Multi-robot Cooperation Based on Continuous Reinforcement Learning with Two State Space Representations | IEEE Conference Publication | IEEE Xplore