Study On Signal Processing Technology Of Sparse Spectrum High Frequency Ground Wave Radar

High-frequency ground-wave radar(HFGWR)has an over-the-horizon range capability,and the detection range can cover an exclusive economic zone of 200 nautical miles.It is widely used in coastal countries for remote monitoring of surface vessels and ocean conditions.However,there are a large number of radio users such as communications and broadcasting in the high-frequency band,the frequency spectrum is very congested,which limits the bandwidth of high-frequency ground-wave radar,resulting in a low range resolution.Usage of sparse spectral signals is a radical solution to the low range resolution problem of highfrequency ground-wave radar systems.Unlike conventional radar signals,the spectrum of a sparse spectral signal is discontinuous,occupying frequency bands which are quiet in the current spectral environment.It appears to be unevenly,sparsely distributed due to the restrained frequency resources for radar operating.Therefore,this thesis focuses on the development of sparse spectrum high-frequency ground-wave radars.Firstly,they are divided into two categories according to the system architecture: one with spectrum sparsity implemented in time domain,that is,whose transmitted signals occupy different frequency bands in different time periods;the other with spectrum sparse implemented in spatial domain,whose transmit channels simultaneously transmit signals with different carrier frequencies and non-overlapping spectrum.The first category only requires one transmit channel.At present,high-frequency ground-wave radar systems using single-transmit-multiple-receive or all-digital phased arrays can be configured as such.The second category is a kind of multi-carrier multiple input multiple output(MIMO)radar,and is also a generalized synthesize impulse and aperture radar(SIAR)with its carrier frequency generalized from uniform,continuous to non-uniform,sparse.In this thesis,we study these two categories of sparse spectrum high-frequency ground-wave radar respectively.The specific work is summarized as follows:In the first part,the commonly used signal waveforms of high-frequency ground-wave radar and its processing methods were summarized.The signal waveform adopted by HF radar is different from general microwave radars.Therefore,the first part firstly analyzes and summarizes the commonly used waveforms of the high-frequency ground-wave radar,including the frequency modulated continuous wave(FMCW),frequency modulated interrupted continuous wave(FMICW)and complementary code pulse signals and their corresponding parameter selecting and range-Doppler processing methods.Firstly,the de-chirping plus two-dimensional fast Fourier transform(FFT)FMCW range-Doppler processing method is introduced.This de-chirping based range processing method is used several times in this thesis.Subsequently,range transform method by directly fast Fourier transforming of the de-chirped signal and by using time gate for FMICW are summarized and compared.Comparing with first two waveforms,phase-coded pulse signals are used less frequently in high-frequency ground-wave radars.However,the complementary code pulse signal can obtain an ideal pulse compression output by using a pair of codes with complementary autocorrelation functions,which is very attractive for high-frequency ground-wave radar.This thesis discusses the application of complementary code signals in HFGWR,and proposes the CLEAN idea based complementary code segmentation pulse compression method for solving the blind area problem which utilizes the complementary feature of the sub pulses of the complementary code sequence generated by the interpolation method.This algorithm guarantees that there are ideal zero sidelobes in the blind area,and no false peak problem existed in full range.In the second part,distance processing of sparse spectrum FMCW(SS-FMCW)is studied.The carrier frequency of SS-FMCW signal hops in fast time domain.It belongs to the category of signal with spectral sparse implemented in time domain.SS-FMCW uses dechirping based processing method similar to conventional FMCW.However,due to the sparsity of signal spectrum,its de-chirping output is no longer a sampling sequence of a single frequency continuous wave,and cannot simply use FFT for range conversion.Here the time rearrangement operation is proposed to transform the SS-FMCW range processing problem to a spectral estimation problem.Due to the sparsity of signal spectrum,sampling sequence of the spectral estimation problem is non-uniformly distributed(discontinuous distributed).In order to solve the spectral estimation problem,this thesis proposes an SS-FMCW range spectrum estimation algorithm using iterative adaptive approach(IAA)and sparse learning via iterative minimization(SLIM).The performance of both algorithms is compared and their fast implementation is given.Computer simulations show that both IAA and SLIM are suitable for SS-FMCW range processing.In the third part,we study the range-velocity processing of SS-FMCW.SS-FMCW brings a high range resolution to the high-frequency ground-wave radar,but at the same time,due to the increase of range resolution,moving of the target during a coherent integration cycle becomes unignorable;on the other hand,due to the hopping of carrier frequency in the fast time dimension,target speed maps to different Doppler frequencies in different fast time segments.Therefore,on the basis of second part,range-velocity joint processing scheme and cascading processing scheme are proposed respectively.Both of the two solutions can solve the problem of range walking during a long time of coherent integration and the fast time Doppler hopping.Advanced spectral estimation algorithms such as IAA and SLIM can be used in both schemes.This thesis analyzes and compares the performance of joint processing and cascading processing scheme in range and velocity domain.So does their computing complexity.Cascading processing scheme makes full use of the characteristics of the SS-FMCW high-frequency ground-wave radar which has non-uniform fast time samples and uniform slow time samples.It is quite suitable to be used in real systems.Range and range-velocity processing methods for SS-FMCW signals can be extended to sparse spectrum frequency modulated interrupted continuous wave(FMICW)signals,which is analyzed and illustrated in this thesis.In the fourth part,the range-azimuth processing of sparse carrier frequency MIMO HFGWR is studied.The object discussed in this part is different from the second and third parts.It is a multi-channel radar system which implements spectral sparsity in the spatial domain.This part deduces the range-azimuth processing model for sparse spectrum MIMO radar system which uses de-chirping range processing or sliding window matched filtering range processing respectively.Their range-azimuth ambiguity functions are analyzed and compared.Subsequently,for range-azimuth coupling problem,the optimization model with sparse and random distribution of carrier frequency is established and its genetic algorithm solution is presented.Random rearrangement carrier frequency distribution optimizing method which suitable for engineering implementation is proposed.Finally,for solving the problem of high range sidelobe in the range-azimuth spectrum due to the sparse distribution of carrier frequencies,the matched matrices of the radar echo in both the de-chirping processing and sliding window matched filtering method are derived.Then IAA and SLIM algorithm introduced in the second part can be directly applied.The simulation experiment takes IAA algorithm as an example to show that when it is utilized in range-azimuth processing of sparse carrier frequency MIMO radar,it can solve the problem of high range sidelobe,and has narrow mainlobes in both range and azimuth domains.