Trajectory-model-based reinforcement learning: Application to bimanual humanoid motor learning with a closed-chain constraint | IEEE Conference Publication | IEEE Xplore