Neural Oscillators with a Sigmoidal Function for the CPG of Biped Robot Walking

Abstract

Two types of neural oscillators with the degree of adaptation, Matsuoka oscillator and McMillen oscillator, are compared each other, and their differences in the system structure are discussed. The Hopf bifurcation analysis is revisited in detail to a half-center oscillator in the latter oscillator. Simulation results are compared each other when the both oscillators are applied as central pattern generator (CPG) controllers for controlling a bipedal robot with five links and six joints.