Periodic Cubic Spline on Motion of Five-Link Human Bipedal Model Using Nonlinear Predictive Control

This paper studies about the performance of periodic cubic spline applied on the nonlinear predictive control (NPC) based on receding-horizon cost function optimization. The five-link human bipedal model includes a torso and two legs which has a thigh and a shank in every leg. The periodic cubic spline is used to achieve smooth walking trajectories of every joint in the bipedal model in the sagittal plane. The results show that the tracking error after using periodic cubic spline is smaller than the tracking error before using periodic cubic spline.