Dynamic Ridesharing Strategy for Shared Autonomous Vehicles Based on Maximum Similarity of Trajectories

As a convergence of “self-driving” and “sharing”, shared autonomous vehicles can offer real-time ridesharing service to the common people, and have potential to transform current traffic patterns. This paper focuses on a new dynamic ridesharing strategy for shared self-driving cars, including system technical framework, workflow to operate for service provider and matching model between customers and cars, according to the similarity of their predicted trajectories, service provider would use fewer vehicles to meet all the requests by matching ride demands and cars, if they have similar trajectories. A computational experiment results showed that, this strategy can increase vehicle occupancy and alleviate congestion.