research article reputation-based secure sensor...

Research ArticleReputation-Based Secure Sensor Localization inWireless Sensor Networks

Jingsha He1 Jing Xu2 Xingye Zhu1 Yuqiang Zhang2 Ting Zhang2 and Wanqing Fu3

1 School of Software Engineering Beijing University of Technology Beijing 100124 China2 College of Computer Science and Technology Beijing University of Technology Beijing 100124 China3 Information Center SINOPEC Research Institute of Petroleum Processing Beijing 100083 China

Correspondence should be addressed to Jingsha He jhebjuteducn

Received 7 March 2014 Accepted 25 April 2014 Published 20 May 2014

Academic Editor Yuxin Mao

Copyright copy 2014 Jingsha He et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Location information of sensor nodes in wireless sensor networks (WSNs) is very important for it makes information that iscollected and reported by the sensor nodes spatially meaningful for applications Since most current sensor localization schemesrely on location information that is provided by beacon nodes for the regular sensor nodes to locate themselves the accuracyof localization depends on the accuracy of location information from the beacon nodes Therefore the security and reliabilityof the beacon nodes become critical in the localization of regular sensor nodes In this paper we propose a reputation-basedsecurity scheme for sensor localization to improve the security and the accuracy of sensor localization in hostile or untrustedenvironments In our proposed scheme the reputation of each beacon node is evaluated based on a reputation evaluation model sothat regular sensor nodes can get credible location information from highly reputable beacon nodes to accomplish localization Wealso perform a set of simulation experiments to demonstrate the effectiveness of the proposed reputation-based security schemeAnd our simulation results show that the proposed security scheme can enhance the security and hence improve the accuracy ofsensor localization in hostile or untrusted environments

1 Introduction

The technologies of wireless sensor networks (WSNs) arebecoming popular along with the rapid advancement ofwireless communication technology more remarkable per-formance of integrated circuits as well as decrease in costand increase in functionality of sensor nodes Since WSNsare a kind of intelligence networks that are able to integratedata collection fusion and transmission such networks havebeen widely used in fields such as military defense industrialand agricultural control urban management environmentmonitoring health care emergency rescue and disasterrelief In addition sensor networks also have a broad prospectof applications in tracking logistics management and spaceexploration Depending on different application scenariosin the above areas researchers have put forward some newtechnology and strategy such as sensor deployment methodssuitable for underwater detection [1] and intelligent moni-toring technologies in Smart Home scenarios [2] In short

the applications of WSNs are being developed to achieveubiquity that can bring more convenience for human beingsin many areas

In most applications sensor nodes are used to collectphysical data such as temperature humidity water levelpressure andwind speed that are sent alongwith the locationinformation to the data center to ensure that the collecteddata have spatial meaning Furthermore the location infor-mation of sensor nodes can also serve as the basis for somenetwork functions such as network configuration and real-time statistics of network coverage Therefore in massivelydeployed WSNs location information of sensor nodes isvery important for enabling many applications which makessensor localization one of the basic services and a coretechnology for WSNs

Since sensor localization in wireless sensor networks(WSNs) is a fundamental technical issue and is criticalfor monitoring applications and for most location-basedrouting protocols and services research in sensor localization

Hindawi Publishing Corporatione Scientific World JournalVolume 2014 Article ID 308341 10 pageshttpdxdoiorg1011552014308341

2 The Scientific World Journal

technology has generated a wide spread interest and variousissues on different aspects have been studied which includeefficiency [3] accuracy [4] and security [5] amongmany hotissues in sensor localization

Current algorithms for sensor localization fall into twocategories range-free algorithms [6] and range-based algo-rithms [7] In a range-free algorithm such as Centroid [8] orCTDV-Hop [9] a node estimates its location using informa-tion of connectivity between different nodes In a range-basedalgorithm a sensor node estimates its own location based oninformation about distances or angles between sensor nodesand through using techniques such as time of arrival (TOA)[10] time difference of arrival (TDOA) [11] received signalstrength indicator (RSSI) [12] and angle of arrival (AOA)[13] as well as methods such as trilateration triangulationor maximum likelihood estimation [14] Among the manydifferent sensor localization algorithms RSSI-based position-ing technology is perhaps the most popular due to its lowcost and easy implementation On the other hand sensorlocalization results can be greatly affected bymalicious nodesin hostile or untrusted environments This is because sensornodes can hardly perform accurate localization if they uselocation information that is provided by untrusted beaconnodes Security in sensor localization has thus received agreat deal of attention along with the development of sensorlocalization technologies for WSNs

In the past few years researchers have proposed severalsecurity strategies for sensor localization from differentaspects Some of the methods implement verification mea-sures to reduce the impact of using unreliable or false locationinformation [15] while some others apply a series of schemesin which temporal spatial and consistent properties areconsidered to deal with distance-consistent spoofing attacks[16] However in these schemes sensor nodes are dividedinto just two types secure and insecure sensor nodes throughthe mechanisms of comparing the nodes and their behavioragainst normal situations However such an approach cannotbe very objective which could cause many false positive andfalse negative results

Meanwhile some other researchers have proposed local-izationmethods that are able to fight against attacks launchedby compromised sensor nodes a problem that is moredifficult to deal with Liu et al proposed robust computingalgorithms to improve the reliability of localization schemes[17] Park and Shin proposed an attack-tolerant localizationprotocol that would perform adaptive management of aprofile for normal localization behavior [18] However thelimitation of these schemes is that they did not consider thesecurity of sensor localization when sensor nodes are joiningand leaving the network along with the passage of time Inaddition they did not pay enough attention to secure sensorlocalization in dynamic wireless networks

As an effective means of ensuring security the notion ofreputation has been introduced and some reputation-basedschemes have since been proposed for sensor localizationSrinivasan et al proposed a distributed reputation-basedbeacon trust system [19] and Xu et al proposed a reputation-based revising scheme for sensor localization which wouldincur high computation cost [20] However complicated

reputation evaluation in the above schemes for sensor local-ization makes it necessary to further improve the efficiencyof evaluation for beacon nodes Any sensor localizationmethod that can achieve good performance should ensure thereliability of location information before such informationcan be actually used for sensor localization

In real applications there may be other types of sensorlocalization methods to fit different application scenariosTherefore specific localization methods in real applicationsneed to be continuously developed and improved basedon orientation methods in order to adapt basic sensorlocalization schemes to the many different network sce-narios Consequently in order to develop effective sensorlocalization methods we should analyze and understandthe main characteristics of specific networks and developproper performance metrics that can be used to measure theperformance of sensor localization schemes In addition weshould also consider limitations of wireless sensor networkssuch as constrained energy supply in the sensor nodes aswell as the complexity of network environments in thedevelopment of effective sensor localization methods

In this paper we propose a novel reputation-based securesensor localization scheme to improve the accuracy of sensorlocalization for WSNs in hostile environments In the pro-posed reputation model the reputation of each beacon nodeis evaluated by each other to ensure that sensor nodes will getcredible location information to perform sensor localizationThe proposed scheme can therefore effectively reduce theimpact of malicious beacon nodes on the localization ofregular sensor nodes by relying on the security mechanismof beacon node evaluation Our simulation results showthat the proposed reputation-based secure sensor localizationscheme can improve the accuracy of sensor localization inhostile or untrusted environments In addition the proposedsecure sensor localization scheme possesses the desirablecharacteristics of expandability and flexibility since it can beused in both static and dynamic networks

