首页|期刊导航|电讯技术|基于改进PPO的雷达抗干扰自适应跳频方法

基于改进PPO的雷达抗干扰自适应跳频方法OA

A Radar Anti-jamming Adaptive Frequency Hopping Method Based on Improved Proximal Policy Optimization(PPO)Algorithm

中文摘要英文摘要

针对现代电子战中雷达系统面临的日益复杂干扰问题,提出了一种基于深度强化学习的频率捷变雷达智能抗干扰决策算法.通过将深度强化学习方法与频率捷变雷达相结合,提出了一种基于近端策略优化的雷达自适应跳频抗干扰方法.首先,将雷达和干扰机之间的相互作用过程建模为部分可观测马尔可夫决策过程,基于检测概率设计了奖励函数.然后,采用近端策略优化(Proximal Policy Optimization,PPO)算法,通过其裁剪操作来解决传统策略梯度算法中策略更新过大的问题.最后,引入多层历史观测信息,并使用门控循环单元来提取序列数据中的长期依赖关系,以解决部分马尔可夫决策过程中信息不足的问题.仿真结果表明,该方法相较于传统近端策略优化算法能够将检测概率提高8%.

In modern electronic warfare,radar systems face increasingly complex jamming challenges.An intelligent anti-jamming decision-making algorithm for frequency-agile radar based on deep reinforcement learning(DRL)is proposed.By integrating DRL with frequency-agile radar technology,an adaptive frequency hopping anti-jamming method is proposed using the proximal policy optimization(PPO)algorithm.First,the interaction between the radar and the jammer is modeled as a partially observable Markov decision process(POMDP),and the reward function is designed based on detection probability.Then,the PPO algorithm is employed,utilizing its clipping mechanism to address the issue of excessive policy updates in traditional policy gradient methods.Finally,multi-layer historical observations are incorporated,and a gated recurrent unit(GRU)is applied to capture long-term dependencies in sequential data,mitigating the information insufficiency in POMDP.Simulation results show that,compared with traditional PPO algorithms,the proposed method improves the detection probability by 8%.

李家强;李淑君;喻庞泽;陈金立;姚昌华

南京信息工程大学 电子与信息工程学院,南京 210044南京信息工程大学 电子与信息工程学院,南京 210044南京信息工程大学 电子与信息工程学院,南京 210044南京信息工程大学 电子与信息工程学院,南京 210044南京信息工程大学 电子与信息工程学院,南京 210044

信息技术与安全科学

频率捷变雷达雷达抗干扰自适应跳频深度强化学习近端策略优化(PPO)

frequency-agile radarradar anti-jammingadaptive frequency hoppingdeep reinforcement learningproximal policy optimization(PPO)

《电讯技术》 2026 (9)

1456-1463,8

10.20079/j.issn.1001-893x.250415006

评论