航空学报 > 2013, Vol. 34 Issue (11): 2580-2589   doi: 10.7527/S1000-6893.2013.0352

基于周期WHT循环滤波的交叠类似LFMCW信号检测与分离

张立民1, 钟兆根1, 王泽众2, 王建雄1   

  1. 1. 海军航空工程学院 电子信息工程系, 山东 烟台 264001;
    2. 海军装备研究院, 北京 102249
  • 收稿日期:2012-11-27 修回日期:2013-07-11 出版日期:2013-11-25 发布日期:2013-08-08
  • 通讯作者: 张立民,Tel.:0535-6635402 E-mail:iamzlm@163.com E-mail:iamzlm@163.com
  • 作者简介:张立民 男,博士,教授,博士生导师。主要研究方向:卫星信号处理、武器系统仿真等。 Tel:0535-6635402 E-mail:iamzlm@163.com;钟兆根 男,博士,讲师。主要研究方向:通信侦察和空间信息对抗。 Tel:0535-6635680 E-mail:zhongzhaogen@163.com
  • 基金资助:

    国家自然科学基金(61102167,60972159);航空科学基金(20085184003)

Detection and Separation of Overlapped Quasi-LFMCW Signals Based on Periodic WHT Recurrent Filter

ZHANG Limin1, ZHONG Zhaogen1, WANG Zezhong2, WANG Jianxiong1   

  1. 1. Department of Electronics and Information Engineering, Naval Aeronautical and Astronautical University, Yantai 264001, China;
    2. Naval Academy of Armament, Beijing 102249, China
  • Received:2012-11-27 Revised:2013-07-11 Online:2013-11-25 Published:2013-08-08
  • Supported by:

    National Natural Science Foundation of China (61102167, 60972159);Aeronautical Science Foundation of China (20085184003)

摘要:

针对非协作方式下交叠类似线性调频连续波(LFMCW)信号的检测与分离问题,以LFMCW信号为例推导了一种基于周期WHT(Wigner-Hough Transform)循环滤波的检测与分离算法,并根据该类信号与LFMCW信号的联系与区别,将其推广应用于类似LFMCW信号的检测与分离。该算法能在较低信噪比条件下实现交叠类似LFMCW信号的分离,且分离后的信号具有抑制噪声的作用;同时针对强信号分离时带来的强信号残留和弱信号损失的问题,设计了一种基于单元平均的窄带频域陷波滤波器,在充分考虑滤除强信号能量的同时,实现了弱信号能量的更好保留。仿真结果表明:当虚警概率为0.01,输入信噪比大于-16 dB时,算法的检测概率大于0.9,相比同条件下的WHT和分数阶Fourier变换(FrFT)分别提高了7 dB和6 dB;对交叠类似LFMCW信号,当最弱信号信噪比大于-7 dB时,就可使分离的相关系数达到0.9以上。

关键词: 周期WHT, 线性调频连续波, 循环滤波, 信号分离, 交叠信号, 特征提取

Abstract:

For weak signal detection and separation of overlapped quasi-LFMCW signals with non-cooperative mode, a new algorithm based on periodic WHT recurrent filter is proposed by using LFMCW signal as an example. Then extend the use of the proposed algorithm to quasi-LFMCW signals with the relationship and difference between them. The proposed algorithm can realize the separation of overlapped quasi-LFMCW signals at low signal-to-noise ratio (SNR), also the separated signals can effectively suppress noise. Aimed at solving the core problem of the high energy signal's strong remainder and the weak energy signal's energy loss after the high energy signal's separation, a narrow frequency excision filter based on cell average is designed, which could not only sufficiently filter the high energy signal, but also keep the weak signal as completely as possible. Simulation results indicate that when false alarm probability is set to 0.01 and SNR is greater than -16 dB, the probability of detection reaches above 0.9. With respect to WHT and FrFT under the same condition, the detection performance is improved by 7 dB and 6 dB respectively. In addition, for overlapped quasi-LFMCW signals, when the weakest signal's SNR is greater than -7 dB, the correlation coefficients between separated and original signal component both exceed 0. 9.

Key words: periodic WHT, LFMCW, recurrent filter, signal separation, overlapped signal, feature extraction

中图分类号: