Hexapod Adaptive Gait Inspired by Human Behavior for Six-Legged Robot Without Force Sensor

Abstract

Legged robots are usually installed with force sensors in order to negotiate with the uneven ground. To eliminate the risk of force sensor failure, adaptive gaits integrating indirect force-estimation are of great importance. Robot Octopus III is a robot designed for carrying a payload in the harsh environment. There is no electric device installed on the lower limb of the robot. The indirect force-estimation method, which is based on its spatial parallel mechanism leg, can estimate the external force exerted on the foot tip. In this paper, an adaptive gait is designed after observing human actions. Experiments are carried out to observe how a human walks through the uneven ground when his/her eyes are covered. A static tripod gait mimics the human behavior during the blind walking. When the foot collides with the obstacle, the robot will adjust the foot’s height and try to overcome the obstacle. Just like human blind walking, the robot foot tries different locations before it steps on somewhere. The gait also detects if the robot is facing a ditch too deep to step or an obstacle is too high to step on. The gait is implemented in the real-time control system. Experiments are carried out to validate the proposed gait. The robot walked through the uneven ground with maximum obstacle height of 0.2m successfully. The proposed gait enables the robot without foot tip force sensors to walk through the field with obstacles.