Stable biped locomotion using central pattern generators based on MATSUOKA neural oscillators

Abstract

In this paper we address a control framework using central pattern generators (CPG) for the stable rhythmic biped locomotion. The mutually inhibiting and coupled nonlinear Matsuoka Neural Oscillators are used to generate the required signals to realize the coordinated movement of a musculoskeletal model of biped robot with 6 degrees of freedom. The proposed design of the CPG is verified through simulation. We present a comparative study between the proposed model and Taga's CPG model to find more stable and effective walking for the biped robot.