The remainder of this paper is structured as follows InSection 2 we present a reputation model in which we firstdescribe the network model and then propose a reputa-tion evaluation model In Section 3 we present our sensorlocalization scheme which is based on the evaluation ofthe reputation of beacon nodes In Section 4 we describethe simulation that we have performed and present thesimulation results Finally in Section 5 we conclude thispaper in which we also discuss some future work

2 The Reputation Model

In hostile network environments which most current WSNdeployments would assume regular sensor nodes need tobe confronted with security threats during the process ofsensor localization If a sensor node can identify the securityand credibility of location information that it receives andsubsequently use the information appropriately the accuracyof sensor localization can be greatly improved or ensured insuch environments Therefore in order to develop effectivesensor localization schemes we should understand the main

The Scientific World Journal 3

characteristics of the specific networks as well as the per-formance goals of the localization schemes To achieve theabove objective we need to consider such characteristics asresource constraints in the sensor nodes and the complexityof the environment where the sensor nodes are deployedAny sensor localization scheme must be effectively workingin a specific WSN after the above-mentioned factors areconsidered in the design

To achieve the above goal we first propose a reputationscheme to be used in the sensor localization scheme we willpropose later in this paper to deal with a hostile deploymentenvironment in which malicious nodes can be dropped intothe network at will and regular sensor nodes can also be easilycompromised to make them behave in a malicious mannerWe call the scheme that we propose the reputation-basedlocalization scheme (RBL) The main characteristics of thereputation model and the RBL are as follows

(1) The proposed secure sensor localization scheme isdeveloped based on a reputation model and on theevaluation of all the beacon nodes for deriving areputation value for each and every beacon node

(2) In the reputation model the reputation of eachbeacon node is evaluated and consequently used byregular sensor nodes to determine the credibilityof the location information provided by the beaconnode

(3) In the reputation model the reputation of eachbeaconnode is updated continuouslywith the passageof time if sensor localization needs to be carried outfrom time to time

In the following sections we will first describe the net-work model followed by the threat model and the reputationmodel

21 The Network Model The WSN under considerationis composed of beacon nodes and regular sensor nodesBeacon nodes are capable of positioning themselves (eg bydetermining their positions through GPS) while the regularsensor nodes need to locate their own positions based onposition information from other nodes especially from thebeacon nodes

Our sensor localizationmethod in this paper requires thata regular sensor node first estimates its relative position tosome of the beacon nodes through the means of receivingsignals from creditable beacon nodes and by computing thedistances between them using a signal attenuation formulaThen the sensor node estimates its position using the maxi-mum likelihood estimation method [21] after it has collectedenough position information

22 The Threat Model An analysis of the network modeldescribed above indicates that position information receivedfrom beacon nodes and the estimation of relative positionsbetween a regular sensor node and the referenced beaconnodes can determine the accuracy of sensor localizationThere are however two primary types of security threats forthe network model as described below

U998400

1

B998400

2

U1

B2

B1

B3

Figure 1 An example of sensor localization in hostile environments

(1) Sending false beacon information if malicious bea-con nodes send false position information suchinformation received by regular sensor nodesmay notbe accurate Then the estimated position of a regularsensor node cannot be guaranteed to be accurate andwill lose its credibility because it is calculated basedon received false information frombeacon nodesTheimpact to sensor localization from this type of attacksis shown in Figure 1 We can see from the figure that1198611015840

2is the false position of beacon node 119861

2 which

makes sensor node 1198801receive a false localization

result 11988010158401

(2) Obstructing physical property if malicious nodesinterfere with normal signals from beacon nodes noregular sensor node would be able to estimate itsrelative position to the beacon nodes accurately bythe means of signal attenuation leading to reducedaccuracy for sensor localization

Consequently the scheme that we propose in this paperneeds to deal with the potential threats that result from theabove two types of attacks in order to improve the accuracyof sensor localization in WSNs

23 The Proposed Reputation Model To deal with the abovesecurity threats we propose a novel reputation model forsensor localization inWSNs In the reputationmodel beaconnodes evaluate each other using information such as thecharacteristics about the perception of positions and providethe evaluation results to the regular sensor nodes Theregular sensor nodes use the evaluation results provided bythe beacon nodes to rank the beacon nodes and base thecredibility of the location information provided by beaconnodes on such ranking

First let us make the following assumptions in ourreputation model

(1) The reputation value for each and every beacon nodeis a number between 0 and 1 indicating values fromthe lowest to the highest reputations


(2) The reputation value for each and every beacon nodeis initialized to be 05 a medium value to start with

(3) In the reputation model each beacon node performsevaluation only on its neighboring beacon nodes thatis the beacon nodes that are one hop away from it

Pseudocode 1 contains the pseudocode for our proposedreputation model for beacon node 119861

119895and sensor node 119880

119898

The details of the evaluation procedure in the proposedreputation model are as follows

(1) Beacon node 119861119894sends its coordinate (119909

119894 119910119894) to its

neighboring beacon nodes(2) Each neighboring beacon node to 119861

119894will calculate

its distance to 119861119894using the received coordinate infor-

mation and the signal strength information inde-pendently Let 119897

119861119895119894denote the distance between 119861

119895

and 119861119894based on the coordinate information and let

119889119861119895119894

denote the distance based on the signal strengthinformation 119861

119895can then calculate 119897

119861119895119894using the

coordinate information from 119861119894and calculate 119889

119861119895119894

through a signal strength ranging algorithm based onthe strength of the signals received from 119861

119894

(3) All the neighboring beacon nodes evaluate the rep-utation of 119861

119894 The value of reputation evaluation is

determined using (1) in which 119877119905119861119895119894

and 119877119905+Δ119905119861119895119894

denotethe reputation values on 119861

119894by 119861119895at times 119905 and 119905 +Δ119905

respectively and Δ119905 denotes the time interval of tworeputation values Let Δ119889 be the threshold for thedistance that is the error that can be tolerated forthe distance and let 120572 be the weight of the evaluationvalue which is determined using (2)

119877119905+Δ119905

119861119895119894= 120572 times 119877

119905

119861119895119894+ (1 minus 120572)

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816le Δ119889

119877119905+Δ119905

119861119895119894= (1 minus 120572) times 119877

119905

119861119895119894

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816gt Δ119889

(1)

120572 =

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816

119897119861119895119894+ 119889119861119895119894

(2)

(4) The neighboring sensor nodes get the reputationvalues for 119861

119894from the beacon nodes Each regular

sensor node collects the evaluation values from all theneighboring beacon nodes and computes the averagereputation value using (3) in which 119877119905+Δ119905

119880119898 119861119894and 119877119905+Δ119905

119861119896119894

denote the reputation value on beacon node 119861119894from

a sensor node 119880119898and that on 119861

119894evaluated by 119861

119896at

time 119905 + Δ119905 where 119861119896is a neighboring beacon node

to 119861119894and 119899 is the number of such neighboring nodes

Consider

119877119905+Δ119905

119880119898 119861119894=

sum119899

119896=1119877119905+Δ119905

119861119896119894

119899 (3)

(5) Every regular sensor node ranks the neighboringbeacon nodes from high to low based on the receivedreputation values

