Trajectory planning and posture adjustment of a quadruped robot for obstacle striding

Abstract

This paper proposes a control method of trajectory planning and posture adjustment for quadruped robots' obstacle striding. The addressed mixed parabola method plans the travelling paths over obstacles in Cartesian coordinate system. The constraints on velocities, accelerations, and jerks at waypoints are employed to generate the time-efficient smooth cubic spline joint trajectories by nonlinear optimization technique. The planned trajectories maximize the compliance and flexibility of joint movements. To guarantee the static stability for the pitch-pitch type quadruped robots, a posture adjustment strategy based on the potential energy is investigated to regulate the trajectory of Center of Gravity (COG). The experimental results on quadruped robot FROG-I (Four-legged Robot for Optimal Gaits) prove the feasibility of the proposed approach of striding a board in natural environment.