Joint Optimizing High-Speed Cruise Control and Multi-Hop Communication in Platoons: A Reinforcement Learning Approach

A platoon is a group of vehicles traveling in the same direction at a reasonable speed while maintaining a reasonable and safe distance. Robust model predictive control addresses vehicle stability, but evolving communication topology is a significant challenge for high-speed autonomous vehicles. A multi-hop communication model is proposed to improve communication quality within power and channel constraints, considering the Doppler effect and changes in vehicle spacing due to high speeds. This paper uses a Markov decision process to model communication and designs a reward function in reinforcement learning to enhance communication quality while reducing power usage. The efficiency of the learning process is enhanced by narrowing the search range for the signal transmission power of vehicles, informed by the characteristics of the communication model. The proposed model and algorithm’s effectiveness and feasibility are confirmed through thorough simulations.