3 The Sensor Localization Scheme

The sensor localization scheme in this paper uses theproposed reputation evaluation scheme described above inwhich the reputation model is relied upon by the beaconnodes to evaluate each other In the illustration below we usethe RSSI ranging technology for sensor localization althoughthe same reputation scheme can be applied equally to TOATDOA and AOA ranging methods in practical applications

After receiving the evaluation results a regular sensornode will select credible beacon nodes based on the repu-tation values Afterwards the sensor node will measure thedistance to the credible beacon nodes using the RSSI rangingtechnology and estimate its location through maximumlikelihood estimation The main steps in our localizationscheme are described as follows

(i) Every beacon node provides its location informationto all the neighbor nodes As shown in Figure 2beacon node 119861

1sends its position coordinate (119909

1 1199101)

to all the neighboring nodes

(ii) Beacon nodes in the network will evaluate each otherusing the proposed reputation model and each willsend its evaluation results to all the neighboringnodes As shown in Figure 2 beacon nodes 119861

2 1198613 1198614

and 1198615evaluate the reputation of 119861

1after receiving

the location information from 1198611using the proposed

reputation model and each will send the evaluationresult to their neighboring sensor nodes includingnode 119880

1

(iii) Each regular sensor node will select credible bea-con nodes based on the results from the reputationevaluation Sensor node 119880

1computes the reputa-

tion value for 1198611and collects the reputation values

from neighboring beacon nodes using (3) Then 1198801

ranks the neighboring beacon nodes according tothe reputation values in the order of high to lowbased on which it selects the credible beacon nodesaccordingly

(iv) Regular sensor nodes estimate their relative positionsto the credible beacon nodes using the signal attenu-ation formula in RRSI [12]

(v) Regular sensor nodes calculate their coordinatesusing maximum likelihood estimation Supposethat the number of credible neighboring beaconnodes around 119880

1is 119901 with coordinates (119909

1 1199101)

(1199092 1199102) (119909

119901 119910119901) respectively and the distances

between 1198801(1199091198801 1199101198801) and the beacon nodes are

1198891 1198892 119889

119901 respectively then the position of 119880

1

can be calculated using the following

(1199091198801minus 119909119894)2

+ (1199101198801minus 119910119894)2

= 1198892

119894 119894 = 1 2 119901 (4)

In addition 119901 distance equations about 1198801and the 119901

beacon nodes are listed as in (5) that result from subtracting


Beacon nodes evaluate each other between neighbor beacon nodes119877119905

119861119895119894= 05

while(true)

120572 =

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816

119897119861119895119894+ 119889119861119895119894

if100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816le Δ119889

119877119905+Δ119905

119861119895119894= 120572 times 119877

119905

119861119895119894+ (1 minus 120572)

else119877119905+Δ119905

119861119895119894= (1 minus 120572) times 119877

119905

119861119895119894

sleep(Δ119905) Beacon nodes send the reputation value to their neighbor sensor nodesSend (node 119861

119895 node 119880

119898 reputation value)

Sensor nodes compute the reputation value of their neighbor beacon nodes

119877119905+Δ119905

119880119898 119861119894=

sum119899

119896=1119877119905+Δ119905

119861119896119894

119899

Pseudocode 1 Pseudocode for the reputation model

B2

U2

U3

U6

B5

B4

B3

B1

U10

U9

U7

U5

U1

U4

U8

Figure 2 The network topology

the last equation from each of the first 119901 minus 1 equationsConsider

1199092

1minus 1199092

119901minus 2 (119909

1minus 119909119901) 1199091198801+ 1199102

1minus 1199102

119901

minus 2 (1199101minus 119910119901) 1199101198801= 1198892

1minus 1198892

119901

sdot sdot sdot

1199092

119901minus1minus 1199092

119901minus 2 (119909

119901minus1minus 119909119901) 1199091198801+ 1199102


119901

minus 2 (119910119901minus1minus 119910119901) 1199101198801= 1198892


119901

(5)

1198801(1199091198801 1199101198801) can then be calculated using the following

1198801= 119860minus1119887 (6)

Thematrices in (6) can then be expressed as the followingexpressions

119860 = 2[

[

1199091minus 119909119901

1199101minus 119910119901

sdot sdot sdot

119909119901minus1minus 119909119901119910119901minus1minus 119910119901

]

]

119887 =[[

[

1199092

1minus 1199092

119901+ 1199102

1minus 1199102

119901minus 1198892

1+ 1198892

119901

sdot sdot sdot

1199092


119901+ 1199102


119901minus 1198892

119901minus1+ 1198892

119901

]]

]

1198801= [1199091198801

1199101198801

]

(7)

The final solution to (6) can be obtained using thefollowing

119880 = (119860119879119860)minus1

119860119879119887 (8)

From the above steps we can see that a regular sensornode would treat the location information from neighboringbeacon nodes differently according to the result of reputationevaluation There is no need to determine the positionrelationship between regular sensor nodes and beacon nodesthat have low reputation values which are required in thesignal attenuation formula resulting in reducing a certainamount of computational overhead

4 Simulation and Analysis

We have performed some simulation on wireless sensorslocalization with the proposed RBL to evaluate the perfor-mance of the scheme

The network configuration for the simulation is set upas follows The regular sensor nodes and beacon nodesare deployed randomly in an area of 650m times 600m Thetransmission radius of each beacon and sensor node is set


0 5 10 15 20

00

05

10

15

20

25Lo

caliz

atio

n er

ror

ID of the regular sensor nodes

Without reputation evaluationWith reputation evaluation

(a)


0 5 10 15 20000204060810121416182022

Loca

lizat

ion

erro

r


(b)

Figure 3 Sensor localization error with a different number of malicious beacon nodes (a) 10 normal and 5 malicious beacon nodes (b) 10normal and 10 malicious beacon nodes

at 200m There exist some malicious beacon nodes thatrandomly send out false location information

Localization error is one important indicator of the per-formance in sensor localization forWSNs which is calculatedusing (9) In the formula (119909

119880119898 119910119880119898) and (1199091015840

119880119898 1199101015840

119880119898) denote

themeasured coordinates and the actual coordinates for node119880119898 respectively 119877 denotes the transmission radius of the

nodes and 119890119898is the localization error Consider

119890119898=

radic(119909119880119898minus 1199091015840

119880119898)2

+ (119910119880119898minus 1199101015840

119880119898)2

119877

(9)

The localization error from the simulation for 20 sensornodes is shown in Figure 3 We can see from the figurethat reputation evaluation is effective for reducing localiza-tion error in hostile environments and the improvement ismore significant as the number of malicious beacon nodesincreases

There are two types of threats in sensor localizationattacks targeted at the nodes and attacks targeted at thelocation information RBL evaluates the credibility of beaconnodes by evaluating the location information that beaconnodes provide in order to reduce the influence of compro-mised beacon nodes on localization results and to resist thethreat of location information tampering by the maliciousbeacon nodes To measure the capability of RBL on counter-ing the above security threats in our evaluation we deploy 40regular sensor nodes to expand the scale of our experimentin which wemeasure the average localization error using (10)where 119873 denotes the number of regular sensor nodes in thenetwork The average localization error from our simulationon a network in which there exist one or more compromisedbeacon nodes is shown in Figure 4We can see from the figurethat although the average localization error fluctuates withthe number and the locations of the regular sensor nodes

the result of RBL is much better than that of the primarylocalization scheme (PLS) using RRSI in which no evaluationof beacon nodes is performed Consider

119890 =sum119873

119894=1119890119894

119873 (10)

