High-Accuracy Path Tracking for Unmanned Vehicle Navigation: A Hierarchical Deep Reinforcement Learning Approach

This paper proposes a novel hierarchical deep reinforcement learning-based path tracking algorithm for unmanned vehicle navigation. To achieve effective acceleration and steering control, our approach divides the multi-variable control task into distinct sub-tasks handled at different levels, enabling more efficient learning and improved policy optimization. Simulation results show that the proposed algorithm achieves superior tracking accuracy compared to existing methods. Moreover, training models on complex reference paths allows the proposed algorithm to generalize effectively, maintaining high performance across diverse, untrained driving environments. These findings reveal the robustness and adaptability of the proposed algorithm, demonstrating the potential for application in diverse autonomous driving scenarios.