BigDataandMobility asaService
Editedby HaoranZhang
CenterforSpatialInformationScience,TheUniversityofTokyo,Kashiwa, Chiba,Japan
XuanSong
DepartmentofComputerScienceandEngineering,SouthUniversityof ScienceandTechnology,Nanshan,China
CenterforSpatialInformationScience,TheUniversityofTokyo,Kashiwa, Chiba,Japan
RyosukeShibasaki
CenterforSpatialInformationScience,TheUniversityofTokyo,Kashiwa, Chiba,Japan
Elsevier
Radarweg29,POBox211,1000AEAmsterdam,Netherlands
TheBoulevard,LangfordLane,Kidlington,OxfordOX51GB,UnitedKingdom 50HampshireStreet,5thFloor,Cambridge,MA02139,UnitedStates
Copyright©2022ElsevierInc.Allrightsreserved.
Nopartofthispublicationmaybereproducedortransmittedinanyformorbyanymeans, electronicormechanical,includingphotocopying,recording,oranyinformationstorageand retrievalsystem,withoutpermissioninwritingfromthepublisher.Detailsonhowtoseek permission,furtherinformationaboutthePublisher’spermissionspoliciesandourarrangements withorganizationssuchastheCopyrightClearanceCenterandtheCopyrightLicensingAgency,can befoundatourwebsite: www.elsevier.com/permissions
Thisbookandtheindividualcontributionscontainedinitareprotectedundercopyrightbythe Publisher(otherthanasmaybenotedherein).
Notices
Knowledgeandbestpracticeinthisfieldareconstantlychanging.Asnewresearchandexperience broadenourunderstanding,changesinresearchmethods,professionalpractices,ormedical treatmentmaybecomenecessary.
Practitionersandresearchersmustalwaysrelyontheirownexperienceandknowledgeinevaluating andusinganyinformation,methods,compounds,orexperimentsdescribedherein.Inusingsuch informationormethodstheyshouldbemindfuloftheirownsafetyandthesafetyofothers, includingpartiesforwhomtheyhaveaprofessionalresponsibility.
Tothefullestextentofthelaw,neitherthePublishernortheauthors,contributors,oreditors, assumeanyliabilityforanyinjuryand/ordamagetopersonsorpropertyasamatterofproducts liability,negligenceorotherwise,orfromanyuseoroperationofanymethods,products, instructions,orideascontainedinthematerialherein.
LibraryofCongressCataloging-in-PublicationData
AcatalogrecordforthisbookisavailablefromtheLibraryofCongress
BritishLibraryCataloguing-in-PublicationData
AcataloguerecordforthisbookisavailablefromtheBritishLibrary
ISBN:978-0-323-90169-7
ForinformationonallElsevierpublications visitourwebsiteat https://www.elsevier.com/books-and-journals
Publisher: JoeHayton
EditorialProjectManager: HowiM.DeRamos
ProductionProjectManager: PunithavathyGovindaradjane
CoverDesigner:MilesHitchen
TypesetbySTRAIVE,India
Contributorsxi
Introductionxiii
ShreyasBharule,HaoranZhang,andRyosukeShibasaki
1.MaaSsystemdevelopmentandAPPs 1 WenjingLi,RyosukeShibasaki,HaoranZhang,and JinyuChen
1.ThedevelopmenthistoryofMaaS 1 1.1Theconception1 1.2Theearlyapplication1 1.3MaaSalliance2
1.4Development3 1.5Revolutionandinnovation4
2.ThecategoryofMaaSsystem 5
2.1Level0:Nointegration6
2.2Level1:Informationintegration6
2.3Level2:Integrationofbookingandpayment6
2.4Level3:Integrationoftheserviceoffering6 2.5Level4:Integrationofsocietalgoals7
3.Studycase 7 3.1UbiGo10 3.2Whim12 3.3Moovit15 3.4Uber17
4.FuturedevelopmenttrendofMaaSsystem 21 4.1Data-integrated21 4.2Future-oriented22 4.3Sustainable23 References 23
2.Spatio-temporaldatapreprocessingtechnologies 25 JinyuChen,HaoranZhang,WenjingLi,and RyosukeShibasaki
1.Introduction 25 2.RawGPSdataandworkflowofdatapreprocessing 26
3.Keytechnologiesandcorrespondingapplication 27 3.1Outlierremovement27
3.2Staylocationdetection29 3.3Travelsegmentation30
3.4Travelmodedetection31 3.5Mapmatching33 3.6Summary35
4.Casestudy 35
4.1Staylocationdetection:Lifepatternanalysis35
4.2Travelsegmentationandmodedetection:Ride-sharing potentialanalysis48
4.3Mapmatching:EstimationofurbanscalePMemission60
5.Conclusion
3.Travelsimilarityestimationandclustering 77
YuhaoYao,RyosukeShibasaki,andHaoranZhang
2.Trajectorysimilarity 79
2.1Point-to-pointdistancemetric80
2.2Similarityfunctionoftrajectory82 2.3Trajectoryclustering87
3.Travelpatternsimilarity 90 3.1Travelpatternextraction91
3.2Travelpatternexpression92
3.3Travelpatternclustering93
4.Origin-destinationmatrixsimilarity 93
4.1VolumedifferencefocusedODsimilaritymeasure95
4.2Image-basedODsimilaritymeasure96
4.3Transformingdistance-basedODsimilaritymeasure97
4.4ODtableausimilaritymeasure:Mobsimilarity98
5.Casestudy 103
5.1CDR-basedtravelestimationaccuracyanalysis103 5.2Metrousagepatternclustering106
6.Conclusionandfuturedirections 106 References 108
4.DatafusiontechnologiesforMaaS 113
YiSui,HaoranZhang,WenxiaoJiang,RenchengSun, andFengjingShao
