Locomotion Control Method for Humanoid Robot Based on United Hierarchical Reinforcement Learning | IEEE Conference Publication | IEEE Xplore