Privacy-Preserving Deep Reinforcement Learning in Vehicle AdHoc Networks

Abstract

The increasing number of road vehicles results in more fatalities and accidents. Thus, the manufacturing industry is working on driver safety to secure and safe transportation in Vehicle Adhoc networks. In addition, the mobile vehicles run in the geographical zone and communicate roadside units over the wireless medium with a certain radius. The Internet of Vehicles has become a new network type where vehicles communicate with the application over public networks. This results in an increase in data exploration and threats related to network security. We propose the deep reinforcement learning method to sensitize the private information for a given vehicle connect over Vehicle Adhoc networks, maintaining a balance between security and privacy through any sanitization process. Furthermore, we provide a set of recommendations and potential applications for the Vehicle Adhoc networks as use cases.