Efficient reinforcement learning for humanoid whole-body control | IEEE Conference Publication | IEEE Xplore