Improving the Dynamic Window Approach for Path Planning in Multi-Legged Robots

Path planning is a crucial component of navigation and obstacle avoidance for mobile robots in complex terrains. In response to this necessity, this paper designs a multi-legged robot navigation and obstacle avoidance system based on an improved A* algorithm and an enhanced Dynamic Window Approach (DWA) algorithm. The system uses Jetson Nano as the core controller, running an Ubuntu operating system with the ROS (Robot Operating System) installed. The multi-legged robot continuously scans the surrounding environment with a LiDAR sensor and utilizes an adaptive Monte Carlo localization method to achieve autonomous navigation and obstacle avoidance. Finally, the proposed method is validated through simulations and experimental debugging. The results demonstrate that the system can effectively and accurately localize itself, achieving autonomous navigation and obstacle avoidance.