1.Introduction 113
2.Dataformula 115
2.1Attributeandeventdata115 2.2Trajectorydata116
2.3Origin-destination(OD)tripdata117 2.4Correlationnetwork117 2.5Environmentaldata118
3.CategoriesofdatafusionmethodsinMaaS 119
4.Datafusionbasedondeeplearning 122
4.1Fundamentalbuildingunitsofdeeplearningnetwork122
4.2Fusionstrategy130
5.Decomposition-basedmethods 135
6.ChallengingproblemsofdatafusioninMaaS 137
6.1Dataquality137
6.2Modelcomplexity137
6.3Datafusionincomparativeanalysis138
7.Conclusions 138 Acknowledgments 138 References 138
5.Data-drivenoptimizationtechnologiesforMaaS 143 HaoningXi
1.Overviewofdata-drivenoptimizationfortheurbanmobility system 143
1.1Data-drivendispatchingmethodsforon-demand ridesharing143
1.2Data-drivenschedulingmethodsforpublictransit147
1.3Data-drivenrebalancingmethodsfor bicycle-sharing148
2.OverviewofthegeneralconceptinMaaSSystem 149
2.1OverviewoftheMaaSsystems149
2.2OverviewofdatainMaaSsystems150
3.MobilityresourceallocationinMaaSsystem 152
3.1MobilityresourceallocationframeworkinMaaS152
3.2Data-drivenonlinestochasticresourceallocation problems157
4.Data-drivenoptimizationtechnologiesforresourceallocation inMaaS 157
4.1Sampleaverageapproximation158
4.2Robustoptimization159
4.3Predictiveanalysisandprescriptiveanalysis161
4.4Machinelearning-basedrobustoptimization162
5.Real-worldapplicationandcasestudy 164
5.1Problemdescription164
5.2Methodology165
5.3Resultsanddiscussion165
6.Conclusions 170 References 171
6.Data-drivenestimationforurbantravel shareability 177
QingYu,WeifengLi,andDongyuanYang
1.Introduction 177
1.1Theemergenceofsharingtransportation177
1.2Thesignificanceofshareabilityestimation178
1.3Chapterorganization178
2.Emergingsharingtransportationmode 179
2.1Bicyclesharing180
2.2Ridesharingandtaxisharing181
2.3Customizedbus182
2.4Characteristicsofsharingtransportationmodes182
3.Backgroundtotraditionaldataandtheirlimitations 183
4.Newandemergingsourceofdata 183
4.1Trackandtracedata184
4.2Geographicinformationdata185
4.3Advantagesanddisadvantagesofnewdatasources186
5.Emergingformofkeytechnologies 187
5.1Agent-basedmodeling187
5.2HowABMcanbeappliedinshareabilityestimation188
6.CasestudyofABMinurbanshareabilityestimation 190
6.1Dynamicelectricfenceforbicyclesharing190
6.2ABMsimulation191
6.3Dataandstudyarea192
6.4Resultofsimulation192
6.5Evaluationoftheresult194
7.Opportunitiesandchallenges 197
7.1Dataacquisition197
7.2Demandprediction198
7.3DesignimprovementofABM198
7.4Accelerationoflarge-scaleABM198
8.Conclusions
7.DataminingtechnologiesforMobility-as-a-Service (MaaS) 203
Wen-LongShang,HaoranZhang,andYiSui
1.IntroductionofdataminingtechnologiesinMaaSsystem 203
2.DataminingtechnologiesinMaaSsystem 204
2.1Whatisdatamining?204
2.2Objectofdatamining205
2.3Classicalstepsofdatamining205
2.4Typesoftransportationdata207
3.MethodologiesofdataminingtechnologiesusedinMaaS system 209
3.1Supportvectormachine209
3.2Linearregression213
3.3Decisiontree216
3.4Clusteringanalysis219
4.CasestudyofdataminingforMaaS:BikesharinginBeijing duringCovid-19pandemic 223
5.Summaryofchapter 227 References 228
8.MaaSandIoT:Concepts,methodologies,and applications
HongbinXie,XuanSong,andHaoranZhang
1.Introduction
2.Overviewoftheconcept
2.1Overviewofthegeneralconcept230
2.2ChallengesofIoTapplicationinMaaS231
3.Keytechnologiesandmethodologies 231
3.1Intelligenttransportationequipment231
3.2CommunicationprotocolsfortheInternetofThings232
3.3MicroservicesbasedontheInternetofThings232
3.4CloudcomputingbasedontheInternetofThings234
3.5Edgecomputing235
3.6SecuritytechnologiesfortheInternetofThings236
4.Applicationandcasestudy
5.Conclusionandfuturedirections
9.MaaSsystemvisualization
ChuangYang,RenheJiang,andRyosukeShibasaki
1.Overviewofthegeneralconcept
2.ThekeyvisualizationtechnologiesinMaaSfordifferent stakeholders
2.1Theperspectiveofdemandersofmobility247
2.2Theperspectiveofsupplieroftransportationservice248
2.3Theperspectiveofcitymanager252
3.Real-worldapplicationandcasestudy 254
3.1Casefordemandersofmobility254
3.2Caseforsupplieroftransportationservice255
3.3Caseforcitymanager257
3.4Open-sourcevisualizationtoolsandlibraries258
4.Conclusionandfuturedirections 261
10.MaaSforsustainableurbandevelopment
XiaoyaSong,RongGuo,andHaoranZhang
1.Introduction
2.MaaSinteractedwithurbantrafficandspace
2.1Urbantrafficstructure267
2.2Urbanspatialstructure269
3.StrategiesforMaaSinurbansustainabledevelopmentat multiplescales 269
3.1Macroscale:Synergybetweenurbanagglomerationsand metropolitanareas270
3.2Mesoscale:Optimizationofinternalresourcesincities271
