Using reward-weighted imitation for robot Reinforcement Learning | IEEE Conference Publication | IEEE Xplore