A model of learning free bipedal walking in indefinite environment constraints self-emergence/self-satisfaction paradigm

We propose a bipedal walking model based on C-SESS paradigm. This model only needs desired velocity to walk. It can begin to walk from a standstill and change to steady state immediately. A walk is maintained by the real-time control when mechanical perturbations are added. Our results suggest that the generated pattern can adopt to unpredictable environment. This model based on C-SESS paradigm is easily extensible to voluntary movement control model.