3.3Microscale:Therefinementofurbanstreets271
4.Casestudy
5.Conclusion
Contributors
Numbersinparenthesisindicatethepagesonwhichtheauthors’contributionsbegin
ShreyasBharule (xiii),LocationMindInc.,Tokyo,Japan
JinyuChen (1,25),CenterforSpatialInformationScience,TheUniversityofTokyo, Kashiwa,Chiba,Japan
RongGuo (265),KeyLaboratoryofColdRegionUrbanandRuralHumanSettlement EnvironmentScienceandTechnology,MinistryofIndustryandInformation Technology,SchoolofArchitecture,HarbinInstituteofTechnology,Harbin,China
RenheJiang (245),CenterforSpatialInformationScience,TheUniversityofTokyo, Kashiwa,Chiba,Japan
WenxiaoJiang (113),CenterforSpatialInformationScience,TheUniversityofTokyo, Kashiwa,Chiba,Japan
WeifengLi (117),KeyLaboratoryofRoadandTrafficEngineeringoftheMinistryof Education,TongjiUniversity,Shanghai,China
WenjingLi (1,25),CenterforSpatialInformationScience,TheUniversityofTokyo, Kashiwa,Chiba,Japan
Wen-LongShang (203),BeijingKeyLaboratoryofTrafficEngineering,Collegeof MetropolitanTransportation,BeijingUniversityofTechnology,Beijing,China
FengjingShao (113),InstituteofSmartCityandBigDataTechnology,Qingdao,China
RyosukeShibasaki (xiii,1,25,77,245),CenterforSpatialInformationScience,The UniversityofTokyo,Kashiwa,Chiba,Japan
XiaoyaSong (265),KeyLaboratoryofColdRegionUrbanandRuralHumanSettlement EnvironmentScienceandTechnology,MinistryofIndustryandInformation Technology,SchoolofArchitecture,HarbinInstituteofTechnology,Harbin,China
XuanSong (229),DepartmentofComputerScienceandEngineering,SouthUniversity ofScienceandTechnology,Nanshan,China;CenterforSpatialInformationScience, TheUniversityofTokyo,Kashiwa,Chiba,Japan
YiSui (113,203),CollegeofComputerScienceandTechnology,QingdaoUniversity, Qingdao,China;CenterforSpatialInformationScience,TheUniversityofTokyo, Kashiwa,Chiba,Japan;InstituteofSmartCityandBigDataTechnology,Qingdao, China
RenchengSun (113),CollegeofComputerScienceandTechnology,Qingdao University,Qingdao,China;CenterforSpatialInformationScience,TheUniversity ofTokyo,Kashiwa,Chiba,Japan;InstituteofSmartCityandBigDataTechnology, Qingdao,China
HaoningXi (143),SchoolofCivilandEnvironmentalEngineering,UniversityofNew SouthWales,Sydney;Data61,CSIRO,Canberra,Australia
HongbinXie (229),DepartmentofComputerScienceandEngineering,South UniversityofScienceandTechnology,Nanshan,China
ChuangYang (245),CenterforSpatialInformationScience,TheUniversityofTokyo, Kashiwa,Chiba,Japan
DongyuanYang (177),KeyLaboratoryofRoadandTrafficEngineeringoftheMinistry ofEducation,TongjiUniversity,Shanghai,China
YuhaoYao (77),CenterforSpatialInformationScience,TheUniversityofTokyo, Kashiwa,Chiba,Japan
QingYu (177),KeyLaboratoryofRoadandTrafficEngineeringoftheMinistryof Education,TongjiUniversity,Shanghai,China
HaoranZhang (xiii,1,25,77,113,203,229,265),CenterforSpatialInformation Science,TheUniversityofTokyo,Kashiwa,Chiba,Japan
Introduction
ShreyasBharulea,HaoranZhangb,andRyosukeShibasakib
aLocationMindInc.,Tokyo,Japan, bCenterforSpatialInformationScience,TheUniversityof Tokyo,Kashiwa,Chiba,Japan
1Background
Datawasoriginallymanual.Transportationdataforthelongesttimewerecollectedthroughclickersatjunctions.However,inthelastdecade,wehaveseena rapidtransformationofthemobilitylandscape.Theconvergenceoftechnologiessuchassmartphones,sensors,theInternetofThings,andappswithin smartphoneshaverevolutionizedhowwelookateveryformoftransportation, particularlytointerpretandpredictthemovementofgoodsandpeopleinreal time.Inrecentyears,mobilityservicessuchasUberfortaxiservice,DHLfor logistics,andothershaveusedvarioustechnologiesthatcollectivelyimpact howwemoveinlong-andshort-haultransportation.
Also,thenumberofsensorsalsousedtocollectdata,bothstationaryand mobile,hasincreasedmultifold.Theyaregeneratingandcollectingrelational datatoexamineissuesandresolvechallenges.Suchaconvergenceoftechnology andtheresultantemergingservicesarechanginghowcitiesrespondtonewer typesofgoodsandpeoplemovementsthroughMobilityasaService(MaaS). Thestatedadvantagesofusingtheseservicesareconvenience,real-timeinformation,andconnectivitytogowhereverthetravelerwantstogo.Thoughsuch possibilitiesaremarketedaslimitless,therearetechnicallimitations.
Bigdata-drivenMaaSdevelopmentisanemergingareabothinacademic andindustrialaspects.Thoughseveralresearchstudiesandtechnicalreports areavailable,aclearlinktounderstandbigdatainMaaSappearsvagueand fragmented.Inthisbook,weaimtofillthisgapbysystematicallysummarizing theknowledgeinthisfield—thechaptersinthebookoutlinetheareastostreamlinetheunderstandingofbigdataanalyticsforMaaS.
