基于双模特征的低信噪比雷达信号调制类型识别方法研究

打开文本图片集

中图分类号： TN971+ ;TN974 文献标志码：A

Abstract： The improved average complex envelope features are employed to classify common radar signal modulation types，while utilizing time-frequency features to compensate for the insuffciency of single features in fully characterizing signals.By fusing these two types offeatures，a dual-mode feature set is constructed.A corresponding classifier is designed for the dual-mode features to improve classification accuracy under low signal-to-noise ratio（SNR） conditions.

Key words：complex envelope feature; time-frequency feature; convolutional neural network

0 引言

有效的雷达信号侦察是雷达对抗的前提和基础，对于制电磁权的获取有至关重要的影响[1]。（剩余7669字）