An Impact Evaluation of Strategic Cyberattacks on Autonomous Vehicles: Safety, Mobility, and Energy Consumption

While security has always been an important factor in the transportation industry, the increasing development and availability of autonomous driving technologies have brought a higher level of importance to safety and security due to malicious cyberattacks. Autonomous vehicles (AVs) are prone to cyberattacks which may not only have negative effects on the vehicle but also further degrade the performance of traffic flow. Following our recent work, in this study we consider the adversarial fashion and stealthy nature of potential attacks on a platoon of vehicles to study their impacts on traffic performance. Specifically, we first present a general framework describing mixed traffic involving AVs and human-driven vehicles (HVs) based on car-following dynamics. Using this framework as the baseline scenario for simulating a platoon of 10 vehicles, we introduce a variety of linear and nonlinear attacks on AVs in additive and multiplicative fashion. The impacts of strategically designed attacks on traffic safety, mobility, and energy consumption are holistically examined, using metrics like time-to-collision (TTC), average speed variation (ASV), and fuel consumption (FC). A series of simulations are conducted across a range of AV market penetration rates (MPRs). Results show that strategically designed attacks tend to increase in impact with the growth of MPR, amplifying traffic instability and vehicle FC while remaining stealthy (without causing direct collisions). More severe, nonstrategic attacks result in greater levels of traffic oscillations and FC while also jeopardizing the safety of some vehicles, with collisions in certain scenarios. This study offers useful insights into understanding the impacts of strategic attacks on future automated transportation systems.