摘要
针对现有基于雷达脉冲重复间隔(PRI)调制模式识别性能受特征选择和门限设置影响较大的问题,提出一种基于长短时记忆网络的识别方法。该方法以脉冲时间序列为输入,不需要人工选择特征和设置门限,通过构建信号参数特征自动提取及PRI调制模式一体化识别的循环神经网络结构,训练获得最优的网络参数完成雷达信号调制模式识别。仿真结果表明:该方法识别精度高,并对脉冲丢失、测量噪声影响及脉冲数量少等情况具有良好的适应能力,适合于工程应用。
To solve the problem that the performance of pulse repetition interval(PRI)modulation pattern recognition method is affected by the feature selection and threshold setting,a PRI recognition method based on the long short-term memory is proposed.This method takes pulse time series as input,and it is unnecessary to select features and set thresholds manually.The recurrent neural network of PRI modulation pattern identification which can extract the signal parameter characteristics automatically is designed.The optimal network parameters are obtained by training to complete the recognition.The simulation results show that the method has high recognition accuracy and good adaptability to pulse loss,measurement noise,and small number of pulses,so it is suitable for engineering application.
作者
孟磊
曲卫
马爽
刘元华
MENG Lei;QU Wei;MA Shuang;LIU Yuanhua(Graduate School,Space Engineering University,Beijing 101416,China;Department of Electronic and Optical Engineering,Space Engineering University,Beijing 101416,China;Beijing Institute of Remote Sensing,Beijing 100192,China)
出处
《现代雷达》
CSCD
北大核心
2021年第1期50-57,共8页
Modern Radar
关键词
循环神经网络
长短时记忆网络
脉冲重复间隔调制
模式识别
recurrent neural network
long short-term memory network
pulse repetition interval modulation
pattern recognition
