Real-time Transfers for Improving Efficiency of Ridesharing Services in the Environment with Connected and Self-driving Vehicles

Current ridesharing services typically support one-hop delivery. With potential vehicle transfers, the ridesharing efficiency can be enhanced. This paper develops a real-time transfer approach based on a centralized pubic taxicab scheduling system. The approach can improve the traffic efficiency by leveraging real-time trajectory information of vehicles maintained in the system. The system can search for a suitable location for the passenger who could be transferred, and calculate the subsequent routes for original and transfer taxicabs. The simulation results based on the road networks of Midtown Manhattan demonstrate the effectiveness of the developed scheme compared to the transfer-disallowed scheme. Several passengers can be transferred to other vehicles to save more travel time and travel distance. Taxicabs on the roads can be used more efficiently.