Reinforcement Learning-based Car-Following Control for Autonomous Vehicles with OTFS

The orthogonal time frequency space (OTFS) modulation is regarded as a promising technology to fully release the potential of integrated sensing and communication (ISAC) systems. In this paper, we propose a reinforcement learning (RL)-based car-following control, i.e., adaptive cruise control (ACC), a method using the popular OTFS-ISAC technology in vehicular networks. In particular, the sensing parameters can be inferred from the wireless communication channels in the delay-Doppler (DD) domain, enabling a comprehensive understanding of the surrounding environment. Next, we design an RL-based framework for dynamic car-following with autonomous vehicles. Finally, the simulation results show the effectiveness of the proposed RL-based control method in terms of safety and decision-making efficacy.