Study On Signal Processing Of Interference And Sea Clutter For HF Surface Wave Radar

In recent years, HF Surface Wave Radar (HF-SWR) has been widely used in military and civilian fields. Its detection range is up to 300-400Km which can compensate blind spots of conventional microwave radar at low altitude and that of Sky Wave Over-the-Horizon Radar in near ranges. HF-SWR take the advantages of long distance and early warning surveillance, anti-stealth, anti-altitude penetration and ARM(Anti-Radiation Missiles). As a new kind of HF-SWR, coast-ship bistatic/multistatic Surface Wave Over-The-Horizon Radar can adopt Synthetic Impulse and Aperture Radar(SIAR) mechanism and can be installed receiver on maneuvering boats, so the system has good maneuverability during military operation over ocean. HF interference and sea clutter are two adverse factors during target detection of HF-SWR. The former includes the short-wave radio communications, lightning, meteor trail echoes, ionospheric interference and so on. The latter relate to spectrum broadening of sea clutter caused by factors such as bistatic structure and receiver movement. This dissertation is focused on suppression processing of HF interference and sea clutter in HF-SWR and SIAR. The main contents and innovation are summarized as follows:1. Correlation and energy distribution characteristics of Radio Frequency interference were analyzed. Meanwhile, characteristics of transient interference including thunder lightning and meteor trails and ionospheric interference are briefly summarized. Simulation results show that the correlation and energy distribution properties of Radio Frequency interference have distinct characteristics in time and range domain. When using correlation characteristics, interference suppression can be obtained by a number of training samples from interested ranges, and computing load is smaller than suppression processing from time domain. When using characteristic of interference energy distribution, suppression can be obtained by excising process in fast time domain but which has side effectives such as target energy loss and windowing effects to sea clutter and consequently high side-lobe in frequency domain. Correlation coefficients of lightning interference between ranges are small and appear time of interference changes between frequency-modulation periods in the low-pass filtering output. Lightning energy are distributed in all Range-Doppler cell and can be detected and removed in time domain by its difference from target and clutter echoes. 2. RF interference suppression methods are analyzed and improved based on single channel echo. During excising suppression, excising method in frequency domain is proposed for coding-modulation system which is similar to time domain excising in FMICW system. During eigen-subspace suppression, interference subspace is well estimation by adopting negative range training samples and range weights for correlation matrix calculation. During cancellation suppression, estimation algorithm for characteristic frequency of RF interference is proposed according to signal parameters and power distribution location. Then Construction-Cancellation suppression method is proposed based on Wiener filter where the reference signal coherent to interference is constructed by the carrier frequency estimated. Finally the comprehensive suppression procedure based on Construction-Cancellation suppression is proposed for composite interference comprise of single-frequency and narrowband components. Simulation results show that comprehensive suppression procedure can get effective results than other methods and with higher target gain. Suppression algorithms above are suitable for conventional ground wave radar and SIAR.3. Transient interference suppression algorithms were introduced and improved based on the difference between lightning and Meteor trail echoes. The algorithm combining wavelet analysis and matrix decomposition for lightning suppression is proposed. First, wavelet analysis is applied for interference detection and then the sample matrix is constructed adaptively by training data near interference component. Then SVD is applied to the matrix by process of which interference components are removed and target energy is effectively accumulated using AR extrapolation. This method can reduce computing load and increase recovery accuracy of target component when compared to the process of uniform truncation for training matrix construction and decomposition which is regardless of interference information. In Meteor trail suppression, interference covariance matrix is constructed adaptively in sub-CPI and an improved adaptive beamforming algorithm introducing clutter constraints is presented which do not destroy time coherence of the first-order Sea clutter after interference suppression and benefits target detection in frequency domain subsequently.4. With Synthetic Impulse and Aperture Radar signal processing the ionosphere interference is analyzed. Using space-time snapshots constructed jointly through array elements and fast time samples, interference adaptive suppression algorithm is proposed by the virtue of random constraint to sea clutter in space-time domain. Meanwhile, main component of sea clutter is effectively traced by constructing Hankel matrix and instantaneous frequency is estimated by SVD decomposition. Simulation and experimental results validate the suppression algorithm for ionospheric interference.5. Sea clutter spreading spectrum is analyzed under the condition of receiver moving in Synthetic Impulse and Aperture Radar, and joint space-time adaptive processing for non-stationary clutter between range cells is discussed. Based on two dimensional distribution characteristics of sea clutter spectrum, space-time adaptive processing (STAP) algorithm is applied to suppress spreading sea clutter: space-time interpolation is applied to non-uniform training samples according to relationship between clutter spectrum and system geometry parameters , thenafter, Direct Data Domain method is adopt to get enough training samples for Sea clutter STAP.6. Characteristics of nonstationary sea clutter sequence are analyzed and high resolution spectral estimation methods are introduced. To improve the accuracy of frequency-domain analysis for non-stationary radar echoes, joint estimation of multi-parameter method in time domain is proposed for the number of stable AR processes and corresponding filter coefficients. During optimal estimation of signal parameters for nonstationary sequence, wavelet analysis for singular signal detection has reduced the searching ranges which contain borderlines of AR sections. Thenafter, maneuvering target detection method according to time-domain characteristics of sea clutter is proposed after target motion parameter estimation process by Wigner-Ville time-frequency analysis (TFA) and Radon transform. Simulation shows that time-domain parameter estimation and maneuvering target detection methods for high-frequency ground wave radar echo can effectively reduce the first-order spectrum width and suppress second-order spectrum and then improve the detection capability to maneuvering target.