Toward Privacy-Aware Multimodal Transportation: Convergence to Network Equilibrium under Differential Privacy

Abstract

The development of multimodal mobility systems (MMSs) to improve transportation services in underserved areas has garnered increasing attention. In order for MMSs to identify demand and provide flexible mobility services to each consumer in real time, a vast array of private mobile data, including the locations of consumers and vehicles, trip records, and other characteristics, must be collected. The widespread deployment of MMSs will expose customers and service providers to the risk of privacy breaches if sensitive information is not carefully managed. Therefore, integrating privacy-preserving techniques and technology into MMSs has become a pressing research area. In this paper, we first create a novel joint differential privacy mechanism for sharing user data in the adaptive hyperpath choice problem. Commuters can learn the optimal hyperpaths, the combination of multiple modes of transport and their routes, based on this encrypted data. Finally, we evaluate the performance of the proposed method via numerical simulations. The suggested mechanism concurrently accomplishes the objectives of preventing retracing of the sensitive user identity in the event of data breaches and ensuring that the iterative hyperpath choice can asymptotically converge to the network equilibrium.

Keywords: Multimodal mobility systems, differential privacy, network equilibrium, hyperpath choice, privacy-preserving mechanism