2Bigdata:Definition,history,today
Bigdataismadeupoflarger,complexdatasetsthatoriginatefromincreasingly newsourcesofdata.Thesedatasetsaresomassivethattheabilitiesoftraditionaldataprocessingsoftwareareinsufficienttomanagethem.Nevertheless, thesemassivevolumesofdatacanbeusedtoaddressbusinessproblemsthat wouldnothavebeentackledbefore.
Asimpledefinitionofbigdataorganizedaroundbigdata’sthreeVscouldbe large volumes ofdatathatcontainimmense variety andaregeneratedatincreasing velocity.Recently,sixotherVsandoneChavebeenaddedtodefinethe truthfulnessandmeaningfulnessofdataas veracity,infrequencyas validity, extracting value fromthecollecteddata,replicabilityintheformof visualizations, processabilityina virtual cloudplatform,data variability,and complexity,purelyincomputationform.
Thoughtheconceptofbigdataisstillemerging,theoriginsoflargedatasets gobacktothe1970swhentheworldoforganizeddatainthefirstdatacenters wasemerging.Inaddition,thedevelopmentoftherelationaldatabasefurther consolidatedtheconcept.Around2005,researchersidentifiedjusthowmuch datausersgeneratedthroughSNS,videostreamingservices,andotheronline services.Inparalleltothesedevelopments,dataorganizationandstoragesystemsweredevelopedaswell.Hadoop,anopen-sourceframeworkcreatedspecificallytostoreandanalyzelargedatasets,wasdevelopedaroundthesame year.Thisdevelopmentofopen-sourceframeworkssuchasHadoop,Spark, andotherswasessentialforthegrowthofbigdatabecausetheymakedatastoragecheaperandenhancetheeaseofengagingwithbigdatasets.
Nevertheless,thevolumeofbigdatahasbeenatanall-timehigh,asisthe dependencyonitfordecisionmaking.Usersofthesystemarestillgenerating significantamountsofdata.However,itisnotjusthumansthatareengaging withsystems.Evolutioninsmartphoneandsensortechnologynowallows devicestocommunicatethroughtheInternetofThings(IoT)network.Besides, advancementsindevicesandcommunicationnetworksystemshaveeaseddata gatheringforvariousindicatorsanddrawingperformanceinsights.Moreover, theemergenceofartificialintelligenceandmachinelearninghaschurnedout diversityandcross-operabilityacrossvariousplatforms.
Furthermore,cloudcomputinghasexpandedbigdatapossibilitiesintruly elasticscalability.Suchscalabilityisofcriticalimportanceinanalyzingand supportingdemandsinservices.Wefocusonthisaspectofbigdatatosolve challengesthathinderfuturedevelopment,particularlyinunderstandingthe emergingareaof MaaS.
3MaaS:Definition,history,today
MaaSisanewconceptinthetransportsector;itprovidesanewwayofthinking abouthowthedeliveryandconsumptionoftransport(ormobility)aremanaged. Inordertohaveacommonstartingpointwithinthebookprojectgroup,itwas importanttohaveacommondefinitionforMaaS.AlthoughMaaSisregardedas amajorparadigmshiftintransportationtowardmoreenvironmentallyfriendly andefficientlyusedtransportmodes,describingallimportantMaaSfacetsis relevantfordeterminingthemainresearchscope.
Thoughthereisrelativelylittleliteratureontheplanningandconceptsof MaaSsystems,MaaSasaconceptwasfirstdescribedin1996asan“intelligent
informationassistant”fortravelneeds.WedescribeMaaSasashiftawayfrom privatelyownedmodesoftransportationandtowardmobilitysolutionsthatare consumedasaservice.
TheonesignificantapplicationofMaaSistheemergenceofsharedtransportationsystems(STS).STSsareenabledbycombiningtransportationservicesfromperson-to-persontransportationprovidersthroughaunified gatewaythatcreatesandmanagesthetrip.ThekeyconceptbehindSTSisto offertravelersmobilitysolutionsbasedontheirtravelneeds.Forlocalauthoritiesandpolicymakers,thepotentialtouseSTStosourcedataontravelmovementscouldopenthedoortonewtransportmanagementtools,planningfor sustainablegrowth,andmoreefficientuseofcapacity.
UK’sTransportSystemsCatapultpredictedthatbytheendof2020,more than50billionconnecteddeviceswouldcollectmorethan2.3ZBofdataglobally.Inthecomingdecade,thesenumbersareexpectedtoincreaseeachyear. STSleveragesthisopportunityandisanexampleofabusinessmodelsupported bythegrowthinsmartphoneuse.STSisadigital,data-drivenservicethatuses severaltechnologicalcapabilitiesassociatedwithintelligentmobilityandinnovation.Thesystemreliesonbuildinganecosystemofstakeholdersthatagrees tomanagethesupplyanddemandoftheservicesthattravelersdemand.
4BigdataXMaaS
Bigdata-drivenMaaSdevelopmentisanemergingtopicbothinacademicand industrialaspects.Therefore,currentlyallstudiesaboutbigdatainMaaSare fragmented.Noworkhassummarizedthesystemicknowledgeinthisfield. Thisbookisdesignedtofillthisgap.
Thebookorganizescornerstonetechnologiesinthesphereofbigdataand MaaS,focusingonintroducinghowtoscreenandprocessthepotentialvalue fromthe“deluge”ofunverified,noisy,andsometimesincompleteinformation forMaaSdevelopment.Furthermore,chaptersarewrittenintheformofsummariesoffrontiertechnologiesapplicableinMaaS,includingMaaSsystem developmentandAPPs,spatiotemporaldatapreprocessingtechnologies,travel similarityestimationandclustering,datafusiontechnologiesforMaaS,datadrivenoptimizationtechnologiesforMaaS,data-drivenestimationforurban travelshareability,MaaSsystemdataminingtechnologies,IoTtechnologies forMaaS,MaaSsystemvisualization,andMaaSforurbansustainabledevelopment(Fig.1).
