IRL-DAL: Safe and Adaptive Trajectory Planning for Autonomous Driving via Energy-Guided Diffusion Models

TL;DR

This paper introduces IRL-DAL, a diffusion-based inverse reinforcement learning framework for autonomous driving that enhances safety, robustness, and adaptability through imitation learning, environment feedback, and a safety supervision model.

Contribution

It presents a novel combination of diffusion models, IRL, and adaptive perception to improve safe and adaptive trajectory planning in autonomous vehicles.

Findings

Achieved a 96% success rate in safe navigation tasks.

Reduced collision rate to 0.05 per 1,000 steps.

Established a new benchmark for safe autonomous driving in simulation.

Abstract

This paper proposes a novel inverse reinforcement learning framework using a diffusion-based adaptive lookahead planner (IRL-DAL) for autonomous vehicles. Training begins with imitation from an expert finite state machine (FSM) controller to provide a stable initialization. Environment terms are combined with an IRL discriminator signal to align with expert goals. Reinforcement learning (RL) is then performed with a hybrid reward that combines diffuse environmental feedback and targeted IRL rewards. A conditional diffusion model, which acts as a safety supervisor, plans safe paths. It stays in its lane, avoids obstacles, and moves smoothly. Then, a learnable adaptive mask (LAM) improves perception. It shifts visual attention based on vehicle speed and nearby hazards. After FSM-based imitation, the policy is fine-tuned with Proximal Policy Optimization (PPO). Training is run in the Webots simulator with a two-stage curriculum. A 96\% success rate is reached, and collisions are reduced to 0.05 per 1k steps, marking a new benchmark for safe navigation. By applying the proposed approach, the agent not only drives in lane but also handles unsafe conditions at an expert level, increasing robustness.We make our code publicly available.