基于GRU神经网络的自适应跳频技术研究  

Research on Adaptive Frequency Hopping Technology Based on GRU Neural Network

作　　者：何雨桐 朱立东[1] 施文军 HE Yutong;ZHU Lidong;SHI Wenjun(National Key Laboratory of Science and Technology on Communications,UESTC,Chengdu 611731,China)

机构地区：[1]电子科技大学通信抗干扰技术国家级重点实验室,四川成都611731

出　　处：《无线电通信技术》2022年第6期1074-1079,共6页Radio Communications Technology

基　　金：国家自然科学基金(61871422)。

摘　　要：近年来,许多自适应干扰技术将重点转移到了跳频系统的同步频率集上。发射机时钟信息高位部分跳频作为控制信息确定的相关码在组帧模式下做跳频同步时,同步频率集的切换是以s为单位的,这就导致其易于捕获和遭受干扰。一旦同步频率集被捕获和干扰,通信系统就会面临崩溃。针对新型干扰对抗技术,首先利用神经网络对系统跳频图案进行训练,并模拟干扰方使用神经网络预测我方跳频图案的过程;然后对比LSTM网络和GRU网络应用于跳频图案预测的性能差异,针对神经网络的预测结果改进跳频图案设计,加入自适应同步频率集切换,观测改进后的跳频图案抗截获能力,并采用GRU神经网络对自适应跳频图案做预测。仿真结果表明,通过预测干扰方的行为来规避同步频率集被捕获的方案可以取得良好的抗干扰性能。In recent years,many adaptive jamming techniques have shifted their focus to synchronizing frequency sets for frequency hopping systems.When the related code determined by the transmitter clock information high-order part as the control information performs frequency hopping synchronization in the framing mode,the switching of the synchronization frequency set is in seconds,which makes it easy to acquire and suffer from interference.Once a synchronized set of frequencies is captured and jammed,the communication system faces collapse.Aiming at the new interference countermeasure technology,the neural network is used to train the frequency hopping pattern of the system.And the process of using the neural network to predict the frequency hopping pattern of the jammer is simulated.Then the performance difference between the LSTM network and the GRU network applied to the frequency hopping pattern prediction is compared.And the design of the frequency hopping pattern is improved according to the prediction result of the neural network,adaptive synchronization frequency set switching is added,the anti-interception ability of the improved frequency hopping pattern is observed,and the GRU neural network is used to predict the adaptive frequency hopping pattern.Simulation results show that the scheme of avoiding the capture of the synchronous frequency set by predicting the behavior of the interference can achieve good anti-jamming performance.

关 键 词：跳频通信 自适应跳频 抗干扰 神经网络 机器学习 

分 类 号：TN914.41[电子电信—通信与信息系统]