WithintheambitofMaaS,thebookoffersdemonstratedanswerstoseveral questions,suchas:
a) Howtoefficientlyextractandeffectivelyutilizekeyfeatureinformationof high-dimensionalcitypeopleflowdata?
b) Howtoefficientlyexpressthemassiveandlarge-scalepeoplemovement andpredictmobilitysharingpotentialatanurbanscale?
c) Whatsolutionscanfacilitateethical,secure,andcontrolledmobilitysharing indifferenttransportationtypes?
d) Howtobuildsharedtransportationservicesbeneficialtocitizens,businesses,andsocietysafelyandresponsibly?
e) Howtoapplynoveltechnologiestosharedtransportationmanagementand MaaSdevelopment?
Furthermore,chaptersinthebookareorganizedtohelpaudiences understand:
a) Asystematicoutlinetodefineandreinventdata-drivenmobilitymodelsby studyingurbandynamics,urbanmobility,transportationbehavior,andsharingpotential.
b) Withinaglobalurbanmobilityframework,howwecancharacterizethe natureofdata-enabledMaaS.
c) UnderstandingtheexistingpositiveandsuccessfulMaaScanbeidentified inthestudieddomainanditcanbedeterminedhowbesttoapplyitforpracticalsuccess.
Collectively,theknowledgeinthisbookisofimmensesignificancefor stakeholdersinMaaSandthoseplanningtoentertheindustry,suchas researchers,engineers,operators,companyadministrators,andpolicymakers inrelatedfields,tocomprehensivelyunderstandcurrenttechnologyinfrastructureknowledgestructuresandlimitations.
FIG.1 Chapteroverviews.
5Summary
EachchapterdealswithonecomponentofbigdataconcerningMaaS.However, theoveralltakeawaysforthereadersofthisbookareseveral.First,thebook outlinesmethodstoassesstheurbanmobilitycharacteristicsaswellasthe potentialtoadoptsharedtransportationandmatchtheirperformanceasconditionsandconstraintsinatransportsystem.Theunderlyingassumptionisthat transportsystemsareundergoingautomationandarehighlydependentonsoftware,navigationsystems,andconnectivity.Second,thebookintroduceshowto designMaaSplatformsforSTSthatadapttotheevolvingmobilityenvironment,newtypesoftransportation,andusersbasedonintegratedsolutionsthat utilizethesensingandcommunicationcapabilitiestotacklethemajorchallengestheMaaSindustryfaces.Finally,thebookaimstodemonstrate:
a) Inadditiontopolicyandmarketfactors,themostsignificantdevelopment bottleneckistechnicallimitation.Thisbooksystematicallysummarizesthe currentfundamentaltechnologiesofMaaStohelppromotethedevelopment oftheMaaSindustry.Inparticular,crowddata’smultimodeurbanand inflow-outflowisofimmensevalueforurbantrafficadministrationand sharedtransportationmanagementtomonitormorediverseaspectsof mobilityandassociatedinformationforpolicyandmarketformulation.
b) TheintroductionoftechnologiesfromatechnicalstandpointcanhelppeopleunderstandthepotentialofMaaSinthefuturecitymoreintuitivelythan theconceptualoverview.Therefore,inthecontextofubiquitousdata analytics-enabledMaaS,whenthereareeffectiveenablingstructuresand processesforstakeholderparticipationandengagement,thegovernance anddecision-makingprocessofsharedtransportationserviceswillbemore effectiveandsustainable.
c) Artificialintelligence(AI),machinelearning,andbigdataareveryhotand popularbutlesser-understoodtopics.Ubiquitousdata-basedservicesbased onmoredisruptivetechnologiessuchasAIandbigdataaremorelikelyto succeedandbesustainedthansharedtransportationservicesbasedonfrugal technologiesandinflexiblemethods.Thus,thisbookiswrittentohelpthe audienceunderstandMaaStechnologiesfromtheperspectivesofAI researchersanddatascientists.
Weorganizethechaptersinthebooktocreateaneasy-to-interprethandbook thatintroducesthefundamentaltechnologiesforutilizinghumanmobilitydata todeveloptheMaaSsystemandcontributetotheliteratureandpolicydialogue aroundbigdataanalyticstoprovideMaaS.
MaaSsystemdevelopment andAPPs
WenjingLi,RyosukeShibasaki,HaoranZhang,andJinyuChen CenterforSpatialInformationScience,TheUniversityofTokyo,Kashiwa,Chiba,Japan
1ThedevelopmenthistoryofMaaS
1.1Theconception
Althoughtheterm“MobilityasaService(MaaS)”hasonlyattractedwidespreadattentioninrecentyears,theconceptofMaaShadappearedalong timeago.
AsimilarconceptofMaaSdatesbackto1996,theENTERConferencein Innsbruck,Austria [1],withtheideaofintegratingtravelservicesintoanintelligentplatform.Atthattime,NicoTschanzandHans-DieterZimmermannen visionedan“intelligentinformationassistant”platformonwhichpeoplecould searchandbooktripwaysanddomanyothertravel-relatedthingslikebooking hotelsandbuyingtickets.ConsideringthattheWorldWideWebjustdeveloped in1990andMicrosoftjustmadeitsfirstwebbrowser,InternetExplorer,toaccess theinternetin1995,oneyearbefore1996,NicoTschanzandHans-DieterZimmermann’sideawaswayaheadoftheirtime.TheirideawouldeventuallyformulateintotheMaaSsystemthatwearemovingtowardimplementingtoday.
