Defense Strategies Against Adversarial Jamming Attacks via Deep Reinforcement Learning

As the applications of deep reinforcement learning (DRL) in wireless communication grow, sensitivity of DRL based wireless communication strategies against adversarial attacks has started to draw more attention. In order to address such sensitivity and alleviate the resulting security concerns, we in this paper study defense strategies against DRL-based jamming attacker on a DRL-based dynamic multichannel access agent. To defend the jamming attacks, we propose three diversified defense strategies: proportional-integral-derivative (PID) control, the use of an imitation attacker and the development of orthogonal policies. We design these strategies and evaluate their performances.