrix and D is the ECEF three-vector coordinate ofthe origin of the local system. The inverse conversion, is given by
v = RT (V - D)
where RT is the transpose of R (which is equal to its inverse).
[2]
To state the transformation between a local system and ECEF coordinates,let e be the longitude, A be the geodetic latitude and A' be the geocentriclatitude of the local coordinate system. Then the rotation matrix is givenby:
R =-sin e
cos eo
- cos e sin A- sin e sin A
cos A
C-l
cos e cos Asin e cos Asin A
[3]
To compute D, we need to consider the following geometry:
z
•
b
0 xI
/I
I
I/
//
Let a be the equatorial radius, b be the polar radius, and £2
2 2a - b
= ------ be2a
the numerical eccentricity. From the equation for an elipse, 1
it can be shown that:
Dz = [d (l-£2
) + h] sin A[4 ]
Dx = (d + h) cos A cos 6
D = (d + h) cos A sin 6y
where:
ad = ------------
~-g2 sin2A
and h is the height (altitude) of the local system.[4] convert local coordinates to ECEF coordinates.
Thus equations [3] and•
Using Rand D, conversion from local coordinates to ECEF coordinates andback can be accomplished. If, for example, the coordinates of a satellite aregiven in a horizon system (azimuth, elevation, and slant range) thesecoordinates can be converted to local x,y,z and then to ECEF. A difficultyarises, however, when converting the Mode S Facility reported position of atarget to its own local coordinates. The difficulty arises because targetreports give altitude rather than elevation angle. Because of this, the local
C-2
..
,
z-axis coordinate of the target is not directly known, but must be computed,and the computation depends upon the model of the earth. The exact solutionusing an ellipsoid model of the earth involves solving a fourth orderequation. A simple spherical model allows a direct calculation and inpractice has shown to be accurate to 0.1 meters to a range of 30 nm. Thegeometry and result is shown below•
h2 - h 2 + 2(h-h ) r - /ssez = --------------------------
2(re + hs )
where:
h is the reported altitude,
hs is the altitude of the local system,
r is the slant range, and
a+br e = is the average earth radius.
2
tI u. S. GOVERNMENT PRINTING OFFICE: 198"00700-000--54
C-3
top related