<ul><li>Reinforcement Learning (RL) has shown excellent performance in solving decision-making and control problems of autonomous driving.</li><li>However, current RL methods face challenges in achieving multi-objective compatibility for autonomous driving.</li><li>To address this, a Multi-objective Ensemble-Critic RL method with Hybrid Parametrized Action is proposed.</li><li>Experimental results demonstrate that the method improves driving efficiency, action consistency, and safety while increasing training efficiency.</li></ul>

Hybrid Action Based Reinforcement Learning for Multi-Objective Compatible Autonomous Driving

Discover more