Stability analysis of a simple walking model driven by a nonlinear oscillator

Walking motion of a human is achieved through the interaction between the dynamics of the human mechanical system and the rhythmic signals of the central pattern generator (CPG). We analyze dynamic properties of a simple walking model of a biped robot driven by a nonlinear oscillator. The simple model consists of a hip and two legs which are connected at the hip and has touch sensors at the tips of the legs. The swing leg is controlled by the rhythmic signal of the oscillator. The oscillator receives feedback signals from the touch sensors and modifies the walking motion according to the signals. From the analysis of the stability of the periodic walking motion using a Poincare map, it is revealed that the simple model has the stabilization property and moreover the stability region is enlarged due to the signal feedbacks.