Humanoid robot locomotion

Abstract

The objective of this paper is to propose a locomotion control method for humanoid robot to realize stable walking, which consisting of optimal joint generation and feedback error compensation. Using energy consumption as criteria, the proposed method can generate a series of joint trajectory for different walking steps and speed. Furthermore, to improve system stability and adaptability, a fuzzy-GA is trained to compensate system errors. Based on sensor information, humanoid robot can get environment and inherent situation and use the proposed method to adjust joint trajectory. The effectiveness of the method is shown with a 28 DOF humanoid robot in ADAMS.