Impact of V2V Communication on Robustness of Autonomous Driving Systems

Autonomous Driving Systems (ADS) hold significant promise in preventing traffic accidents and alleviating urban traffic congestion. A critical component of the ADS software stack is vehicular networking, which enables vehicles to achieve mutual perception and cooperative driving through wireless communication technology. Ensuring the reliability of this communication is crucial for the safety of autonomous driving systems. This paper focuses on the robustness of mutual perception in vehicle communication, specifically investigating the impact of variations and disturbances in the communication range on safety and potential hazards. Through an empirical study, we examine the degree of information sharing among vehicles as a variable and observe the likelihood of collisions during interactions within an autonomous driving cluster. Our experimental outcomes reveal that variations in the interconnection system can lead to significantly different planning consequences, underscoring the potential risks associated with ADS under imperfect communication conditions.