无线传感器网络模型设计英文文献翻译精说课讲解.docx

无线传感器网络模型设计英文文献翻译精说课讲解.docx

《无线传感器网络模型设计英文文献翻译精说课讲解.docx》由会员分享，可在线阅读，更多相关《无线传感器网络模型设计英文文献翻译精说课讲解.docx（15页珍藏版）》请在冰点文库上搜索。

无线传感器网络模型设计英文文献翻译精说课讲解

ModelDesignofWirelessSensorNetworkbasedonScale-FreeNetworkTheory

ABSTRACT

Thekeyissueofresearchesonwirelesssensornetworksistobalancetheenergycostsacrossthewholenetworkandtoenhancetherobustnessinordertoextendthesurvivaltimeofthewholesensornetwork.Asaspecialcomplexnetworklimitedespeciallybytheenvironment,sensornetworkismuchdifferentfromthetraditionalcomplexnetworks,suchasInternetnetwork,ecologicalnetwork,socialnetworkandetc.Itisnecessarytointroduceawayofhowtostudywirelesssensornetworkbycomplexnetworktheoryandanalysismethods,thekeyofwhichliesinasuccessfulmodelingwhichisabletomakecomplexnetworktheoryandanalysismethodsmoresuitablefortheapplicationofwirelesssensornetworkinordertoachievetheoptimizationofsomecertainnetworkcharacteristicsofwirelesssensornetwork.Basedongenerationrulesoftraditionalscale-freenetworks,thispaperaddedseveralrestrictionstotheimprovedmodel.Thesimulationresultshowsthatimprovementsmadeinthispaperhavemadetheentirenetworkhaveabetterrobustnesstotherandomfailureandtheenergycostsaremorebalancedandreasonable.Thisimprovedmodelwhichisbasedonthecomplexnetworktheoryprovesmoreapplicabletotheresearchofwirelesssensornetwork.

Key-words:

Wirelesssensornetwork;Complexnetwork;Scale-freenetwork

I.INTRODUCTION

Inrecentyears,wirelesssensornetworkshaveattractedmoreandmorerelatedresearchersforitsadvantages.Sensornodesareusuallylow-powerandnon-rechargeable.Theintegrityoftheoriginalnetworkswillbedestroyedandothernodeswillhavemorebusinessburdenfordatatransmissioniftheenergyofsomecertainnodesdeplete.Thekeyissueofsensornetworkresearchistobalancetheenergyconsumptionofallsensornodesandtominimizetheimpactofrandomfailureofsensornodesorrandomattackstosensornodesontheentirenetwork[1].

ComplexnetworktheoryhasbeenforsometimesincefirstproposedbyBarabasiandAlbertin1998,butcomplexnetworktheoryandanalysismethodappliedtowirelesssensornetworksresearchisseriouslyrareanddevelopsinslowprogress.Asaspecialcomplexnetworklimitedespeciallybytheenvironment,sensornetworkismuchdifferentfromthetraditionalcomplexnetwork,andtheexistingcomplexnetworktheoryandanalysismethodscannotbedirectlyappliedtoanalyzesensornetworks.Basedonscale-freenetworktheory（BAmodel[2],（1thispaperaddedarandomdamagemechanismtoeachsensornodewhendeployedinthegenerationrule;（2consideringtherealstatementofwirelesssensornetworks,aminimumandmaxinumrestrictiononsensorcommunicationradiuswasaddedtoeachsensornode;（3inordertomaintainabalancedenergycomsuptionoftheentirenetwork,thispaperaddedalimiteddegreeofsaturationvaluetoeachsensornode.Thisimprovedscale-freemodelnotonlyhasthementionedimprovementsabove,butalsohaslotsofadvantagesoftraditionalscale-freenetworks,suchasthegoodabilitytoresistrandomattacks,sothattheexistingtheoryandanalysismethodsofcomplexnetworkwillbemoresuitablefortheresearchesofwirelesssensornetwork.

II.PROGRESSOFRELATEDRESEARCH

HailinZhuandHongLuohaveproposedtwocomplexnetworks-basedmodelsforwirelesssensornetworks[3],thefirstofwhichnamedEnergy-awareevolutionmodel（EAEMcanorganizethenetworksinanenergy-efficientway,andcanproducescale-freenetworkswhichcanimprovethenetworksrelianceagainstrandomfailureofthesensornodes.InthesecondmodelnamedEnergy-balancedevolutionmodel（EBEM,themaximumnumberoflinksforeachnodeisintroducedintothealgorithm,whichcanmakeenergyconsumptionmorebalancedthanthepreviousmodel（EAEM.

CHENLijunandMAOYingchihaveproposedatopologycontrolofwirelesssensornetworksunderanaveragedegreeconstraint[4].Inthepreconditionofthetopologyconnectivityofwirelesssensornetworks,howtosolvethesparsenessofthenetworktopologyisaveryimportantprobleminalargenumberofsensornodesdeployedrandomly.Theyprovedtheirproposedschemecandecreaseworkingnodes,guaranteenetworktopologysparseness,predigestroutingcomplexityandprolongnetworksurvivalperiod.

LEIMingandLIDeshihaveproposedaresearchonself-organizationreliabilityofwirelesssensornetwork[5],whichaimingonthetwosituations:

deficiencyofWSNnodesandunderexternalattack,analyzestheerrortoleranceabilityofdifferenttopologiesofWSN,andeventuallyobtainsoptimizedself—organizedtopologicalmodelsofWSNandproposesarefinedroutingalgorithmbasedonWSN.

