Autonomous gait switching method and experiments of a hexapod walking robot for Mars environment with multiple terrains

Abstract

Mars exploration significantly advances our understanding of planetary evolution, the origin of life, and possibilities for Earth’s future. It also holds potential for discovering new mineral resources, energy sources, and potential settlement sites. Navigating Mars’ complex environment and unknown terrain is a formidable challenge, particularly for autonomous exploration. The hexapod walking robot, inspired by ant morphology, emerges as a robust solution. This design offers diverse gait options, mechanical redundancy, high fault tolerance, and stability, rendering it well suited for Martian terrain. This paper details the development of an ant-inspired hexapod robot, emphasizing its terrain adaptability on Mars. A novel terrain detection method utilizing a convolutional neural network enables efficient identification of varied terrain types through semantic segmentation of visual images. Additionally, we introduce a comprehensive motion performance evaluation index for the hexapod robot, including speed and stability. These metrics facilitate effective performance assessment in different environments. A key innovation is the proposed gait switching method for the hexapod robot. This approach allows seamless transition between gaits while in motion, enhancing the robot's ability to traverse challenging terrains. The experimental results validate the effectiveness of this method. Utilizing gait switching leads to a significant improvement in robot performance and stability—58.5% and 41.4% better than using tripod and amble gaits, respectively. Compared to single tripod, amble, and wave gaits, the comprehensive motion performance indices of the robot improved by 36.3%, 30.6%, and 41.1%, respectively. This study can provide new ideas and methods for the motion evaluation and adaptive gait switching of multilegged robots in complex terrains. It significantly enhances the mobility and adaptability of such robots in challenging environments, contributing valuable knowledge to the field of planetary exploration robotics.