机器学习辅助共享出行.pdf

1、ITAP AI WorkshopJune 4 2023Machine Learning Aided Shared MobilityMachine Learning Aided Shared MobilityDepartment of Civil&Coastal EngineeringLili Du,Ph.D.,Associate Professor University of Florida TRB Innovations in Travel Analysis and Planning 2023 workshop:W8:Demystifying Machine Learning June 4,

2、2023ITAP AI WorkshopJune 4,20232Data in Transportation SystemVast and diverse spatiotemporal traffic dataFixed point traffic data from loop detectors,camera,etc.Trajectory data from probe vehicles.Traffic Prediction(Data+Physical Model+ML)DataDataTraffic and events prediction including sports,constr

3、uction,and accidents.Hybrid mobility serviceConnected and autonomous vehicle control.Traffic data communication through V2V or V2I.Mobility data from mobile appsTraffic Domain Knowledge Traffic Management and Control(Data+Physical Model+ML+Operations Research)Traffic congestion mitigation:EV/CAV rou

4、ting and platooning,ridesharing;information provision.Data+physical model extract significant features for AI and ML optimization and controlML Aided Traffic Prediction,Control,and ManagementHKUST Seminar May 2nd,2022Ride-hailing(solo trip)causes congestionThe advantages of autonomous vehicles(AVs)L

5、ow occupancy leads to congestionSolution:Promote Carpooling/ridesharingRidesharing(such as Uber and Lyft)improves mobilityand reduces private car ownership rateLarger capacityRoute flexibilityIncreased safetyLess ParkingPeng,Wang,and Lili Du.Investigating Optimal Carpool Scheme by a Semi-Centralized

6、 Ride-Matching Approach.IEEE Transactions on Intelligent Transportation Systems(2022).DOI:10.1109/TITS.2021.3135648Community Learning Based Semi-Centralized Optimal Carpool/Ridesharing SchemeHKUST Seminar May 2nd,2022Research ChallengesCarpooling(i.e.,ridesharing)may often cause detour,currently hav