III.IMPROVEDSCALE-FREEMODELFORWSN

Becauseofthelimitedenergyandtheevilapplicationenvironment,wirelesssensornetworksmayeasilycollapsewhensomecertainsensornodesareofenergydepletionordestructionbythenature,andevensomesensornodeshavebeendamagedwhendeployed.Thereisalsoarestrictiononmaxinumandmininumcommunicationradiusofsensornodesratherthantheotherknownscale-freenetworkssuchasInternetnetwork,whichhasnorestrictiononcommunicationradius.Tohaveabalancedenergyconsumption,itisnecessarytosetupasaturationvaluelimiteddegreeofeachsensornode[6].

Inresponsetothesepoints,basedonthetraditionalscale-freemodel,thispaperhasmadethefollowingimprovementsintheprocessofmodelestablishment:

（1Alargenumberofresearcheshaveshownthatmanycomplexnetworksinnaturearenotonlytheresultfrominternalforces,butalsotheresultfromexternalforceswhichshouldnotbeignoredtoformanentirecomplexnetwork.Nodefailuremaynotonlyoccourbynodeenergydepletionorrandomattackstothemwhensensornetworksareintheworkingprogress,butalsooccourbyexternalforces,suchasbythenature,whendeployed.Inthispaper,amechanismofsmallprobabilityofrandomdamagehasbeenaddedtotheformationofsensornetworks.

（2UnlikeInternetnetworkwheretwonodesareabletoconnectdirectlytoeachotherandtheirconnectionareneverlimitedbytheirreallocation,sensornetwork,twonodesinwhichconnecttoeachotherbythewayofmulti-hop,sothateachnodehasamaximumoflengthrestrictionontheircommunicationradius.Toensurethesparseofthewholenetwork,theremustalsobeaminimumoflengthrestrictionontheircommunicationradius.Inthispaper,alengthrestrictiononcommunicationradiusofsensornodeshasbeenproposedintheimprovedmodel.

（3Insensornetwork,ifthereexistsasensornodewithaseriouslyhighdegree,whoseenergyconsumptionisveryquickly,itwillbeseriouslybad.Thewholesensornetworkwouldsurelycollapseifenoughenergywerenotsupportedtothecertainnode.Toavoidthissituation,thispaperhassetupasaturationvaluelimiteddegreeofeachsensornode.Byaddingthementionedrestrictionsabovetotheformationofthescale-freemodel,thenewimprovedmodelwillbemoreinlinewiththerealstatementofsensornetwork.Complexnetworktheoryandanalysismethodswillbemoreappropriatewhenusedtoresearchandanalyzethesensornetwork.

IV.DESCRIPTIONOFTHEIMPROVEDALGORITHM

Thespecificalgorithmoftheimprovedmodelformationaredescribedasfollows:

（1Agivenregion（assumedtobesquareisdividedintoHS*HSbigsquares（namedasBS;

（2EachBS（assumedtobesquareisdividedintoLS*LSsmallsquares（namedasSS,andeachSScanhaveonlyonenodeinitscoverageregion;

（3m0backbonenodesareinitiallygeneratedasarandomgraph,andthenanewnodewillbeaddedtothenetworktoconnecttheexistingmnodeswithmedgesateachtimeinterval.（m

（4Thenewlygeneratednodev,hasacertainprobabilityofPetobedamageddirectlysothatitwillneverbeconnectedwithanyexistingnodes;

（5Thenewlygeneratednodevconnectswiththeexistingnodei,whichobeyesdependent-preferenceruleandissurelylimitedbythedegreeofthecertainsaturationvalue.

（6Thedistancedivbetweenthenewlygeneratednodevconnectsandtheexistingnodeishallbeshorterthanthemaximumdmaxofthecommunicationradiusofsensornodes.

Aboveall,theprobabilitythattheexistingnodeiwillbeconnectedwiththenewlygeneratednodevcanbeshownasfollows:

Inordertocomputeitconveniently,hereassumedthatfewnodeshadreachedthedegreeofsaturationvaluekimax.Thatis,NisveryminimalinEqs.

（1sothatitcanbeignoredhere.AndinEqs.

i

Nj1akKj

π=≈∑0N=m1t+-（2

WithThevaryingratewithtimeofki,weget:

0m112iiiitj

jkamkamkmtmtmkδπδ+-====-∑（3

Whent→∞,

condition:

ki（ti=m,wegetthesolution:

i2,itkta

ββ=（t=m（（4TheprobabilitythatthedegreeofnodeIissmallerthankis:

11{k（tk}P{t}iimtPkββ<=>（5

Thetimeintervalwheneachnewlygeneratednodeconnectedintothenetworkisequal,sothatprobabilitydensityoftiisaconstantparameter:

01（tiPmt

=+1/βwereplaceitintoEqs.（5,thenweget:

11111{k（tk}P{t}1（timtkiiitmtPPkββββ=<=>=-∑（6

1101（tm

mtkββ-+Soweget:

110（k（tk21（k.iPmtPkmtkββ

δδ<==+（7Whent→∞,weget:

2（k2mrPk-=（8Inwhich1

2=1+=1+aγβ,andthedegreedistributionwegetandthedegreedistributionoftraditionalscale-freenetworkaresimilar.Approximately,ithasnothingtodowiththetimeparametertandthequantityofedgesmgeneratedateachtimeinterval.

maxP{dd}iv≤couldbecalculatedbythemaxinumrestrictiondmaxoncommunication

radiusofeachsensornodeandtheareaoftheentirecoverageregionS,that

ismaxP{dd}iv≤=2S

dπThenw

ereplacemaxP{dd}iv≤=2

Sdπ

anda=max