CPG-based motion generation for bipedal robot using stochastic optimization

This paper proposes a stochastic approach to CPG-based motion generation of bipedal robot. We make an extension to a motion control system based on neuro-musculo-skeletal model into three dimensions, and give an optimization of control parameter by simulated annealing. This paper also provides some empirical verification of our approach by performance of various walking patterns adaptive to different link structures and the bias of objectives, such as walking distance, gait stability, and so on.