Towards training an agent in augmented reality world with reinforcement learning | IEEE Conference Publication | IEEE Xplore