A Q-learning based adaptive congestion control for V2V communication in VANET

Vehicular ad hoc networks (VANETs) require timely delivery of periodic basic safety messages (BSMs) containing critical vehicle status information, as well as event-driven notifications to ensure road safety and improve traffic flow. The limited channel capacity of the wireless medium and high message rates needed for adequate situational awareness create a dilemma between the conflicting goals of congestion control and awareness control algorithms. To ensure reliable delivery, vehicles need to interact with a complex and dynamic environment to determine the appropriate message rate and power for their transmissions at any given time. In this paper, we propose an innovative framework where vehicles use reinforcement learning (RL) to intelligently select their transmission parameters based on the current channel conditions. Our simulation results indicate that RL methods can provide a flexible solution for adaptive congestion control by designing the appropriate reward function.