2017/03/06 by Justin Miller, Jonathan P. How, Miller, Justin +1
Engineering · Social Sciences · #FOS: Computer and information sciences #Human Mobility and Location-Based Analysis #Robotics (cs.RO) #Transportation Planning and Optimization #Transportation and Mobility Innovations
paper · pdf · doi:10.48550/arxiv.1703.02130
openalex publication_date 2017/03/06 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
In autonomous Mobility on Demand (MOD) systems, customers request rides from\na fleet of shared vehicles that can be automatically positioned in response to\ncustomer demand. Recent approaches to MOD systems have focused on environments\nwhere customers can only request rides through an app or by waiting at a\nstation. This paper develops MOD fleet management approaches for ride hailing,\nwhere customers may instead request rides simply by hailing a passing vehicle,\nan approach of particular importance for campus MOD systems. The challenge for\nride hailing is that customer demand is not explicitly provided as it would be\nwith an app, but rather customers are only served if a vehicle happens to be\nlocated at the arrival location. This work focuses on maximizing the number of\nserved hailing customers in an MOD system by learning and utilizing customer\ndemand. A Bayesian framework is used to define a novel customer demand model\nwhich incorporates observed pedestrian traffic to estimate customer arrival\nlocations with a quantification of uncertainty. An exploration planner is\nproposed which routes MOD vehicles in order to reduce arrival rate uncertainty.\nA robust ride hailing fleet management planner is proposed which routes\nvehicles under the presence of uncertainty using a chance-constrained\nformulation. Simulation of a real-world MOD system on MIT's campus demonstrates\nthe effectiveness of the planners. The customer demand model and exploration\nplanner are demonstrated to reduce estimation error over time and the ride\nhailing planner is shown to improve the fraction of served customers in the\nsystem by 73 % over a baseline exploration approach.\n