Research On Non-Stationary Processing Methods Of Low Probability Of Intercept Signals

LPI radar is an important developing aspect of modern radar technology. It is difficult to detect, identify, localize and track the LPI radar signal using traditional interception approaches. How to detect the LPI radar signal and estimate the signal parameters in low signal to noise ratio condition is very important in modern Electronic Warfare (EW). The research in this field is belongs to the frontier of EW technology. Many new LPI radar signal detection and parameter estimation algorithms have been deeply and widely studied in some degree in this dissertation. This work was partially supported by the project of National 973 Program. The main concerned signals are Linear Frequency Modulated (LFM) signal, Polynomial Phase Signal (PPS) and Binary Phase Shift Keying (BPSK) signal. The contributions of this dissertation are summarized as follows: * The technology of LPI radar has been introduced, and a classification of LPI radar signals is proposed. The current status of LPI signal processing has also been introduced.* For the problem of LFM signal detection, a method named as Multi-LFM signal time-frequency representation basing on Radon-ambiguity transform filtering method is proposed. For the different amplitude of Multi-LFM signal detection, a method based on recursive filtering Radon-ambiguity transform is proposed. Moreover, Multi-LFM signal time-frequency representation basing on recursive filtering Radon-ambiguity transform method is proposed.When the start point of the received signal record and that of the LFM signal are not coincident, the LFM signal parameter estimation problem is discussed. A method combining segmented correlation dechirping with local searching dechirping is proposed for low SNR LFM signal parameter estimation. After the analysis of Radon-Wigner transform, the local Radon-Wigner transform method is proposed. A method based on optimization methods and named as segmented correlation local recursive searching dechirping method is proposed for LFM signal parameter estimation. The validity of the estimation results got by dechirping method is* discussed on the condition that the beginning of the received signal and the beginning of the LFM signal are not coinciedent.* After the analysis of filter banks time-frequency detection method, a method based on the filter banks short time Radon-Wigner transform is proposed for PPS detection. Following the introduction of fractional Fourier transform (FrFT) and short time FrFT (STFrFT), a method based on CLEAN approach and STFrFT time-frequency representation is proposed for Multi-PPS detection in low SNR condition. * For the problem of PPS parameter estimation, the DPT method for PPS parameter estimation is introduced. According to the special structure of the 3rd order PPS, a method based on combining the segmented correlation with the LFM parameter estimation is proposed. A method based on instantaneous frequency fitting is proposed for the 3rd order PPS parameter estimation. * Following the spectral correlation theory and wavelet transform theory, a method named as wavelet spectral correlation method is proposed for BPSK signal detection and chip rate estimation. At last, the whole dissertation is concluded. And some further research frontiers of LPI radar signal detection and parameters estimation are pointed out.