Cooperative Adaptive Cruise Control Based on Reinforcement Learning for Heavy-Duty BEVs

Advanced driver assistance systems (ADAS) are playing an increasingly important role in supporting the driver to create safer and more efficient driving conditions. Among all ADAS, adaptive cruise control (ACC) is a system that provides consistent aid, especially in highway mobility, guaranteeing safety by minimizing the possible risk of collision due to variations in the speed of the vehicle in front, automatically adjusting the vehicle velocity and maintaining the correct spacing. Theoretically, this type of system also makes it possible to optimize road throughput, increasing its capacity and reducing traffic congestion. However, it was found in practice that the current generation of ACC systems does not guarantee the so-called string stability of a vehicle platoon and can therefore lead to an actual decrease in traffic capacity. To overcome these issues, new cooperative adaptive cruise control (CACC) systems are being proposed that exploit vehicle-to-vehicle (V2V) connectivity, which can provide additional safety and robustness guarantees and introduce the possibility of concretely improving traffic flow stability.