Deep Reinforcement Learning for Decentralized Multi-Robot Control: A DQN Approach to Robustness and Information Integration