StudyonUtilisationofProbeDatainMetropolitanCitiesinIndia.pdf智能交通世界大会ITS智慧城市社区人工智能A.pdf

《StudyonUtilisationofProbeDatainMetropolitanCitiesinIndia.pdf智能交通世界大会ITS智慧城市社区人工智能A.pdf》由会员分享，可在线阅读，更多相关《StudyonUtilisationofProbeDatainMetropolitanCitiesinIndia.pdf智能交通世界大会ITS智慧城市社区人工智能A.pdf（17页珍藏版）》请在淘文阁 - 分享文档赚钱的网站上搜索。

1、-最新资料推荐最新资料推荐-StudyonUtilisationofProbeDatainMetropolitanStudyonUtilisationofProbeDatainMetropolitanCitiesinIndia.pdfCitiesinIndia.pdf 智能交通世界大会智能交通世界大会 ITSITS 智慧城市智慧城市社区人工智能社区人工智能 AIAI 物联网物联网 ITIT 报告课件教案报告课件教案 22 nd ITS World Congress,Bordeaux,France,59 October2019 Paper number ITS-1852 Study on Uti

2、lisation of ProbeData in Metropolitan Cities in India Noboru Kondo,HiroyaTotani,Takatsugu Deguchi,1.East Nippon Expressway CompanyLimited,Japan,3-3-2Kasumigaseki,Chiyoda-ku,Tokyo100-8979,+81-3-3506-0298,n.kondo.aae-nexco.co.jp 2.NipponKoei Co.Ltd.,Japan 3.Zenrin Data Com Co.Ltd.,JapanAbstract Bengal

3、uru is a capital city of Karnataka state inIndia.It is known as IT capital and important industryhub insouth India.The traffic congestion is becoming serious problemin Bengaluru due to rapid urbanisation and it hinders smoothbusiness activities.Under such situation,developing ITS inBengaluru metropo

4、litan area is planned,assisted by JapaneseInternational Cooperation Agency(JICA).It aims to mitigatethe traffic congestion by guiding vehicles from congested routetoalternativerouteandenhancingmodalshiftthroughdisseminating realtime traffic information.The probe system1 1/1717is considered to collec

5、t realtime traffic data because of theadvantagethatitdoesnotrequirealargeamountofinfrastructurecostwithlessmaintenanceinsteadofinstalling a number of roadside equipment.A technical studyon probe data(referred to as the Study hereinafter)wasconducted using sample probe data available in Bengaluru toe

6、xamineusabilityoftheprobedataunderthelocationcondition.Keywords:Busprobedata,taxiprobedata,Bengaluru in IndiaStudy PurposeThe probe system isconsidered as one of major data collection methods for realtimetraffic information provision for ITS in Bengaluru.The probesystem is widely prevailed in the de

7、veloped countries and theprobe data is used for generating congestion information.However,such system has not been in existence yet in India.A number of issues are expected to prepare and operate the probesystemandutilisetheprobedatatogeneratetrafficcongestion information in the developing countries

8、 such asIndia because of the local conditions which are significantlydifferent from the developed countries.The examples of thelocalconditionsarelimitedavailability/reliability ofcommunication,trafficcharacteristiccomposedofapproximately70%oftwowheelers,underdevelopedroad-最新资料推荐最新资料推荐-infrastructure

9、,and etc.It is important to identify thetechnical issues,understand how they affect and overcome theseissues to use the probe data as a reliable data source inBengaluru.This paper describes the usability of bus probe dataas data source for realtime congestion information provisionby comparing with t

10、axi probe Study on Utilisation of Probe Datain Metropolitan Cities in India 2 data.Background and Reasonof Selecting Bus Probe for Data Source Bengaluru MetropolitanTransport Corporation(referred to as BMTC hereinafter)is agovernmental city bus company in Bengaluru.They are operatingapproximately 6,

11、700 buses(referred to as city bus hereinafter)in Bengaluru.A bus location system is planned to develop byBMTC and GPS devices will be installed on 6,700 buses.Theconcept of the information provision system in Bengaluru is toutilise the probe data obtained from these city buses as animmediate method

12、inearly stage andto use the probe data fromother vehicles such as taxi,commercial vehicles and etc.inthe next step as necessary.The reasons of selection of busprobe data as first step are:Covering the major roads in thecity at relatively constant speed The city buses run on thedetermined major route

13、s in the city as scheduled.On the other3 3/1717hand,the running routes of other vehicles such as taxi are notdetermined.They can travel on any roads e.g.minor roads inthe residential area.In addition,their behaviours are nowherein sight.They may stop by at the gas stations,stay aroundhotels for a lo

14、ng-time,and etc.The basic idea of theinformation provision system in Bengaluru is to initiatecovering the major roads in the city in the beginning stage andto gradually increase density of data collection later.In thelight of the basic idea,the probe data of the city buses isconsidered more appropri

15、ate.Expected large amount of probedata from city bus As stated above,there are approximately6,700 buses in the city.When the GPS devices are installed onall these buses,a large amount of probe data is expected tobe obtained.Furthermore,the figure below shows the runningfrequency of the city buses on

