The Design of Central Pattern Generators Based on the Matsuoka Oscillator to Generate Rhythmic Human-Like Movement for Biped Robots

Abstract

We propose a controller based on a central pattern generator (CPG) network of mutually coupled Matsuoka nonlinear neural oscillators to generate rhythmic human-like movement for biped robots. The parameters of mutually inhibited and coupled Matsuoka oscillators and the necessary interconnection coupling coefficients within the CPG network directly influence the generation of the required rhythmic signals related to targeted motion. Our objective is to analyze the mutually coupled neuron models of Matsuoka oscillators to realize an efficient CPG design that leads to have dynamic, stable, sustained rhythmic movement with robust gaits for bipedal robots. We discuss the design of a CPG model with new interconnection coupling links and its inhibitation coefficients for a CPG-based controller. The new design was studied through interaction between simulated interconnection coupling dynamics with six links and a musculoskeletal model with the 6 degrees of freedom (DOFs) of a biped robot. We used the weighted outputs of mutually inhibited oscillators as torques to actuate joints. We verified the effectiveness of our proposal through simulation and compared the results to those of Taga's CPG model, confirming better, more efficient generation of stable rhythmic walking at different speeds and robustness in response to disturbances.