Parameter tuning of CPGs for hexapod gaits based on Genetic Algorithm

In the Central Pattern Generator (CPG) inspired locomotion control, it is difficult to set appropriate weight values of oscillator network and initial values of each oscillator to generate the desired pattern with specific amplitude, frequency and phase lag. In this paper, a novel method tuning the parameters of CPGs for hexapod gaits based on Genetic Algorithm (GA) is presented. And the CPG model discussed in this work is based on Van Der Pol (VDP) equation. The process of parameter-tuning has two stages:(i) searching parameter values of single VDP oscillator to generate periodic oscillation sign with specific amplitude and frequency. (ii)searching synaptic weight values of oscillator network and initial values of each oscillator to generate specific phase lag. Two experiments have been taken in MATLAB and the results show the feasibility of the presented method in searching appropriate parameters to generate different hexapod gaits.