An Anti-jamming Intelligent Decision-Making Method for Multi-user Communication Based on Deep Reinforcement Learning

Abstract

In order to effectively cope with the external malicious jamming from jammers and avoid mutual interference caused by competitive channels among users, an intelligent antijamming decision-making method based on deep reinforcement learning for multi-user communication is proposed in this paper. In multiple jamming modes, the base station perceives the current spectrum information of the multi-user and the jammer and makes it as the input of the deep reinforcement learning strategy neural network, then selects the joint action according to the dynamic greedy algorithm, which will help the users intelligently select the communication frequency band via the feedback from the base station. The simulation results show that the proposed method effectively overcome the communication jamming induced by external jammers and internal users. Moreover, the normalized throughput of each user in the proposed algorithm is much higher than the random frequency hopping and fixed ε deep reinforcement learning algorithm with faster convergence speed.