Determination of the Central Pattern Generator Parameters by a Neuro-Fuzzy Evolutionary Algorithm

Abstract

The selection of appropriate parameters of a Central Pattern Generator (CPG) facilitates its use on applications in robotics. A Central Pattern Generator is class of oscillator represented by a system of n differential equations of first order with m parameters, finding those m parameters to match a specific behavior is a complex problem to solve. For the development of applications in robotics, such as locomotion, the CPG have been applied as a decentralized control system. In this work, a Neuro-Fuzzy Evolutionary procedure is proposed to determine the parameters of a CPG for a specific cases of 2, 3, 4, 5 and 6 neurons. The developed methodology is presented in detail. The proposed algorithm guarantees a given response in the frequency and amplitude into the ranges required.