An LLM-Explainable DRL Framework for Passenger-Directed Autonomous Driving

This article presents a novel framework that integrates deep reinforcement learning (DRL) with large language model (LLM) explainability for autonomous driving systems. The DRL agents, trained using a Dueling Double Deep Q-Network, effectively adapt to driving requests such as "fast," "comfort," and "stop," while LLM modules provide real-time explanations of the agents' behaviors to passengers. This approach enhances public trust in autonomous vehicles by improving transparency and safety, making it significant for practitioners focused on developing explainable AI in transportation.