1.2Theearlyapplication
Inthefirst15yearsofthe2000s,somediscussionsaboutMaaShaveemerged andsomeforward-lookingcompanieshavealreadystartedtheirexplorationin theMaaSbusiness.
In2006,thestartupBlaBlaCarwassetupinFrance.Itdevelopedalongdistanceride-sharingwebsiteplatformthatconnectsdriversandpassengers whowillsharethecostoflongcarjourneys.In2008,thesecondversionofBlaBlaCarwebsitewaslaunched.Itincludedacommunitymoduleallowingusers toshowtheirprofilesandbiographies,aswellasrecommendtheirpreferences. In2012,anonlinereservationservicewasadded.Althoughtheword“MaaS” BigDataandMobilityasaService. https://doi.org/10.1016/B978-0-323-90169-7.00002-6
hadnotappearedyet,BlaBlaCarhasshowntheMaaScharacteristicthatintegratedthemobilityservices.
AnotherearlyexplorerisZimride(predecessorofLyft),foundedin2007.It startedwiththeideathatmatchesdriversandpassengerswhowanttoshare ridesthroughsocialnetworkservicestoeliminatetheanxietythattheydo notknoweachotherbefore.TheearlyusersofZimrideweremainlyuniversity studentsandcompanyworkers.Zimrideonlyconnectedpeoplewhogotothe samecampusorworkatthesamecompany.
Uber,nowoneofthelargestfirmsofMaaSeconomy,isfoundedin2009.It isdevelopedforreducingtransportationfeesbysharingrides.Thebetaversion waslaunchedin2010.Originally,theapplicationonlyalloweduserstocallfor luxuryvehiclesthatthepricewashigherthanthatofataxi.Until2012,Uber alloweduserstorequestregulartaxisandpersonalvehicleswithbackground checks.
InJune2012,Agrion(the“Energy&SustainableDevelopment”brandofthe EBG,aleadingthink-tankondigitalinnovationinFrance)sponsoredahalf-day conferenceinSanFrancisco,USnamed“E-MobilityasaService.”Thediscussion topicsincludedhowtolinktheprivateautomobileandthepubliclyfundedand operatedtransitsystem,thepotentialoftheintegrationbetweenthesmartphone andsharedvehicles,theimpactofshared-usevehicles,andsoforth.The “E-MobilityasaService,”whichisverysimilartotheconceptofMaaS,highlights adigitallyconnectedseamlessmulti-modaltransportationnetwork.Throughrealtimeconnectivityfromasmartphone,mobilityservicescouldbeubiquitousand seamless.Atthatyearthatthesmartphonesbegantospreadandreplacefeature phones,thisideagavepeopleahugeimaginationofthecommercialprospects.
From2013to2014,anintegratedmobilityserviceexperimentnamedUbiGo launchedinacommercialpilotofGothenburg,Sweden.Itisthefirst-ever developmentofwhattodayiscalledMaaS.Thisprojectaimedtoexplore hownewbusinessmodelscanreducetheuseofprivatecarsandhowthe seamlessnessandmultimodalityuseofinformationtechnologycanpromote sustainabletravel.TheUbiGoachievedthisbycombiningpublictransport, bike-sharing,car-sharing,carrentalservice,taxiallintooneapplicationand oneserviceonasmartphone.Undertheprepaidmonthlysubscription,theusers canaccessallthesetravelservices.Thepublictrialoperatedlastfor6months andinvolved195users.Theservicewaswelcomedbycitizens.However,itwas discontinuedduetothelackofthegovernmentlevelsupportforthird-party on-sellingofpublictransporttickets.
Thoughalittletortuous,theseearlybusinessexplorationsandapplications helpedtodeveloptheMaaSconceptintoafeasiblefuture.
1.3MaaSalliance
InIntelligentTransportationSystem(ITS)Congressheldin2014inHelsinki, Finland,theword“MobilityasaService”wasformallyproposed.In2015,ITS
CongressinBordeaux,France,“MobilityasaService”becamethepopular topicofdiscussion.Atthatcongress,morethan20Europeanorganizations jointlyformedtheworld’sfirstMaaSAlliance.Itsmembersincludetransportationserviceproviders,publictransportationoperators,MaaSserviceoperators,ITsystemproviders,usergroups,governmentorganizations,etc.The foundationofMaaSAllianceisanincrediblysignificanteventthattheMaaS conceptiswidelyconcerned.Amongpeople,skepticismsalsobegantoarise, raisingquestionssuchas“howcoulditwork?”andwhetherornotitwasa “completerevolutionorsimplysomechangestointegrateintothecurrenttravel business”.In2016,ITSCongress,thereweresixsubforumsdedicatedtodiscussingthetechnology,businessmodel,andengineeringapplicationissues facedbyMaaS.
1.4Development
Since2016,explosivedevelopmentsofsmartphonetechnologyhavelifted MaaStoahigherintegrationlevel.
Whimwasreleasedin2016,markingthebeginningofaprototypesystem withahigherintegrationlevel.Thisplatformintegratedpublictransit,bikes,escooters,ferry,taxis,andaffordablerentalcarsinoneapp.Theuserscould searchandchoosethepreferredmodeoftransportontheirsmartphone.Trips couldbepaidontheapppertimeorbyseasonalticketsubscription.
In2017,DIDIChuxing,theMaaSgiantinChina,expandedtheirbusinessoftaxi-hailingandride-sharingtobike-sharingservicesande-bikesharingservices.Carmaintenance,refueling,andrechargingserviceswere alsoprovidedtodriver-partnersan dindependentvehicleownersonthe platform.
