Safe Reinforcement Learning with Scene Decomposition for Navigating Complex Urban Environments
This addresses the problem of safe and efficient autonomous navigation in urban environments with multiple traffic participants, representing an incremental improvement over existing approaches.
The paper tackles autonomous vehicle navigation in complex urban intersections by proposing a modular decision-making algorithm that combines safe reinforcement learning with scene decomposition. The result is an algorithm that empirically outperforms both rule-based methods and existing reinforcement learning techniques on complex intersection scenarios.
Navigating urban environments represents a complex task for automated vehicles. They must reach their goal safely and efficiently while considering a multitude of traffic participants. We propose a modular decision making algorithm to autonomously navigate intersections, addressing challenges of existing rule-based and reinforcement learning (RL) approaches. We first present a safe RL algorithm relying on a model-checker to ensure safety guarantees. To make the decision strategy robust to perception errors and occlusions, we introduce a belief update technique using a learning based approach. Finally, we use a scene decomposition approach to scale our algorithm to environments with multiple traffic participants. We empirically demonstrate that our algorithm outperforms rule-based methods and reinforcement learning techniques on a complex intersection scenario.