Since WSNs possess the characteristics of dynamicnetwork topology an advanced secure sensor localizationscheme should not only be able to ensure the security ofsensor localization in a static network but also be able tohandle the cases of nodes joining leaving and removing fromthe networkWe have performed some simulations on sensorlocalization for the above scenarios

In order to expand the coverage of beacon nodes ina network so as to make more regular sensor nodes theneighbors of the beacon nodes in the network consequentlyimproving the utilization of beacon information we canincrease the signal transmission power of the beacon nodesto effectively expand the signal transmission radius of thebeacon nodes

In the simulations we first deploy 20 regular nodes and 4normal beacon nodes in the area Then we add more beaconnodes into the network at the rate of one node per minutestarting at the moment of 15min with normal and maliciousbeacon nodes being added alternately Malicious beaconnodes that are added into the network would send out falseposition information randomly while normal beacon nodesalways send out their real position information Figure 5shows the average localization error for regular sensor nodesduring the first sevenminutes fromwhich we can see that theaverage localization error for regular sensor nodes fluctuatesnoticeably in the primary localization scheme but exhibits agood performance in our proposed RBL

We have also performed some simulations to evaluate theimpact of nodes leaving the network on sensor localization


0 5 10 15 20 25 30 35 40 45000002004006008010012014016018020022024026028030032034036038040

Aver

age l

ocal

izat

ion

erro

r

Number of the regular sensor nodes

PLSRBL

Figure 4 Average sensor localization error with a varying numberof regular sensor nodes

0 1 2 3 4 5 6 700

02

04

06

08

10

12

14

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

Figure 5 Average sensor localization error as beacon nodes areadded into the network

In the simulation we first deploy 20 regular nodes 4 normalbeacon nodes and 4 malicious beacon nodes in the areaThen we remove the beacon nodes from the network atthe rate of one node per 1 minute starting at the momentof 15min with normal and malicious beacon nodes beingremoved alternately Again normal beacon nodes alwaysclaim their real positions while malicious beacon nodeswould send out false position information randomly Figure 6shows the average sensor localization error for regular sensornodes during the process from which we can see that theaverage localization error of RBL is noticeably lower thanthat of PLS in most cases However when the number ofnormal beacon nodes falls below three in the whole network

0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

Figure 6 Average sensor localization error as beacon nodes areremoved from the network

Table 1 State situations for the beacon nodes

Categories 1 2 3 4Situations 119877119875

1cap 1198621198752

1198771198752cap 1198621198751

1198771198752cap 1198621198752

1198771198752cap 1198621198753

the advantage would disappear which seems to be a limita-tion of the current RBL

Lastly we evaluate the impact of status change amongthe existing beacon nodes on regular sensor nodes under theassumption that the total number of beacon nodes remainsthe same Four possibilities exist for such status change asillustrated in Table 1 in which 119877119875

1and 119877119875

2represent the

cases in which a beacon node does not change its realposition and changes its real position respectively and 119862119875

1

1198621198752 and 119862119875

3represent the cases in which a beacon node

does not change its claimed position information changesits claimed position information randomly and changes itsclaimed position information consistently respectively

We perform evaluations for four scenarios In the firstevaluation we deploy 20 regular sensor nodes and 8 normalbeacon nodes in the area and then change the status of 4beacon nodes to the state that corresponds to situation 1 inTable 1 gradually during a 4-minute time period starting atthe moment of 15min In the second evaluation we deploy20 regular sensor nodes and 8 normal beacon nodes in thearea and then change the status of 4 beacon nodes to the statethat corresponds to situation 2 in Table 1 gradually duringa 4-minute time period starting at the moment of 15minIn the third evaluation we deploy 20 regular sensor nodesand 8 normal beacon nodes in the area and then changethe status of 4 beacon nodes to the state that corresponds tosituation 3 in Table 1 gradually during a 4-minute time periodstarting at the moment of 15min In the last evaluation wedeploy 20 regular sensor nodes and 8 normal beacon nodes in


0 1 2 3 4 500

02

04

06

08

10

12

14

16

18Av

erag

e loc

aliz

atio

n er

ror

Time (min)

PLSRBL

(a)

0 1 2 3 4 500

02

04

06

08

10

12

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(b)

0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(c)

0 1 2 3 4 500

01

02

03

04

05

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(d)

Figure 7 Average sensor localization error as beacon nodes change their status in various scenarios

the area and then change the status of 4 beacon nodes tothe state that corresponds to situation 4 in Table 1 graduallyduring a 4-minute time period starting at the moment of15min

Figures 7(a) 7(b) 7(c) and 7(d) show the results of theevaluations that correspond to the above four evaluationscenarios We can see from the figure that RBL can effectivelyfilter out abnormal (or malicious) beacon nodes when someof the beacon nodes change their status in an unpredictablemanner which demonstrates that RBL is an effective schemefor secure sensor localization for WSNs which clearly showsthat RBL can improve the accuracy of sensor localization inhostile or untrusted environments

In summary so far we have performed three setsof simulation experiments to verify the performance and

the effectiveness of the proposed security scheme for sen-sor localization in hostile or untrusted environments Inthe first one we evaluated the performance of localizingregular sensor nodes in the presence of a varying numberof malicious beacon nodes In the second one we evaluatedthe average sensor localization error for different numbersof regular sensor nodes in hostile or untrusted environmentIn the third one we evaluated sensor localization resultswhen new beacon nodes join the network existing beaconnodes leave the network and existing beacon nodes changetheir status that determines how they would make claimson their positions It is clear that the purpose of the lastexperiment is to evaluate the influence on the localizationof regular sensor nodes due to changes on the credibilityof the beacon nodes That is the first two experiments are


mainly aimed at showing the performance of RBL on securesensor localization in static WSNs while the third one isaimed at verifying the effectiveness of RBL on secure sensorlocalization in dynamic WSNs All the simulation resultsthat we have obtained clearly show that RBL can reduce theeffect ofmalicious beacon nodes on the localization of regularsensor node thus allowing us to conclude that RBL caneffectively improve the security and the accuracy of sensorlocalization inWSNsThe experiments also indicate that RBLcan scale well with the size of the network and can be appliedin dynamicWSNs especially when new sensor nodes can joinand existing sensor nodes can leave networkswith the passageof time

5 Conclusion

In this paper we proposed a novel reputation model forregular sensor nodes to evaluate the credibility of beaconnodes in sensor localization In the model beacon nodesfirst evaluate each other and then provide the evaluationresults to regular sensor nodes for them to determine thecredibility of beaconnodes to ensure that theywill receive anduse credible position information from the beacon nodes inlocating their own positions The proposed security schemecan improve the accuracy of sensor localization in hostileor untrusted environments The scheme can help to ensurethe reliability of received location information under thescenario of signal attenuation by minimizing the effects offalse location information as well as interfering signals causedby malicious beacon nodes

In the future we will extend our security scheme tocounter other types ofmalicious attacks in sensor localizationwithout incurring too much additional computational costand communication overhead and to apply our reputation-based sensor localization scheme to different network envi-ronments to further verify and improve the scheme We willalso study the impact on evaluation due to other factorsof sensor nodes to further improve the performance andusability of our secure sensor localization scheme in WSNs

