Walking Pattern Generation Using Quintic Spline Function Based on Human Motion Capture

Abstract

In this paper, we propose a biped gait generation method based on human motion capture. The data of human waist are obtained through the motion capture system. Then we analyze the position, velocity and acceleration curve of the waist trajectory and obtain the relationship between the maximum speed of waist and step size. The quintic spline function is used to plan the waist joint trajectory of the robot. We also analyze the trajectory of the swinging foot. The process is divided into five stages and the spline parameters of each interpolation point are given by motion capture. Then we compare quintic spline method with LIPM interpolation, MPC and DCM. The advantages of quintic spline method are analyzed. This method is finally applied to our robot and achieve a very stable gait.