2018ITSCongressinCopenhagen,Denmarkprovidedparticipantswiththe useofanexperimentalMaaSprototypesystemMinRejseplan.Itwasasimilar projectbutbuiltonacasefromDenmark.Itofferedusersvarioustraveloptions togetfromAtoB—train,metroferries,sharingservices,andevenaselfdrivingshuttlebus.Electronicticketswithallpublicmodesinincludingspecial discountsforsharingservicesandtaxiswouldshowontheapp.Inadditionto theregularpublictransportservicesalloverDenmark,Rejseplanenalsointegratedseveraldemand-responsiveandsharedmobilityservicessuchasFlextrafik,taxisaswellascarpooling.
InSeptember2019,Berlin’spublictransportauthorityBerlinerVerkehrsbetriebe(BVG)launchedacity-ownedMaaSprojectnamed“Jelbi”,togetherwith aLithuanianmobilitystartupTraffic.Jelbiintegratedbus,metro,bike,motor scooter,ride-sharing,taxiinoneapp.Alltheavailablewaystogettothedestinationwereshowedandcomparedclearlybydurationandpricesothatusers couldbookjustwhattheyneedfortheoccasion,weatherconditions,price,or themood.
1.5Revolutionandinnovation
Meantime,withthedevelopmentofMaaSandtheexplosionintheamountof information,MaaSismorecloselyintegratedwithbigdata.Developmentssuch asIoTTechnologies,artificialintelligencealsopromotethistrend.Massive datawascollectedfromplatformssuchassmartphones,desktops,andotherdigitaldevicestoessentiallyreconstructMaaSservices.Manydata-drivenhightechsolutionsarecomingforthinMaaSsystems.
From2016,UberstartedtoestablishUberAILabs,aresearcharmdedicated tocutting-edgeresearchofMaaSbusinessinartificialintelligenceandmachine learning.Basedonmassiveriderecords,UberusesAIformatchingdriversand riders,routeoptimization,riskassessment,safetyprocesses,andsoforth.
In2017,MoovitlaunchedasuiteofMaaSsolutions,poweredbyAIandbig data,coveringsimulation,real-timeoperations,andoptimization.Itshowspositivesignificanceingrowingridership,increasingoperationalefficiency,and reducingurbancongestion.
In2017,DidiChuxingofficiallylaunchedAILabstoexpanditsbusinesson AI-driveninnovation.In2018,DidiChuxinglaunchedanintegratedsolution forsmartcitytrafficmanagementnamed“DiDiSmartTransportationBrain”. LeveragingcloudcomputingandAI-basedtechnologies,itcombinesvideo cameras,GPSdata,andothersensordatafromDidi’scarswithdatafromgovernmentandotherpartners.Thissmartbraincanhelpimprovetheefficiencyof drivingschedules.Forexample,utilizingreal-timemobilitydataandpredictive algorithms,itanticipatestrafficjamsandroutesitsownridersanddrivers aroundexistingtraffictoalleviatetrafficcongestion.Ride-hailingsupplyand demandindifferentregions,weatherdatesandtimecanalsobeforecasted. Thesmartbraincanalsopromotearangeoftransportationinfrastructure improvements.Forexample,byanalyzingdatafrombillionsofridesofDidi’s drivers,thesystemcouldpredictwhenandwherethetrafficjamislikelytoform andadjusttrafficlightssignaltiming.Congestionduringrushhoursdropped afterapplyingthissolution.
In2018,HaconofSiemensdemonstratedtheMyMobilitypluginfor HAFASapplication.Thispluginenableslearningtheusers’behaviorandhabits fromhistoricalrecordsdatabymachinelearningtechniquestodeliverpersonalizedandproactivemobilityrecommendations.
In2020,anintelligentbackendsystempoweredbyPTVVisumwasintegratedintotheoperationsoftheMaaSserviceofthetrafficcontrolcenterin Hamburg,Germany.Thissoftwareisnowusedforsimulatingandcalculating theimpactoftrafficdisruptions.ThefrontendsystemofROADS,asoftwarefor the coordination ofconstructionmeasuresandanalysisofchangesintraffic flow,willalsobeintegratedintothesystem.Thissystemisnowusedforcongestionforecastingforitslong-termandshort-termtrafficplanningbasedon bigdata.
Alsoin2020,ToyotaMotorCorporationannouncedtosetupanewbusiness alliancewithNTTDATA,aleadingITservicesproviderinJapan.Thisalliance aimstoaccelerateToyota’sMaaSinitiativesbyoperatingitsclouddatacenter forbigdatacollectedfromconnectedcars.
2ThecategoryofMaaSsystem
Indifferentstudies,MaaSsystemcanbecategorizedaccordingtodifferent ways.Forexample,accordingto (1)transportationtype,MaaScanbecategorizedintobike-hailing,ride-sharing,bus-sharing,trainservices,micromobility,etc.;accordingto (2)servicestype,MaaScanbecategorizedinto journeymanagement,journeyplanning,booking,navigation,flexibletransactions,personalizedcustomization,etc.;accordingto (3)solutiontype,MaaS canbecategorizedintotechnologyplatforms,navigationsolutions,payment engines,ticketingsolutions,telecomconnectivityproviders,etc.;according to (4)transportationtype,MaaScanbecategorizedintopublicorprivate; accordingto (5)businessmodel,MaaScanbecategorizedintobusinessto-business,business-to-government,business-to-consumer,andpeer-to-peer; accordingto (6)operatingsystem,MaaScanbecategorizedintoAndroid, iOS,andPC.
Sochor,Arby [2] characterizedMaaSsystemfromLevels0to4ascharacterizedbydifferentlevelsofservicesintegration,thatisnointegration,integrationofinformation,integrationofbookingandpayment,integrationofthe serviceoffer,andintegrationofsocietalgoals(Fig.1).Serviceintegrationis thepreconditionofMaaS.Basedonthisfoundation,theintegrationofinformation,data,platform,etc.arerealized.ThecategorywayofSochor,Arby [2] is representativeandhasbeenwidelyrecognizedbyMaaS-relatedresearch.
