Research on Resource Allocation Strategy of One-to-Many Radar Jamming Based on Reinforcement Learning
SHANG Xi1，
YANG Gewen2，
DAI Shaohuai2，
JIANG Yilin1
1. School of Information and Communication Engineering， Harbin Engineering University， Harbin 150001， Heilongjiang， China；2. Shanghai Electro-Mechanical Engineering Institute， Shanghai 201109， China
Abstract：Aiming at the interference penetration of the jammer in the case of one-jammer to multi-radar, a reinforcement learning-based interference resource allocation method in the case of one-jammer to multi-radar interference is proposed. The interference radiation energy ratio and penetration distance ratio are introduced as evaluation indicators, and the dynamically adjusted reward values are used for DQN (deep Q network) and Dueling-DQN algorithms to enhance the convergence ability of the algorithm. By building a one-jammer to multi-radar interference penetration scenario, DQN and Dueling-DQN algorithms were verified, the experimental results verify the feasibility and difference of the two algorithms, and realize the resource allocation ability for interference resources in interference power, duration, interference pattern and interference radar selection, and meet the real-time and dynamic interference resource allocation requirement in the case of one-jammer to multi-radar.
尚熙, 杨革文, 戴少怀, 蒋伊琳. 基于强化学习的一对多雷达干扰资源分配策略研究[J]. 空天防御, 2022, 5(1): 94-101.
SHANG Xi, YANG Gewen, DAI Shaohuai, JIANG Yilin. Research on Resource Allocation Strategy of One-to-Many Radar Jamming Based on Reinforcement Learning. Air & Space Defense, 2022, 5(1): 94-101.
