Privacy-Preserving Task Scheduling for Time-Sharing Services of Autonomous Vehicles

Sharing of autonomous vehicles between multiple users can potentially be the ultimate solution for increasing the efficiency of the transportation system. In a time-sharing scenario, autonomous vehicles (AV) owners share their vehicles to others at their unwanted times without incurring any human efforts. However, such sharing service requires the disclosure of users' locations and route information, raising severe privacy concerns and issues. In this paper, we propose a privacy-preserving task scheduling scheme for time-sharing services of autonomous vehicles. First, we design a matching scheme that finds the feasible requesters for each available AV. Then, we propose a scheduling scheme using different approaches for assigning requesters to the AVs based on different system parameters. Both schemes work efficiently without requiring users to share their exact locations and route details while maximizing the AV owner profit and minimizing the requester cost. Specifically, our schemes enable an untrusted matching server to match owners and requesters using a set of intermediate destination (ID) locations in the travel paths. Only if the service can be given to the requester efficiently, the owner and the requester share the details of the routes. All the calculations for verification of the feasibility of the service are done on the untrusted server. We show the accuracy of our proposed scheme through extensive simulations on real data. The results confirm that our traffic-based ID selection scheme, with consideration of the traffic information and patterns in the area, outperforms the baseline scheme where the IDs are chosen randomly. Furthermore, the effectiveness of different scheduling schemes including greedy-based and first come first served are evaluated.