Message Interval Control for Efficient Resource Utilization in Dynamic Platoons Using Multi-Agent Reinforcement Learning

This paper addresses the pressing need for efficient communication and stable operation within autonomous vehicle platoons. An integrated approach is proposed for aperiodic message transmission with efficient resource utilization through Multi-Agent Reinforcement Learning (MARL). The proposed system optimizes message transmission timing, considering parameters like age of information, vehicle speed, and acceleration, which are achieved through the Q Learning algorithm. A robust control function adapts to real-world uncertainties, ensuring safe vehicle spacing and stable platooning dynamics. Simulation results indicate increased communication intervals efficiently utilize available resource blocks among six vehicles.