Stochastic policy gradient reinforcement learning on a simple 3D biped | IEEE Conference Publication | IEEE Xplore