Notations

119861119894 Beacon node 119894

119880119898 Sensor node119898

Δ119905 The time interval of the two reputationvalues

119897119861119895119894 The distance between beacon node 119895 and 119894

based on the coordinate information119889119861119895119894 The distance between beacon node 119895 and 119894

based on the ranging techniques (such asreceived signal strength indicator)

119877119905

119861119896119894 The reputation value on beacon node 119894

from a beacon node 119896119877119905

119880119898 119861119894 The reputation value on beacon node 119894

from a sensor node119898 at time 119905(119909119901 119910119901) The coordinate of beacon node 119901

(119909119880119898 119910119880119898) The coordinate of sensor node119898

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

The work in this paper has been supported by NationalNatural Science Foundation of China (Grant no 61272500)andBeijingNatural Science Foundation (Grant no 4142008)

References

[1] E F Golen B Yuan and N Shenoy ldquoAn evolutionary approachto underwater sensor deploymentrdquo International Journal ofComputational Intelligence Systems vol 2 no 10 pp 184ndash2012009

[2] J C Augusto J Liu P McCullagh H Wang and J-B YangldquoManagement of uncertainty and spatio-temporal aspects formonitoring and diagnosis in a smart homerdquo InternationalJournal of Computational Intelligence Systems vol 1 no 4 pp361ndash378 2008

[3] S-K Yang and K-F Ssu ldquoAn energy efficient protocol for targetlocalization in wireless sensor networksrdquo World Academy ofScience Engineering and Technology vol 56 no 8 pp 398ndash4072009

[4] M Boushaba A Hafid and A Benslimane ldquoHigh accuracylocalization method using AoA in sensor networksrdquo ComputerNetworks vol 53 no 18 pp 3076ndash3088 2009

[5] R Sugihara and R K Gupta ldquoSensor localization with deter-ministic accuracy guaranteerdquo in Proceedings of the IEEE INFO-COM pp 1772ndash1780 April 2011

[6] J Park Y Lim K Lee and Y-H Choi ldquoA polygonal methodfor ranging-based localization in an indoor wireless sensornetworkrdquoWireless Personal Communications vol 60 no 3 pp521ndash532 2011

[7] Y W E Chan and B H Soong ldquoA new lower bound on range-free localization algorithms in wireless sensor networksrdquo IEEECommunications Letters vol 15 no 1 pp 16ndash18 2011

[8] N Bulusu J Heidemann and D Estrin ldquoGPS-less low-costoutdoor localization for very small devicesrdquo IEEE PersonalCommunications vol 7 no 5 pp 28ndash34 2000

[9] HWuMDeng L XiaoWWei andAGao ldquoCosine theorem-based DV-hop localization algorithm in wireless sensor net-worksrdquo Information Technology Journal vol 10 no 2 pp 239ndash245 2011

[10] I Guvenc and C-C Chong ldquoA survey on TOA based wirelesslocalization andNLOSmitigation techniquesrdquo IEEE Communi-cations Surveys and Tutorials vol 11 no 3 pp 107ndash124 2009

[11] A Savvides C-C Han and M B Strivastava ldquoDynamicfine-grained localization in ad-hoc networks of sensorsrdquo inProceedings of the 7th Annual International Conference onMobile Computing and Networking pp 166ndash179 July 2001

[12] P Bahl and V N Padmanabhan ldquoRADAR an in-building RF-based user location and tracking systemrdquo in Proceedings ofthe 19th Annual Joint Conference of the IEEE Computer andCommunications Societies pp 775ndash784 March 2000

[13] D Niculescu and B Nath ldquoAd hoc positioning system (APS)using AOArdquo in Proceedings of the 22nd Annual Joint Conferenceon the IEEE Computer and Communications Societies pp 1734ndash1743 April 2003


[14] Y Zhang L Bao S-H Yang MWelling and DWu ldquoLocaliza-tion algorithms for wireless sensor retrievalrdquoComputer Journalvol 53 no 10 pp 1594ndash1605 2010

[15] S Capkun K B Rasmussen M Cagalj and M SrivastavaldquoSecure location verification with hidden and mobile basestationsrdquo IEEE Transactions on Mobile Computing vol 7 no 4pp 470ndash483 2008

[16] H Chen W Lou and Z Wang ldquoA novel secure localizationapproach in wireless sensor networksrdquo EURASIP Journal onWireless Communications and Networking vol 2010 pp 1ndash122010

[17] D Liu P Ning A Liu C Wang and W K Du ldquoAttack-resistant location estimation in wireless sensor networksrdquoACMTransactions on Information and System Security vol 11 no 7pp 22ndash39 2008

[18] T Park and K G Shin ldquoAttack-tolerant localization via iterativeverification of locations in sensor networksrdquo Transactions onEmbedded Computing Systems vol 8 no 12 pp 1ndash24 2008

[19] A Srinivasan J Teitelbaum and W Jie ldquoDRBTS distributedreputation-based beacon trust systemrdquo in Proceedings of the 2ndIEEE International Symposium on Dependable Autonomic andSecure Computing (DASC rsquo06) pp 277ndash283 October 2006

[20] X Xu H Jiang L Huang H Xu and M Xiao ldquoA reputation-based revising scheme for localization in wireless sensor net-worksrdquo in Proceedings of the IEEEWireless Communications andNetworking Conference (WCNC rsquo10) pp 1ndash6 April 2010

[21] R-I Rusnac and A S Gontean ldquoMaximum Likelihood Esti-mation Algorithm evaluation for wireless sensor networksrdquo inProceedings of the 12th International Symposiumon Symbolic andNumeric Algorithms for Scientific Computing (SYNASC rsquo10) pp95ndash98 September 2010

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Active and Passive Electronic Components

Control Scienceand Engineering

Journal of



RotatingMachinery


Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics


Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of


Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation



DistributedSensor Networks



technology has generated a wide spread interest and variousissues on different aspects have been studied which includeefficiency [3] accuracy [4] and security [5] amongmany hotissues in sensor localization

Current algorithms for sensor localization fall into twocategories range-free algorithms [6] and range-based algo-rithms [7] In a range-free algorithm such as Centroid [8] orCTDV-Hop [9] a node estimates its location using informa-tion of connectivity between different nodes In a range-basedalgorithm a sensor node estimates its own location based oninformation about distances or angles between sensor nodesand through using techniques such as time of arrival (TOA)[10] time difference of arrival (TDOA) [11] received signalstrength indicator (RSSI) [12] and angle of arrival (AOA)[13] as well as methods such as trilateration triangulationor maximum likelihood estimation [14] Among the manydifferent sensor localization algorithms RSSI-based position-ing technology is perhaps the most popular due to its lowcost and easy implementation On the other hand sensorlocalization results can be greatly affected bymalicious nodesin hostile or untrusted environments This is because sensornodes can hardly perform accurate localization if they uselocation information that is provided by untrusted beaconnodes Security in sensor localization has thus received agreat deal of attention along with the development of sensorlocalization technologies for WSNs

In the past few years researchers have proposed severalsecurity strategies for sensor localization from differentaspects Some of the methods implement verification mea-sures to reduce the impact of using unreliable or false locationinformation [15] while some others apply a series of schemesin which temporal spatial and consistent properties areconsidered to deal with distance-consistent spoofing attacks[16] However in these schemes sensor nodes are dividedinto just two types secure and insecure sensor nodes throughthe mechanisms of comparing the nodes and their behavioragainst normal situations However such an approach cannotbe very objective which could cause many false positive andfalse negative results

