Adaptive Reconnaissance and Suppressive Jamming Strategy Generation Based on Reinforcement Learning

In modern electronic warfare, developing intelligent and adaptive radar jamming methods to counter frequency agile radar (FAR) is crucial, as FAR can unpredictably change frequencies. Traditional jamming methods are unable to adjust jamming strategies in real time against the frequency hopping patterns of FAR. To address this issue, an adaptive reconnaissance and suppressive jamming strategy generation method of the unmanned aerial vehicle (UAV) based on reinforcement learning (RL) is proposed in this article. First, a system model, including a UAV reconnaissance model and a suppressive jamming model, is established to quantify the reconnaissance and suppressive jamming effectiveness against FAR. Next, the UAV reconnaissance and suppressive jamming process is formulated as a Markov Decision Process (MDP). Especially, the reward function is designed as a weighted sum of the reconnaissance effect factor and the suppressive jamming effect factor to flexibly balance the reconnaissance and suppressive jamming missions of UAV. Finally, based on the Q-Learning algorithm, we propose an adaptive UAV reconnaissance and suppressive jamming strategy that can timely acquire frequency domain information of FAR and achieve efficient suppressive jamming. Simulation results show that, compared to the traditional fixed jamming strategies, the proposed strategy can flexibly counter various FAR frequency hopping cases.