Using a controller based on reinforcement learning for a passive dynamic walking robot | IEEE Conference Publication | IEEE Xplore