Gait planning in 3D robot simulation using ZMP theory

Abstract

This paper presents a method of modeling robot legs with linear inverted pendulum model based on Zero-Moment Point (ZMP) theory. It quantitatively plans the foot's target pose matrix during the walking of a robot. Our method follows principles of inverse kinematics to get the anti-solution of joints based on Newton-Raphson algorithm and Jacobian matrix, therefore, it amends and optimizes the ZMP. We demonstrate the feasibly of our algorithm in robot's gait planning using various simulations.