Verification Of Adversarially Robust Reinforcement Learning Mechanisms In Autonomous Systems