Research on path planning of quadruped robot based on globally mapping localization

Perception and navigation control is core key technology for the legged robot to adapt to complex terrain and achieve autonomous walking. It is the key to distinguish the legged robot from manned equipment and become a ground unmanned system. In this paper, the SLAM globally localization system based on the lidar point cloud is established, and a localization framework based on the combination of topological measurements is proposed to realize the map reconstruction and localization of the wild environment. Path planning based on Dijkstra algorithm is proposed to implement the globally terrain autonomous navigation task of the quadruped robot, and autonomous obstacle avoidance strategy of the local map for quadruped robot based on the artificial potential field theory is applied considering the motion and vibration of legged robot. The experimental results show that the globally path navigation can accurately plan the optimal path and complete the autonomous obstacle avoidance of the local map. The results have demonstrated the effectiveness of the proposed method and realize the globally field autonomous walking of the quadruped robot.