α稳定分布噪声中低截获概率信号侦测方法研究

The low-probability-of-intercept (LPI) signals received by intercept systems in reality are often polluted by α-stable distributed noise. Moreover, there has been little literature to date on interception methods of LPI signals in α-stable noise under non-cooperative conditions. Focusing on the key techniques and difficult issues in LPI signal interception, the project mainly deals with the detection and parameter estimation of LPI signals in the presence of α-stable noise under non-cooperative conditions by using the fractional lower order operation and the clipper techniques to suppress α-stable noise and by employing such modern signal processing techniques as the cyclostationary theory and the time-frequency analysis, etc. The research can provide theoretical and technical support for performance improvement of non-cooperative receivers. The main contents are as follows: 1) The α stable distributed noise characteristics and its suppression methods. 2) For the problem of LPI signal blind detection, the following detection methods are studied respectively. They are the asymptotically optimal chi-square test for DSSS signals, the S transform and the channelization technique for FH signals, as well as the WVD-Hough transform and the LPP for LFM signals. 3) For the blind parameter estimation problem of LPI signals, the following estimation methods are studied respectively. They are the fractional lower order cyclic autocorrelation for DSSS signals, the robust STFT and the channelized spectra enhancement strategy for FH signals, as well as the maximum likelihood estimation and the DPT for LFM signals.

实际中截获接收系统侦收到的低截获概率（LPI）信号常受α稳定分布噪声的影响，而目前国内外非协作方式下、α稳定分布噪声中LPI信号侦测方法方面的研究几近空白。本项目针对LPI信号侦测中的关键技术及难点问题，拟采用分数低阶运算和限幅器等方法抑制α稳定分布噪声，结合循环平稳理论、时频分析等现代信号处理技术，开展非协作方式下、α稳定分布噪声中LPI信号的检测和参数估计方法研究，为提高非协作截获接收系统的性能提供成熟的理论和技术支撑。主要研究内容包括：1）α稳定分布噪声的特性及抑制方法。2）针对LPI信号的盲检测问题，分别研究直扩信号的渐近最优卡方检验，跳频信号的S变换、信道化技术及LFM信号的WVD-Hough变换和LPP等检测方法。3）针对LPI信号的盲参数估计问题，分别研究直扩信号的分数低阶循环自相关，跳频信号的鲁棒STFT、信道化谱增强及LFM信号的最大似然估计和DPT等参数估计方法。

实际中截获接收系统侦收到的低截获概率（LPI）信号常受α稳定分布噪声的影响，本项目重点研究α稳定分布噪声中LPI信号侦测中的关键技术及难点问题，在α稳定分布噪声抑制、直接序列扩频、跳频和LFM三种不同低截获概率信号的盲检测与盲参数估计方面，取得了一系列创新性研究成果。（1）针对α稳定分布噪声，分别研究了基于分数低阶统计量，基于限幅器、基于M估计理论和基于云模型理论的α稳定分布噪声抑制方法。2）针对LPI信号的盲检测问题，分别研究了直扩信号的渐近最优卡方检验，跳频信号的S变换、信道化技术、基于数据可信度加权与时频分析相结合、基于柯西分布最大似然估计，以及LFM信号的WVD-Hough变换、LPP、基于最优L-柯西加权等盲检测方法。3）针对LPI信号的盲参数估计问题，分别研究了直扩信号的分数低阶循环自相关，M估计与循环平稳理论相结合、基于柯西分布最大似然估计，跳频信号的鲁棒STFT、信道化谱增强、基于Meridian滤波、基于最优核时频分析，及LFM信号的最大似然估计、DPT、LVD、SAT等参数估计方法。本项目圆满完成了项目申请书中既定的研究内容，研究成果对于构建脉冲噪声中低截获概率信号侦测的理论体系具有重要的意义，并为提高非协作截获接收系统的性能提供成熟的理论和技术支撑。