Experimental realization of body balance control of quadruped walking robot, pongbot, on slopes

This paper describes the body balance control for ground motion with changing slopes for the quadruped walking robot PongBot. First, the robot is simply modeled as an inverted pendulum model, and the stabilizing torque to balance the inverted pendulum against the ground slope is calculated from the observed state feedback controller. To practically implement the stabilizing torque, a strategy is proposed to distribute the vertical ground reaction forces acting on the supporting feet through the quadratic programming (QP) optimization method. Feedforward and feedback control methods are then used simultaneously to quickly generate the distributed vertical ground reaction forces. Finally, the performance of body-balance control of the quadruped walking robot PongBot is verified on a treadmill with changing slopes.