A Q-Learning Optimized Cluster-Based Framework for Energy-Efficient Communication in 6G VANETs

Abstract

The future of intelligent transportation now rests on the integration of Vehicular Ad-hoc Networks (VANETs) with powerful wireless or 6G wireless standards. However, most of the very features that make 6G revolutionary—immense data rates and dense connectivity—also threaten its sustainability by drastically increasing energy consumption. In this paper, we tackle this problem with a novel framework that combines two intelligent techniques. First of all, our Stable Clustering Algorithm (SCA) creates robust network groups by selecting leaders based on a balanced score of their remaining battery, driving patterns, and connection stability. Second, our Reinforcement Learning-based Adaptive Sleep Scheduling (RASS) protocol utilizes a Q-learning model to make informed decisions, enabling roadside units and vehicles to enter low-power "sleep" modes during periods of inactivity without missing critical data. Extensive simulations show our combined framework is remarkably effective. It reduces total network energy use by up to 41.5% and extends network lifetime by 34.8% compared to modern benchmarks. It must look at EECB and ECRP, all while maintaining the strict quality of service required for ultra-reliable, low-latency applications.