A Physics-Driven Closed-Loop Motion Planning Method for Spherical Multi-Expandable-Limb Robots

The spherical multiexpandable foot robot showcases exceptional characteristics, including omnidirectional movement, wide deformation range, and effective radial force output. These features enable its suitability for applications such as disaster relief, combat reconnaissance, cave exploration, and planetary surface exploration, highlighting its remarkable terrain adaptability. However, existing locomotion planning approaches lack integration with sensor data, limiting real-time control, and resulting in path tracking errors for complex, curved paths. To address this problem, a closed-loop locomotion planning scheme leveraging sim to real techniques is proposed. The proposed scheme dynamically adjusts the subsequent motion plan of the robot based on real-time changes in its center position, resulting in more accurate path tracking. By bridging the gap between simulated training and real-world applications, the sim to real approach empowers the spherical multiexpandable foot robot to overcome challenges, enhance control capabilities, and operate successfully in diverse environments.