Meanwhile some other researchers have proposed local-izationmethods that are able to fight against attacks launchedby compromised sensor nodes a problem that is moredifficult to deal with Liu et al proposed robust computingalgorithms to improve the reliability of localization schemes[17] Park and Shin proposed an attack-tolerant localizationprotocol that would perform adaptive management of aprofile for normal localization behavior [18] However thelimitation of these schemes is that they did not consider thesecurity of sensor localization when sensor nodes are joiningand leaving the network along with the passage of time Inaddition they did not pay enough attention to secure sensorlocalization in dynamic wireless networks

As an effective means of ensuring security the notion ofreputation has been introduced and some reputation-basedschemes have since been proposed for sensor localizationSrinivasan et al proposed a distributed reputation-basedbeacon trust system [19] and Xu et al proposed a reputation-based revising scheme for sensor localization which wouldincur high computation cost [20] However complicated

reputation evaluation in the above schemes for sensor local-ization makes it necessary to further improve the efficiencyof evaluation for beacon nodes Any sensor localizationmethod that can achieve good performance should ensure thereliability of location information before such informationcan be actually used for sensor localization

In real applications there may be other types of sensorlocalization methods to fit different application scenariosTherefore specific localization methods in real applicationsneed to be continuously developed and improved basedon orientation methods in order to adapt basic sensorlocalization schemes to the many different network sce-narios Consequently in order to develop effective sensorlocalization methods we should analyze and understandthe main characteristics of specific networks and developproper performance metrics that can be used to measure theperformance of sensor localization schemes In addition weshould also consider limitations of wireless sensor networkssuch as constrained energy supply in the sensor nodes aswell as the complexity of network environments in thedevelopment of effective sensor localization methods

In this paper we propose a novel reputation-based securesensor localization scheme to improve the accuracy of sensorlocalization for WSNs in hostile environments In the pro-posed reputation model the reputation of each beacon nodeis evaluated by each other to ensure that sensor nodes will getcredible location information to perform sensor localizationThe proposed scheme can therefore effectively reduce theimpact of malicious beacon nodes on the localization ofregular sensor nodes by relying on the security mechanismof beacon node evaluation Our simulation results showthat the proposed reputation-based secure sensor localizationscheme can improve the accuracy of sensor localization inhostile or untrusted environments In addition the proposedsecure sensor localization scheme possesses the desirablecharacteristics of expandability and flexibility since it can beused in both static and dynamic networks

The remainder of this paper is structured as follows InSection 2 we present a reputation model in which we firstdescribe the network model and then propose a reputa-tion evaluation model In Section 3 we present our sensorlocalization scheme which is based on the evaluation ofthe reputation of beacon nodes In Section 4 we describethe simulation that we have performed and present thesimulation results Finally in Section 5 we conclude thispaper in which we also discuss some future work

2 The Reputation Model

In hostile network environments which most current WSNdeployments would assume regular sensor nodes need tobe confronted with security threats during the process ofsensor localization If a sensor node can identify the securityand credibility of location information that it receives andsubsequently use the information appropriately the accuracyof sensor localization can be greatly improved or ensured insuch environments Therefore in order to develop effectivesensor localization schemes we should understand the main











U998400

1

B998400

2

U1

B2

B1

B3




2 which


result 11988010158401











119898



119894 119910119894) to its






119895


119889119861119895119894



119861119895119894using the


119861119895119894


119894




and 119877119905+Δ119905119861119895119894


119894by 119861119895at times 119905 and 119905 +Δ119905


119877119905+Δ119905

119861119895119894= 120572 times 119877

119905

119861119895119894+ (1 minus 120572)

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816le Δ119889

119877119905+Δ119905

119861119895119894= (1 minus 120572) times 119877

119905

119861119895119894

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816gt Δ119889

(1)

120572 =

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816

119897119861119895119894+ 119889119861119895119894

(2)




119880119898 119861119894and 119877119905+Δ119905

119861119896119894




119896at



Consider

119877119905+Δ119905

119880119898 119861119894=

sum119899

119896=1119877119905+Δ119905

119861119896119894

119899 (3)







1 1199101)



2 1198613 1198614


1after receiving



1









1 1199101)

(1199092 1199102) (119909



1198891 1198892 119889


1


(1199091198801minus 119909119894)2

+ (1199101198801minus 119910119894)2

= 1198892

119894 119894 = 1 2 119901 (4)





119861119895119894= 05

while(true)

120572 =

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816

119897119861119895119894+ 119889119861119895119894

if100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816le Δ119889

119877119905+Δ119905

119861119895119894= 120572 times 119877

119905

119861119895119894+ (1 minus 120572)

else119877119905+Δ119905

119861119895119894= (1 minus 120572) times 119877

119905

119861119895119894


119895 node 119880



119877119905+Δ119905

119880119898 119861119894=

sum119899

119896=1119877119905+Δ119905

119861119896119894

119899


B2

U2

U3

U6

B5

B4

B3

B1

U10

U9

U7

U5

U1

U4

U8



1199092

1minus 1199092

119901minus 2 (119909

1minus 119909119901) 1199091198801+ 1199102

1minus 1199102

119901

minus 2 (1199101minus 119910119901) 1199101198801= 1198892

1minus 1198892

119901

sdot sdot sdot

1199092


119901minus 2 (119909

119901minus1minus 119909119901) 1199091198801+ 1199102


119901



119901

(5)


1198801= 119860minus1119887 (6)


119860 = 2[

[

1199091minus 119909119901

1199101minus 119910119901

sdot sdot sdot


]

]

119887 =[[

[

1199092

1minus 1199092

119901+ 1199102

1minus 1199102

119901minus 1198892

1+ 1198892

119901

sdot sdot sdot

1199092


119901+ 1199102


119901minus 1198892

119901minus1+ 1198892

119901

]]

]

1198801= [1199091198801

1199101198801

]

(7)


119880 = (119860119879119860)minus1

119860119879119887 (8)






0 5 10 15 20

00

05

10

15

20

25Lo

caliz

atio

n er

ror



(a)


0 5 10 15 20000204060810121416182022

Loca

lizat

ion

erro

r


(b)




119880119898 119910119880119898) and (1199091015840

119880119898 1199101015840

119880119898) denote



119890119898=

radic(119909119880119898minus 1199091015840

119880119898)2

+ (119910119880119898minus 1199101015840

119880119898)2

119877

(9)




119890 =sum119873

119894=1119890119894

119873 (10)






0 5 10 15 20 25 30 35 40 45000002004006008010012014016018020022024026028030032034036038040

Aver

age l

ocal

izat

ion

erro

r


PLSRBL


0 1 2 3 4 5 6 700

02

04

06

08

10

12

14

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL



0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL




1cap 1198621198752

1198771198752cap 1198621198751

1198771198752cap 1198621198752

1198771198752cap 1198621198753



1and 119877119875

2represent the


1

1198621198752 and 119862119875





0 1 2 3 4 500

02

04

06

08

10

12

14

16

18Av

erag

e loc

aliz

atio

n er

ror

Time (min)

PLSRBL

(a)

0 1 2 3 4 500

02

04

06

08

10

12

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(b)

0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(c)

0 1 2 3 4 500

01

02

03

04

05

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(d)








