Navigation Scheme of Tracked Robots Based on LiDAR SLAM and Improved A* Algorithm

Tracked robots are often deployed in environments characterized by narrow passages and uneven terrains, necessitating efficient and accurate path planning to ensure reliable navigation. This paper presents a comprehensive navigation framework based on the SC-A-LOAM LiDAR SLAM system and the OctoMap-based environment modeling approach. A two-dimensional grid map is constructed, followed by obstacle inflation to enhance environmental safety margins. To improve the quality of global path planning, an enhanced A* algorithm incorporating a directional penalty mechanism is proposed to minimize excessive turning behaviors, while cubic spline interpolation is employed to achieve curvature-continuous path smoothing. Furthermore, to address the challenges posed by dynamic environments, the improved A* algorithm is integrated with the Dynamic Window Approach (DWA), enabling coordinated global path generation and real-time local obstacle avoidance. Both simulation and real-world experimental results demonstrate the effectiveness of the proposed method in enhancing navigation efficiency, trajectory smoothness, and overall system safety for tracked robotic platforms.