Walking Control Framework of a Radial-Type Quadruped Robot

This paper deals with a walking control framework for a radial-type quadruped robot. Our goal is to achieve stable blind walking using a minimum number of sensors. Therefore, this work presents a robot control method using three estimation methods. The first is the contact estimation algorithm which calculates whether the robot's foot has landed. The second is the body state estimation algorithm. The third is the slope estimation algorithm for calculating ground slopes. We generate gait patterns based on contact estimation and user command. In addition to the gait pattern, the robot controls the posture while walking through the force control of the stance feet. Finally, we verify the algorithm using GAZEBO simulation and apply it to ST-Quad, a radial-type quadruped robot, to test walking stability.