5 Conclusion



Notations

119861119894 Beacon node 119894

119880119898 Sensor node119898





119877119905








Acknowledgments


References

























RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation



















U998400

1

B998400

2

U1

B2

B1

B3




2 which


result 11988010158401











119898



119894 119910119894) to its






119895


119889119861119895119894



119861119895119894using the


119861119895119894


119894




and 119877119905+Δ119905119861119895119894


119894by 119861119895at times 119905 and 119905 +Δ119905


119877119905+Δ119905

119861119895119894= 120572 times 119877

119905

119861119895119894+ (1 minus 120572)

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816le Δ119889

119877119905+Δ119905

119861119895119894= (1 minus 120572) times 119877

119905

119861119895119894

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816gt Δ119889

(1)

120572 =

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816

119897119861119895119894+ 119889119861119895119894

(2)




119880119898 119861119894and 119877119905+Δ119905

119861119896119894




119896at



Consider

119877119905+Δ119905

119880119898 119861119894=

sum119899

119896=1119877119905+Δ119905

119861119896119894

119899 (3)







1 1199101)



2 1198613 1198614


1after receiving



1









1 1199101)

(1199092 1199102) (119909



1198891 1198892 119889


1


(1199091198801minus 119909119894)2

+ (1199101198801minus 119910119894)2

= 1198892

119894 119894 = 1 2 119901 (4)





119861119895119894= 05

while(true)

120572 =

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816

119897119861119895119894+ 119889119861119895119894

if100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816le Δ119889

119877119905+Δ119905

119861119895119894= 120572 times 119877

119905

119861119895119894+ (1 minus 120572)

else119877119905+Δ119905

119861119895119894= (1 minus 120572) times 119877

119905

119861119895119894


119895 node 119880



119877119905+Δ119905

119880119898 119861119894=

sum119899

119896=1119877119905+Δ119905

119861119896119894

119899


B2

U2

U3

U6

B5

B4

B3

B1

U10

U9

U7

U5

U1

U4

U8



1199092

1minus 1199092

119901minus 2 (119909

1minus 119909119901) 1199091198801+ 1199102

1minus 1199102

119901

minus 2 (1199101minus 119910119901) 1199101198801= 1198892

1minus 1198892

119901

sdot sdot sdot

1199092


119901minus 2 (119909

119901minus1minus 119909119901) 1199091198801+ 1199102


119901



119901

(5)


1198801= 119860minus1119887 (6)


119860 = 2[

[

1199091minus 119909119901

1199101minus 119910119901

sdot sdot sdot


]

]

119887 =[[

[

1199092

1minus 1199092

119901+ 1199102

1minus 1199102

119901minus 1198892

1+ 1198892

119901

sdot sdot sdot

1199092


119901+ 1199102


119901minus 1198892

119901minus1+ 1198892

119901

]]

]

1198801= [1199091198801

1199101198801

]

(7)


119880 = (119860119879119860)minus1

119860119879119887 (8)






0 5 10 15 20

00

05

10

15

20

25Lo

caliz

atio

n er

ror



(a)


0 5 10 15 20000204060810121416182022

Loca

lizat

ion

erro

r


(b)




119880119898 119910119880119898) and (1199091015840

119880119898 1199101015840

119880119898) denote



119890119898=

radic(119909119880119898minus 1199091015840

119880119898)2

+ (119910119880119898minus 1199101015840

119880119898)2

119877

(9)




119890 =sum119873

119894=1119890119894

119873 (10)






0 5 10 15 20 25 30 35 40 45000002004006008010012014016018020022024026028030032034036038040

Aver

age l

ocal

izat

ion

erro

r


PLSRBL


0 1 2 3 4 5 6 700

02

04

06

08

10

12

14

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL



0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL




1cap 1198621198752

1198771198752cap 1198621198751

1198771198752cap 1198621198752

1198771198752cap 1198621198753



1and 119877119875

2represent the


1

1198621198752 and 119862119875





0 1 2 3 4 500

02

04

06

08

10

12

14

16

18Av

erag

e loc

aliz

atio

n er

ror

Time (min)

PLSRBL

(a)

0 1 2 3 4 500

02

04

06

08

10

12

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(b)

0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(c)

0 1 2 3 4 500

01

02

03

04

05

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(d)








5 Conclusion



Notations

119861119894 Beacon node 119894

119880119898 Sensor node119898





119877119905








Acknowledgments


References

























RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation














119898



119894 119910119894) to its






119895


119889119861119895119894



119861119895119894using the


119861119895119894


119894




and 119877119905+Δ119905119861119895119894


119894by 119861119895at times 119905 and 119905 +Δ119905


119877119905+Δ119905

119861119895119894= 120572 times 119877

119905

119861119895119894+ (1 minus 120572)

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816le Δ119889

119877119905+Δ119905

119861119895119894= (1 minus 120572) times 119877

119905

119861119895119894

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816gt Δ119889

(1)

120572 =

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816

119897119861119895119894+ 119889119861119895119894

(2)




119880119898 119861119894and 119877119905+Δ119905

119861119896119894




119896at



Consider

119877119905+Δ119905

119880119898 119861119894=

sum119899

119896=1119877119905+Δ119905

119861119896119894

119899 (3)







1 1199101)



2 1198613 1198614


1after receiving



1









1 1199101)

(1199092 1199102) (119909



1198891 1198892 119889


1


(1199091198801minus 119909119894)2

+ (1199101198801minus 119910119894)2

= 1198892

119894 119894 = 1 2 119901 (4)





119861119895119894= 05

while(true)

120572 =

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816

119897119861119895119894+ 119889119861119895119894

if100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816le Δ119889

119877119905+Δ119905

119861119895119894= 120572 times 119877

119905

119861119895119894+ (1 minus 120572)

else119877119905+Δ119905

119861119895119894= (1 minus 120572) times 119877

119905

119861119895119894


119895 node 119880



119877119905+Δ119905

119880119898 119861119894=

sum119899

119896=1119877119905+Δ119905

119861119896119894

119899


B2

U2

U3

U6

B5

B4

B3

B1

U10

U9

U7

U5

U1

U4

U8



1199092

1minus 1199092

119901minus 2 (119909

1minus 119909119901) 1199091198801+ 1199102

1minus 1199102

119901

minus 2 (1199101minus 119910119901) 1199101198801= 1198892

1minus 1198892

119901

sdot sdot sdot

1199092


119901minus 2 (119909

119901minus1minus 119909119901) 1199091198801+ 1199102


119901



119901

(5)


1198801= 119860minus1119887 (6)


119860 = 2[

[

1199091minus 119909119901

1199101minus 119910119901

sdot sdot sdot


]

]

119887 =[[

[

1199092

1minus 1199092

119901+ 1199102

1minus 1199102

119901minus 1198892

1+ 1198892

119901

sdot sdot sdot

1199092


119901+ 1199102


119901minus 1198892

119901minus1+ 1198892

119901

]]

]

1198801= [1199091198801

1199101198801

]

(7)


119880 = (119860119879119860)minus1

119860119879119887 (8)






0 5 10 15 20

00

05

10

15

20

25Lo

caliz

atio

n er

ror



(a)


0 5 10 15 20000204060810121416182022

Loca

lizat

ion

erro

r


(b)




119880119898 119910119880119898) and (1199091015840

119880119898 1199101015840

119880119898) denote



119890119898=

radic(119909119880119898minus 1199091015840

119880119898)2

+ (119910119880119898minus 1199101015840

119880119898)2

119877

(9)




