An evolutionary central pattern generator for stable bipedal walking by the increased double support time

Central pattern generator (CPG) consisting of neural oscillators, generates rhythmic signals using simple input signal. It can modify motor patterns to handle environmental perturbations by sensory feedback. In this paper, an evolutionary CPG for stable bipedal walking by the increased double support time is proposed. The proposed CPG generates swing motion of arms as well as ankle and the center of pelvis (COP) motions in Cartesian coordinate system. Sensory feedback pathways in the proposed CPG use force sensing resistor (FSR) signals. The sensory feedback maintains humanoid robot´s balance and prevents it from falling down to the ground. To optimize the parameters of the proposed CPG, evolutionary algorithm is employed. The effectiveness of the scheme is demonstrated by simulations with the Webot model of a small-sized humanoid robot, HSR-IX, developed in the RIT Lab., KAIST.