A Study of Comfort Measuring System Using Taxi Trajectories Li-Ping Tung 1, Tsung-Hsun Chien 2,3, Ting-An Wang 3, Cheng-Yu Lin 3, Shyh-Kang Jeng 2, and Ling-Jyh Chen 3 1 National Chiao Tung University, Taiwan 2 National Taiwan University, Taiwan 3 Academia Sinica, Taiwan 1
Introduction The comfort or rides has been identified as one of the top criteria that affects passengers’ satisfactory with public transportation system. 2 Comfort does matter!!
How to Measuring it? 3 Questionnaire/Interview Professional Instruments Problems: Cost, Timeliness, and Scalability
Internet of Things The idea of IoT is to interconnect state-of-the-art digital products in physical world to provide more powerful applications. intelligent transportation systems remote healthcare systems smart grid systems 4
Vehicles and the Trajectory Data Vehicles are view as parts of Internet of Things GPS devices allow recording the movement track of moving vehicles. The collected trajectory data could be real-time transmitted to the data server via wireless technologies, such as WiMAX and 4G LTE. Applications of trajectory data provide passengers with the expected trip time and fare of a given itinerary predict driving directions supervise urban traffic or serve location-based services 5
Comfort Measuring System Exploit the GPS data Calculate the comfort index by following ISO 2631 Comfort Score: 20 x (6 – CI) 6 Acceleration Level uncomfortablecomfortable
Taxi Trajectory Dataset One of the Taipei service providers Duration: 2010/11/8~2010/11/28 Objects: 200,000 trajectories among about 700 taxis 7 fielddata typedescription idintsequence number micmacchartaxi number longitudedoublelongitude of trajectory latitudedoublelatitude of trajectory speeddoubledriving speed datatime driving time clientontaxiboolload/uoload
Statistical Results of Dataset (1) 8 Among 24 Hours Among a WeekAmong 24 Hours
Statistical Results of Dataset (2) 9 85% is under 30 minutes for passengers saving time low fare for drivers risk of no load in the returning trip 9 Trip Time Driving Time
Comfort Scores in CDF Distribution 10 comfortable
Comfort Scores Analysis - Day and Night 11 without passengerswith passengers Comfort Scores: 1. day > night 2. w/o passengers > w/ passengers
Comfort Scores Analysis – Trip Time 12
Comfort Scores Analysis – Trip Distance 13
Ranking of Load among a Day Ranking lists according to some criteria number of loads comfort score 14
Ranking of Comfort Score 15 The 10 BEST The 10 WORST
Implication from Ranking Lists Track back to the trajectories to understand what happened drivers’ driving behaviors road conditions traffic conditions 16
Conclusions We present a Comfort Measuring System for vehicles equipped with GPS devices. It shows that comfort level varies with trip time/distance w/ and w/o passengers Ranking lists according to comfort score and number of loads Work on spatial-temporal analysis is ongoing (e.g., road conditions, drivers’ behavior, and traffic congestion). 17