FIG.1 SummaryofMaaSintegrationlevel.
2.1Level0:Nointegration
Atthislevel,theservicesproviderprovidesseparateservicesoftransportation mode.Theservicessystemissingle.Itcanprovidematureserviceforcertain travelmodesbutlacksintegrationwithothertravelmodes.Thereisnodata interactionbetweendifferenttransportationmodes.Hellobike(bike-sharing), Hertz(carrental),NipponRent-a-Car(carrental),andSunfleet(car-sharing) areexamplesofservicesprovidersofthislevel.
2.2Level1:Informationintegration
TheMaaSserviceofthislevelhasacentralizedinformationplatformthatcan provideinformationcomparisonandtravelrecommendationsofmultipletransportationmodes.Ithelpsuserstofindthebesttrip.Ittypicallyfocusesonasingletripratherthansinglecustomers.RepresentativeservicesincludeGoogle andAutoNavi.Thestandardizedinformationistypicallyforfreeandeveryuser canuseitsservice.Someinformationsuchaspublictransport,cateringinformation,andreservationinformationwouldbeintegratedintotheservicesplatformofthislevel.However,theplatformoperatoronlyactsasaninformation collectorandaconnectorbetweentheuserandthechosenprovider,ratherthan theservicedirectlyoperator.Theapplicationofthisleveldoesnotreflectthe coreconceptofMaaSandhasnodirecteffectsinpushingtravelerstoabandon privatecarstochoosepublictransport.
2.3Level2:Integrationofbookingandpayment
Applicationofthislevelalsofocusesonasingletrip.Basedonprovidingtravel planningservices,theyprovidetravelerswithsearch,reservation,andpayment servicesofpublictransit,taxi,bike,oranyothertypeoftravelmode.RepresentativeservicesincludeMoovit,HANNOVERmobile,andXiecheng.TheadditionalvalueofLevel2isthatitenablesservicesfrommultipleoperatorseasier tobeaccessedto.Userscanfind,compare,book,andpaywiththesameapp.
Theapplicationofthislevelismainlyresponsibleforticketing,booking, andpurchasebuttheytypicallyarenotdirectlyresponsibleforthepassenger andcargotransportationprocess.Theapplicationofthisleveldoesnotreflect thecoreconceptofMaaStoo.Theyjustprovidetheintegrationofpaymentbut havenodirecteffectsinpushingtravelerstoabandonprivatecarstochoose publictransport.
2.4Level3:Integrationoftheserviceoffering
RepresentativeservicesofthislevelincludeUbiGoandWhim.Applicationof thislevelfocusesontheuser’stotalmobilityneedsratherthanonlysingletrip fromoneplacetoanotherplace.TheMaaSoperatortakestwo-way
responsibilitytobothend-userandsupplier.TheMaaSoperatorbuyservices fromdifferenttransportservicesproviders,thenreorganize,integrate,andsale themtotheend-user.
Thereistwo-sidedataintegrationatthislevel.Ontheonehand,astheMaaS operatortakesresponsibilityfortheservicedeliveredtoitscustomers,theinformationoftheuserside,suchasmobilityneeds,mobilitypreferences,and mobilitypatterns,shouldbecarefullyintegrated.Itisnotonlyforasingleuser butprobablyforthewholefamilyorthewholecommunity.Ontheotherhand, MaaSoperatorsshouldbuildagoodconnectionwiththetransportservicessupplier.Forexample,fortaxi-hailing,theMaaSoperationshouldanalyzethe bookingdatatohelpbettervehiclescheduling.
However,thisdoesnotmeanthatthedataintegrationlevelishigherthan level2.Becausetheservicesofthislevelaretypicallylocal-based,MaaSoperatorneedstofindthebestsupplierofeachmodetodeveloptheservicewith. Duetofewersuppliersandlessinteraction,thecomplexityofthetechnicalintegrationcanbelowerforaLevel3servicethanforaLevel2service.Besides, Level3serviceusuallyneedsmorecooperationwiththeregionalorlocalpublic transportauthoritiestofindpoliticallyacceptablecontractmodels.
2.5Level4:Integrationofsocietalgoals
TheMaaSserviceofthislevelintegrateslocal,regional,nationalpoliciesand goals.Forexample,publictransportationisintegratedintotheMaaScustomizablepackageswithcommercialservices.TheMaaSservicesoperatorcanprovideincentivesfordesiredtravelmodetoinfluencethetravelbehaviorofthe user.Besides,traditionalpublictransportserviceisaone-size-fits-allservice withnon-flexiblebusinessmodels,whileanattractiveMaaSserviceneedsto beunifiedandflexible.
Integrationofsocietalgoalsrequiresthehighintegrationofdata.Itisnot onlyintegratingdatafromtheend-usersideortransportservicessupplierside butalsofrompublicinfrastructure,publicspace,andpublictransportation.The MaaSserviceofthislevelcanhelpreducedprivatecarownership,improvethe transportationefficiencyandpromoteamoresustainableandlivablecity.
3Studycase
ThedevelopmentofMaaShasbroughtnewopportunitiesandchallengesto transportationinvariouscountries.Atpresent,variouscountriesandregions areactivelyinvolvingtheimplementationandapplicationofMaaS.Dozens ofAppserviceshavebeenlaunchedworldwide.Jittrapirom,Caiati [3] summarizedtheprojectsthathavebeenrunorpilotedinthefirstfewyearsofMaaS development.Ontheirbasis,wehaveupdatedtheseprojectsandaddednew applicationresultsinrecentyears. Table1 showsthesummaryofsomeof theintegratedMaaSsystemsaroundtheworld.