Weather-Aware Power Control for Decentralized V2V Communication

Vehicular networks rely on vehicle-to-vehicle (V2V) communication to enable real-time information sharing for collision avoidance, traffic efficiency and future autonomous driving systems. In V2V communication, vehicles constantly exchange messages; therefore, developing a power control method that saves energy and reduces interference is essential. Previous power control mechanisms typically rely on static configurations and neglect dynamic environmental factors such as weather conditions; some require feedback channels that may not be available in distributed V2V scenarios. As a result, those methods struggle to maintain reliable communication in rapidly changing conditions, particularly during adverse weather events. To address these limitations, this paper presents a novel weather-aware reinforcement learning-based power control algorithm to improve the robustness and efficiency of V2V communications. Unlike conventional open-loop power control schemes that rely solely on SINR-based heuristics, our algorithm incorporates weather-induced attenuation and vehicle density into the transmit power decision process to ensure reliable communication during link setup. Furthermore, we propose a reinforcement learning framework in which each vehicle autonomously learns optimal power control strategies over time by observing local SINR estimations, surrounding vehicle densities, and environmental variations. The learning agent dynamically adjusts the transmission power to maximize communication reliability and minimize interference between vehicles. We conducted simulations based on 3GPP TR37.885 scenarios, and the results show that our solution effectively addresses the impact of weather on power control in V2V networks and achieves better reliability and adaptability under challenging weather conditions.