Ensemblebasedtraf clightcontrolforcityzonesusingareducednumberof

sensors
DanPescaru,Daniel-IoanCuriac
FacultyofAutomaticsandComputers,‘‘Politehnica’’UniversityofTimisßoara,Bd.V.ParvanNo.2,Timisoara300223,Romania
articleinfoabstract
Rapidadvancesincomputing,sensingandtelecommunicationtechnologyofferunprece-dentedopportunitiesforarti cialintelligenceconceptstoexpandtheirapplicationsin
the eldoftraf cmanagementandcontrol.Ourmethodologygravitatesaroundapowerfuldecision-makingmethod:ensemble-basedsystems.Thistechniqueisusedtoaccuratelyclassifythenearfuturetraf cconditionsandtomakeef cientdecisionsforadaptingthetraf clightssequenceswithinanurbanareatooptimizethetraf c ows.Theproposedapproachrequiresonlymeasurementsprovidedbytraf csensorslocatedalongtheprinci-palroadsenteringthezone.Thisreducednumberofsensorsareconsideredtobeenoughrelevantforclassifyingthenearfuturestateofthetraf candmoreover,theirmeasure-mentscanbevalidatedthroughanalytical/hardwareredundancy.Ourmethodologyismeanttobeimplementedwithintheframeworkofawirelesssensorandactuatornetworkandiscon rmedbycomputersimulation,includingnormalorabnormaltraf cconditions,forthecentralpartofthecityofTimisoara-Romania.
Ó2014ElsevierLtd.Allrightsreserved.
Articlehistory:
Received7June2013
Receivedinrevisedform11June2014Accepted25June2014
Keywords:
Intelligenttraf clightcontrolEnsemblebasedsystem
Wirelesssensorandactuatornetwork
1.Introduction
Urbantraf crepresentsahighlycomplexphenomenonthatbecomesmoreandmoreamajorconcernforoureverydaylife.Theescalatingdemandforpeopleandgoodsmobilityinurbanareas(withlimitedroadinfrastructure)hascausedfre-quenttraf ccongestion,withvariousundesirableconsequences:delays,energywaste,noise,pollutionorroadaccidents.Overtheyearsadiverserangeofsolutionshadbeenappliedtoreducetheleveloftraf ccongestionandtominimizetheconsequences.Duetotheintricatesetofinteractionsbetweenroadinfrastructure,diversetypesofvehicles,weathercondi-tionsandmultitudeoftechnologiesinvolved,ageneralapproachhasyettobefound.
Withthecontinuousadvancementsincomputing,communicationandsensingtechnology,aseriesofarti cialintelli-genceconceptshadbeenemployedintraf cmanagementandoptimization:expertsystems(FindlerandStapp,1992),pre-diction-basedoptimization(TavladakisandVoulgaris,1999;Liuetal.,2002),fuzzylogic(Tanetal.,1995;Leeetal.,1995),neuralnetworks(Srinivasanetal.,2006;Vlahogiannietal.,2005),evolutionaryalgorithms(Taaleetal.,1998),reinforcementlearning(ThorpeandAndersson,1996;Sutton,1996),etc.Theirdirectoutcomeresidesinanoverallimprovementoftraf c owsinbothnormalandabnormaltraf ccircumstances.
Thispaperproposesanewperspectiveuponadaptivetraf clightcontrolinurbanareasemployingnotasingleexpertsystembutamixtureofexperts(ensemblebasedsystem),inordertoboostthetraf cdecisionaccuracy.Ourapproachisbasedonthefollowingpremises:(i)onlyasmallnumberofroadsenteringacityzonecoagulatethemajorityoftraf c–the Correspondingauthor.Tel.:+40256403227;fax:+40256403214.
E-mailaddresses:dan@cs.upt.ro(D.Pescaru),daniel.curiac@aut.upt.ro(D.-I.Curiac).

doc.guandang.net
0968-090X/Ó2014ElsevierLtd.Allrightsreserved.
Traffic Modelling for Moving-Block Train Control System_专业资料。维普资讯 ... The control system for... 4人阅读 4页 2.00 DSP-based multi-...
LTE QoS (Part 3) - P2P Traffic Bandwidth Control using the LTE QoS_信息与通信_工程科技_专业资料。NETMANIAS TECH-BLOG Please visit www.netmanias.com ...
LTE QoS (Part 3) - P2P Traffic Bandwidth Control using the LTE QoS_信息与通信_工程科技_专业资料。NETMANIAS TECH-BLOG Please visit www.netmanias.com ...
热门文档