Conference ProceedingGeneration of Time-Varying Feedback-Based Wheel Lock Attack Policies with Minimal Knowledge of the Traction DynamicsMohammadi, Alireza; Malik, HafizThere are a variety of ways, such as reflashing of targeted electronic control units (ECUs) to hijacking the control of a fleet of wheeled mobile robots, through which adversaries can execute attacks on the actuators of mobile robots and autonomous vehicles. Independent of the source of cyber-physical infiltration, assessing the physical capabilities of an adversary who has made it to the last stage and is directly controlling the cyber-physical system actuators is of crucial importance. This paper investigates the potentials of an adversary who can directly manipulate the traction dynamics of wheeled mobile robots and autonomous vehicles but has a very limited knowledge of the physical parameters of the traction dynamics. It is shown that the adversary can exploit a new class of closed-loop attack policies that can be executed against the traction dynamics leading to wheel lock conditions. In comparison with a previously proposed wheel lock closed-loop attack policy, the attack policy in this paper relies on less computations and knowledge of the traction dynamics. Furthermore, the proposed attack policy generates smooth actuator input signals and is thus harder to detect. Simulation results using various tire-ground interaction conditions demonstrate the effectiveness of the proposed wheel lock attack policy.Springer2023-08-2010491965Intelligent Computing. SAI 2023. Lecture Notes in Networks and Systemshttps://doi.org/10.1007/978-3-031-37963-5_872035770National Science Foundation