Multi-Level Progressive Reinforcement Learning for Control Policy in Physical Simulations | IEEE Conference Publication | IEEE Xplore