Towards autonomic urban traffic control with collaborative multi-policy reinforcement learning | IEEE Conference Publication | IEEE Xplore