Deep reinforcement learning method for POMDP based tram signal priority | IEEE Conference Publication | IEEE Xplore