The Central Pattern Generator: a paradigm for artificial locomotion

Abstract

 In this paper the fundamental aspects that unify the basic programs for locomotion generation in both vertebrates and invertebrates are examined. Subsequently it is shown that these laws that are commonly included under the biological stereotype of the so-called Central Pattern Generator (CPG), can be efficiently represented and reproduced by means of low-cost, locally interacting nonlinear systems and programmed by soft-computing techniques. To demonstrate the suitability of the approach some simple Central Pattern Generators are implemented on biologically inspired robotic prototypes.