Problem Solving with Deep Q-Learning at IEEE RoboResearch