119890 =sum119873

119894=1119890119894

119873 (10)






0 5 10 15 20 25 30 35 40 45000002004006008010012014016018020022024026028030032034036038040

Aver

age l

ocal

izat

ion

erro

r


PLSRBL


0 1 2 3 4 5 6 700

02

04

06

08

10

12

14

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL



0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL




1cap 1198621198752

1198771198752cap 1198621198751

1198771198752cap 1198621198752

1198771198752cap 1198621198753



1and 119877119875

2represent the


1

1198621198752 and 119862119875





0 1 2 3 4 500

02

04

06

08

10

12

14

16

18Av

erag

e loc

aliz

atio

n er

ror

Time (min)

PLSRBL

(a)

0 1 2 3 4 500

02

04

06

08

10

12

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(b)

0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(c)

0 1 2 3 4 500

01

02

03

04

05

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(d)








5 Conclusion



Notations

119861119894 Beacon node 119894

119880119898 Sensor node119898





119877119905








Acknowledgments


References

























RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation











119861119895119894= 05

while(true)

120572 =

100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816

119897119861119895119894+ 119889119861119895119894

if100381610038161003816100381610038161003816119897119861119895119894minus 119889119861119895119894

100381610038161003816100381610038161003816le Δ119889

119877119905+Δ119905

119861119895119894= 120572 times 119877

119905

119861119895119894+ (1 minus 120572)

else119877119905+Δ119905

119861119895119894= (1 minus 120572) times 119877

119905

119861119895119894


119895 node 119880



119877119905+Δ119905

119880119898 119861119894=

sum119899

119896=1119877119905+Δ119905

119861119896119894

119899


B2

U2

U3

U6

B5

B4

B3

B1

U10

U9

U7

U5

U1

U4

U8



1199092

1minus 1199092

119901minus 2 (119909

1minus 119909119901) 1199091198801+ 1199102

1minus 1199102

119901

minus 2 (1199101minus 119910119901) 1199101198801= 1198892

1minus 1198892

119901

sdot sdot sdot

1199092


119901minus 2 (119909

119901minus1minus 119909119901) 1199091198801+ 1199102


119901



119901

(5)


1198801= 119860minus1119887 (6)


119860 = 2[

[

1199091minus 119909119901

1199101minus 119910119901

sdot sdot sdot


]

]

119887 =[[

[

1199092

1minus 1199092

119901+ 1199102

1minus 1199102

119901minus 1198892

1+ 1198892

119901

sdot sdot sdot

1199092


119901+ 1199102


119901minus 1198892

119901minus1+ 1198892

119901

]]

]

1198801= [1199091198801

1199101198801

]

(7)


119880 = (119860119879119860)minus1

119860119879119887 (8)






0 5 10 15 20

00

05

10

15

20

25Lo

caliz

atio

n er

ror



(a)


0 5 10 15 20000204060810121416182022

Loca

lizat

ion

erro

r


(b)




119880119898 119910119880119898) and (1199091015840

119880119898 1199101015840

119880119898) denote



119890119898=

radic(119909119880119898minus 1199091015840

119880119898)2

+ (119910119880119898minus 1199101015840

119880119898)2

119877

(9)




119890 =sum119873

119894=1119890119894

119873 (10)






0 5 10 15 20 25 30 35 40 45000002004006008010012014016018020022024026028030032034036038040

Aver

age l

ocal

izat

ion

erro

r


PLSRBL


0 1 2 3 4 5 6 700

02

04

06

08

10

12

14

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL



0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL




1cap 1198621198752

1198771198752cap 1198621198751

1198771198752cap 1198621198752

1198771198752cap 1198621198753



1and 119877119875

2represent the


1

1198621198752 and 119862119875





0 1 2 3 4 500

02

04

06

08

10

12

14

16

18Av

erag

e loc

aliz

atio

n er

ror

Time (min)

PLSRBL

(a)

0 1 2 3 4 500

02

04

06

08

10

12

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(b)

0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(c)

0 1 2 3 4 500

01

02

03

04

05

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(d)








5 Conclusion



Notations

119861119894 Beacon node 119894

119880119898 Sensor node119898





119877119905








Acknowledgments


References

























RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










0 5 10 15 20

00

05

10

15

20

25Lo

caliz

atio

n er

ror



(a)


0 5 10 15 20000204060810121416182022

Loca

lizat

ion

erro

r


(b)




119880119898 119910119880119898) and (1199091015840

119880119898 1199101015840

119880119898) denote



119890119898=

radic(119909119880119898minus 1199091015840

119880119898)2

+ (119910119880119898minus 1199101015840

119880119898)2

119877

(9)




119890 =sum119873

119894=1119890119894

119873 (10)






0 5 10 15 20 25 30 35 40 45000002004006008010012014016018020022024026028030032034036038040

Aver

age l

ocal

izat

ion

erro

r


PLSRBL


0 1 2 3 4 5 6 700

02

04

06

08

10

12

14

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL



0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL




1cap 1198621198752

1198771198752cap 1198621198751

1198771198752cap 1198621198752

1198771198752cap 1198621198753



1and 119877119875

2represent the


1

1198621198752 and 119862119875





0 1 2 3 4 500

02

04

06

08

10

12

14

16

18Av

erag

e loc

aliz

atio

n er

ror

Time (min)

PLSRBL

(a)

0 1 2 3 4 500

02

04

06

08

10

12

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(b)

0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(c)

0 1 2 3 4 500

01

02

03

04

05

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(d)








5 Conclusion



Notations

119861119894 Beacon node 119894

119880119898 Sensor node119898





119877119905








Acknowledgments


References

























RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










0 5 10 15 20 25 30 35 40 45000002004006008010012014016018020022024026028030032034036038040

Aver

age l

ocal

izat

ion

erro

r


PLSRBL


0 1 2 3 4 5 6 700

02

04

06

08

10

12

14

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL



0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL




1cap 1198621198752

1198771198752cap 1198621198751

1198771198752cap 1198621198752

1198771198752cap 1198621198753



1and 119877119875

2represent the


1

1198621198752 and 119862119875





0 1 2 3 4 500

02

04

06

08

10

12

14

16

18Av

erag

e loc

aliz

atio

n er

ror

Time (min)

PLSRBL

(a)

0 1 2 3 4 500

02

04

06

08

10

12

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(b)

0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(c)

0 1 2 3 4 500

01

02

03

04

05

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(d)








5 Conclusion



Notations

119861119894 Beacon node 119894

119880119898 Sensor node119898





119877119905








Acknowledgments


References

























RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










0 1 2 3 4 500

02

04

06

08

10

12

14

16

18Av

erag

e loc

aliz

atio

n er

ror

Time (min)

PLSRBL

(a)

0 1 2 3 4 500

02

04

06

08

10

12

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(b)

0 1 2 3 4 500

02

04

06

08

10

12

14

16

18

20

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(c)

0 1 2 3 4 500

01

02

03

04

05

Aver

age l

ocal

izat

ion

erro

r

Time (min)

PLSRBL

(d)








5 Conclusion



Notations

119861119894 Beacon node 119894

119880119898 Sensor node119898





119877119905








Acknowledgments


References

























RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation











5 Conclusion



Notations

119861119894 Beacon node 119894

119880119898 Sensor node119898





119877119905








Acknowledgments


References

























RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation




















RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation









research article reputation-based secure sensor...

Documents