16、 the major roads in the city.Consideringtherunningfrequencyonthemajorroads,sufficient amount of the probe data can be expected.Figure-1Running Frequency of City Bus on Major Roads in BengaluruStudy on Utilisation of Probe Data in Metropolitan Cities inIndia 3 Study Focus The study was carried out fo

17、cusing on thebus(referred to as bus probe data hereinafter)by comparing-最新资料推荐最新资料推荐-with the probe data of taxi(referred to as taxi probe dataeousness in terms of transmission ofpositioning data from GPS device to centre server for promptcollection in terms of constantly stable data volume ProbeDat

18、aUsedfortheStudyDataSource,DataTransmissionFrequency,Number of Probe Vehicle The study used sample probedata obtained from bus,taxi and commercial vehicles and theywere compared.The sample probe data used for the study is shownin the table below.The bus location system of BMTC is nowunder trial oper

19、ation.The GPS devices have been installed on200 buses for the trial.The probe data collected from 200 citybuses was providedby BMTC for this study.The probe datacollected from taxi and commercial vehicle was obtained througha private company.All of the probe data is being used fortheir vehicle monit

20、oring system.The analysis of the study wascarried out on off-line basis using the shared sample probe dataasshownbelow.Table-1DataSource,DataTransmissionFrequencyandProbeVehicleNumberDataSourceDataTransmission Frequency from Probe Vehicle to Centre Number of5 5/1717Probe Vehicle Taxies(*)1 minute 60

21、0 Commercial Vehicle(*)1 minute 1,200 City Bus 10 seconds 200 Total-2,000 Note(*):The data collected from taxi and commercial vehicles aretreated together with and referred to as taxi probe data in thispaper.Period of Sample Probe Data Used for the Study The studyused the sample probe data for four(

22、4)days,which wascollected from 24th(Mon)to 27th(Thu)November 2019.Thereason is that all above probe data,bus,commercial vehicleand taxi probe data,was available during this period forcomparison.Methodology The study took the followingmethodology:Study on Utilisation of Probe Data in MetropolitanCiti

23、es in India 4 Confirmation of Availability of Probe Datain Terms of Time Trend Characteristics Availability of bus andtaxi probe data by hour was confirmed based on the timestampsof the collected probe data.The comparisons of time trendbetween bus and taxi were made for(i)unique vehicle number(*),(i

24、i)collected probe data amount,and(iii)average speed.The data was summarised by half an hour by hour.Note(*):Uniquevehicle number means the number of vehicle,either bus or taxi,which are running at a time.Confirmation of Area Coverage ofProbe Data The probe data coverage depends on the service areaco

25、vered by the mobile communication service provider because-最新资料推荐最新资料推荐-the probe data is sent from GPS device through mobile network.The data coverage is also affected by running routes of theprobe vehicles.The area coverage of the probe data was examinedby plotting on the map.The area coverage of

26、the taxi probe dataand bus probe data was respectively confirmed.Finding TimeDelay Characteristics in Terms of Data Transmission The probedata contains two times:(i)position measurement time(timerecorded by GPS device)and(ii)server time(time recorded onthe server).The difference between these two ti

27、mes is definedas Delay-Time.As the data is written on the server sent fromGPS device through mobile network and Internet line,thedelay-time is affected by such communication lines if occurs.The data continuity was examined focusing on the delay-time forthe specific day.If discontinuity is observed,i

28、t can beassumed that it is the time of occasion when some kind ofproblems occurred on any of server system,Internet line ormobile communication network.The entire Bengaluru was coveredand morning and evening of weekdays were checked.It is notedthat the taxi probe data was used for this analysis beca

29、use itcovers wider area in the city.Examination of Correlationbetween Delay-Time and Congestion(Travel Time)by Visualising7 7/1717Probe Data The specific software was developed to visualise thelocationandspeedofthecollectedprobedata.Itdistinguishes the sample on the map by different colours by thefo

30、llowing steps:(i)All collected probe data is stored indatabase.(ii)The stored probe data is retrieved from databaseby specifying condition such as hour,day,type of probe,i.e.bus probe/taxi probe,etc.(iii)The retrieved result isplotted on the map and they are marked by different colours.It visualises

31、 the summary of probe data in terms of the scaleof delay-time on the map.The visualised scale of delay-timeis calculated by aggregating the probe data weighted bydelay-time.The travel time on the road is also visualised bythe congestion level,i.e.smooth,slow,heavy,critical andalmost stopped.The visu

32、alised travel speed is drawn from thedriving recorder data at identical time of the probe data.Study on Utilisation of Probe Data in Metropolitan Cities inIndia 5 Study Results Availability of Probe Data in Terms ofTime Trend Characteristics Unique Vehicle Number The figurebelow shows the time trend

33、 of number of unique vehicle of busand taxi respectively,the number of probe vehicles which wererunning in a day.It was confirmed that a certain number of taxivehicles were running during late night hours too.On the other-最新资料推荐最新资料推荐-hand,only limited numbers of buses were running during latenight

