摘要
针对传统的通信干扰策略生成算法需要预知通信环境先验知识的缺点,提出了基于高斯扰动探索的智能化干扰策略生成算法,实现未知通信环境下的干扰策略生成。该算法在没有先验知识的前提下,构建以频率、功率、占空比、干扰样式为基础的四维干扰策略搜索空间,结合强化学习的思想,通过高斯扰动探索的方式搜索整个策略组合空间,生成最优的干扰策略,提高了最优干扰策略的选择速度。仿真结果表明,针对不同的通信系统,基于高斯扰动探索的干扰策略智能生成方法能够更快的学习到最优干扰策略。
Aiming at the shortcomings of traditional communication interference strategy generation algorithms for predicting the prior knowledge of communication environmental, this paper proposes an intelligent interference strategy generation algorithm based on Gaussian perturbation exploration to realize the generation of interference strategies in unknown communication environments. Under the premise of no prior knowledge, this algorithm constructs a fourdimensional interference strategy search space based on frequency, power, duty cycle and interference pattern. Combined with the idea of reinforcement learning, the algorithm searches for the entire strategy combination space by Gaussian perturbation. The optimal interference strategy is generated and the selection speed of the optimal interference strategy is improved. Numerous results show that interference strategy intelligent learning based on Gaussian noise exploration can learn the optimal interference strategy faster for different communication systems.
作者
张君毅
张冠杰
杨鸿杰
Zhang Junyi;Zhang Guanjie;Yang Hongjie(The 54th Research Institute of China Electronics Technology Group Corporation, Shijiazhuang 050051, China)
出处
《电子测量技术》
2019年第16期148-153,共6页
Electronic Measurement Technology
基金
国家自然科学基金(61472305)项目资助
关键词
智能干扰
强化学习
通信对抗
intelligent interference
reinforcement learning
communication countermeasure