基于强化学习的功率与信道联合干扰方法研究

Joint Jamming Method of Channel and Power based on Reinforcement Learning

主要研究了一种基于强化学习的无线通信智能干扰方法。目前,传统的干扰大多为单一域干扰。而伴随着越来越复杂的电磁频谱环境和抗干扰技术的不断发展,单一域内干扰难以在复杂环境中起到较好的干扰效果。同时,通信用户的侦察系统会接收干扰机能量值对其进行侦察定位,随即利用相关抗干扰技术减弱干扰。为提升复杂电磁环境中的干扰效果并降低干扰机被定位概率,结合强化学习提出了一种功率与信道联合干扰决策算法。干扰机可以通过学习和训练在动态变化环境中决策出最佳干扰功率和干扰信道。仿真结果表明,在所给条件下,所提算法可以收敛到最佳干扰策略,相较于随机干扰算法和不考虑定位因素的Q学习干扰算法,所提算法在确保干扰有效性的同时,干扰机被定位的概率分别降低了30%和60%。

This paper mainly discusses an intelligent jamming method of wireless communication based on reinforcement learning.At present,most of the traditional interference is single-domain interference.With the development of increasingly complex electromagnetic spectrum environment and anti-jamming technologies,single-domain interference is difficult to play a better jamming effect in complex environment.At the same time,the reconnaissance system of communication users will receive energy value of the jammer for reconnaissance and positioning,and then use the relevant anti-jamming technology to reduce interference.In order to improve the jamming effect in complex electromagnetic environment and reduce the probability of jammer being located,a joint power and channel interference decision algorithm is proposed on the basis of reinforcement learning.The jammer can autonomously decide the best jamming power and jamming channel in the dynamic environment by learning and training.Simulation results indicate that the proposed algorithm can converge to the best interference strategy.Compared with random jamming algorithm and Q-learning jamming algorithm without considering location factors,the proposed algorithm can ensure the effectiveness of jamming and reduce the probability of jamming being located by 30%and 60%respectively.