다관절 해저보행로봇 경로 계획을 위한 선행 연구

Abstract

This study presents a preliminary investigation into implementing RRT* (Rapidly-exploring Random Tree Star) algorithm-based path generation and tracking using a quadruped walking robot. Optimal paths were generated using the RRT* algorithm based on the robot's odometry and estimated position information, considering obstacles, narrow passages, and external forces for efficient path exploration. Experiments were conducted in an indoor corridor environment using the Unitree B1 robot, and the generated paths were optimized through path smoothing techniques. The experimental results demonstrated that the RRT* algorithm can stably generate paths and enable the robot to efficiently navigate to the target destination, even in the presence of obstacles and narrow passages. This study provides a foundation for future applications in underwater mine detection and removal using underwater walking robots.