34、hours.This is considered due to the limited operationof the city bus during night.Figure-2 Time Trend of UniqueVehicle Number Available Probe Data The figure below shows thehourly distribution of the available bus and taxi probe data.Figure-3 Hourly Distribution of Probe Data Collected by Busesand T

35、axies The amount of bus probe data increases during daytimewhilst the amount of taxi probe data stays relatively steadilythroughout a whole day.For example,the difference in amountof the available taxi Study on Utilisation of Probe Data inMetropolitan Cities in India 6 probe data between daytime and

36、midnight is approximately two(2)times.It becomes almost ten(10)times in the case of bus probe data.This implies that mostof the city buses are parking at bus depot during midnight hoursand available probe data will be limited accordingly.Howevermuch larger amount of the bus probe data than taxi prob

37、e datawas collected during daytime even the number of the probevehicle equipped with GPS are limited to 200 city buses.Thisis because of higher frequency of data transmission than taxiand frequent bus operation during daytime in Bengaluru.Eventhough the amount of bus probe data becomes small during9

38、 9/1717midnight,available bus probe data during daytime can beconsidered as a potential major data source as long as thetarget time of information provision is set out to daytime atthe initial stage.Furthermore,it is noted that BMTC isplanning to install GPS devices on 6700 buses in Bengaluru inthen

39、earfuture.However,itshallalsobenotedthatfluctuation in terms of collected amount of bus probe data issignificant during daytime whereas the taxi probe data issteadily collected.Average Speed The figure below shows thetime trend of average speedof taxi and bus.The average speedof taxi is fasterthan b

40、us.However the differenceof theaverage speed between taxi and bus is limited to approximately10%to 15%during daytime.It can be said that such differencemay not be significant in terms of utilising the bus probe dataasdatasourceforinformationprovision.05101520253035404550Time Trend of Average SpeedTa

41、xi Probes BusProbesFigure-4TimeTrendofAverageSpeedStudyonUtilisation of Probe Data in Metropolitan Cities in India 7 AreaCoverage of Probe Data The area coverage of the taxi probe dataand bus probe data was confirmed.The figures below show thecoverage areas during morning peak hours and noon for the

42、 taxiand bus respectively.Figure-5 Area Coverage of-最新资料推荐最新资料推荐-Probe Data in Bengaluru:Taxi and Bus The taxi probe data isdistributed throughout the city whilst the bus probe data isconcentrated on the north area of the city.This is because thatthe taxies travel on various routes including small r

43、oads insuch areas as residential zone in the city.Meanwhile,thenumber of buses equipped with GPS device is limited to 200 busesin this study;accordingly the routes which are covered by thesebuses are limited in the north area of the city.However,itwas confirmed that the major routes of the bus were

44、covered.It is judged that the major routes in the entire Bangaluru willbe covered when GPS are installed on 6,700 buses.It was alsoconfirmed that the mobile network was available in almostentire area of Bengaluru from the result of the area coveredby the taxi probe data.Taxies 7:30-8:00 Buses 7:30-8

45、:00Taxies 13:30-14:00 Buses 13:30-14:00 Study on Utilisation ofProbe Data in Metropolitan Cities in India 8 Time DelayCharacteristics in Terms of Data Transmission The delay timeof each sample of probe data was calculated and categorized.The contribution ratio was assigned as to percentage againstto

46、tal probe data.The result was weighted and represented byweighting value.This is for the purpose of plotting the data1111/1717by the visualisation system,described later.The result issummarised as shown below.Table-2 Delay Time of Probe Data andContribution Ratio Delay Time Number of Probes Contribu

47、tionRatio(%)Weighting Value 000000000010 7,749 15,912 45.5 93.41 000011000020 2,816 16.5 2 000021000030 2,083 12.2 30000310000401,296 7.6 4 0000410000501,095 6.4 5000051000100 873 5.1 6 000101000200 491 491 2.9 2.9 7000201000500 407 407 2.4 2.4 8 000501001000 183 183 1.11.1 9 001001002500 24 24 0.1

48、0.2 10 002501010000 3 3 0.0010001120600 14 14 0.1 120191 240000 1 1 0.0 17,03517,035 100.0 100.0 The total amount of probe data was 17,035.The probe data with less than 10 second ofdelay time accountsfor 45.5%and less than 1 minute 93.4%.Correlation betweenDelay-Time and Congestion(Travel Time)by Vi

49、sualising ProbeData Locations of Longer and Frequent Delay-Time in the CityThe figure below shows the delay-time scale of the probe databy location visualised on the map of Bengaluru(*).The redcoloured in bigger size represents longer and more frequentdelay time.The purple coloured in smaller size i

50、ndicatesshorter and less frequent delay time.It is observed thatlonger and more frequent delay tends to occur on exterior areaof the city and shorter and less frequent delay in the central-最新资料推荐最新资料推荐-area of the city.Note(*):The result is visualised by thesoftware,called Mobmap software.It is a GI