A method for generating clutter signals based on statistical characteristics

Abstract

The present invention is aimed at the field of radar simulator echo signal generation, and particularly relates to an efficient method for generating clutter signals based on statistical characteristics. The present invention adopts a statistical method combined with a grid division method to generate clutter echo signals in a certain area, and uses an amplitude probability density function, a power spectrum density function, and a spatial correlation function to characterize the clutter; considering that the power spectrum position and spectrum width and spatial correlation between each frame of clutter echoes have small-range fluctuations, the present invention introduces the Doppler spectrum defocusing characteristics of moving targets, and based on the generation of the first frame of clutter data, uses a defined defocusing factor to cyclically generate multiple frames of clutter signals, ensuring that the power spectrum and spatial correlation of the clutter change dynamically over time, greatly reducing the system processing time required for the generation of multiple frames of clutter data.

Description

Technical Field

[0001] The invention belongs to the field of echo generation, and particularly relates to a method for efficiently generating clutter signals. Background Art

[0002] Clutter signals need to be generated in radar simulators and radar self-tests. The radar echo signals in existing complex environments contain not only target signals, but also a large number of clutter signals reflected from the surrounding environment and other active interference signals. Since clutter signals vary with the surface environment and radar performance parameters, they cannot be accurately modeled according to the clutter scattering mechanism. At the same time, clutter data exhibits different statistical characteristics under different radar irradiation times. If each frame of signal echo is modeled separately based on the time-varying nature of the clutter statistical characteristics, the amount of clutter data generated will become very large as the simulated radar working time increases, which will reduce the system processing efficiency.

[0003] The existing method for clutter data is to generate an echo signal, then generate a modulation matrix according to the required statistical characteristics, and multiply the echo signal and the modulation matrix row by row to complete the generation of a frame of clutter. Subsequently, multiple frames of data are generated in this way in a cycle, and the generation technology is inefficient. Summary of the invention

[0004] In view of the inefficiency of existing clutter data generation technology, the present invention provides a clutter signal generation method based on a clutter statistical model and the Doppler spectrum defocusing characteristics of a moving target. In order to more comprehensively describe the clutter characteristics, the three statistical characteristic functions of amplitude probability density function, power spectrum density function and spatial correlation function are used to simulate clutter data. Based on the first frame of clutter data generated, the Doppler spectrum defocusing characteristics are used to transform each repeated period clutter signal in the frame into the frequency domain and multiply it by different frequency factors to achieve changes in the power spectrum position and spectrum width. Similarly, a set of slow time dimension data in each distance unit of a frame of data is multiplied by different frequency factors to achieve changes in the spatial correlation of the clutter signal.

[0005] The technical solution of the present invention is: a method for efficiently generating clutter signals based on statistical characteristics.

[0006] Step 1: Determine the range of the clutter area and divide it into clutter units.

[0007] Assuming the radar clutter area distance range [r 1 ,r 2 ], azimuth range According to the equal distance-equal azimuth division method, the division of clutter units is based on the azimuth resolution. and the distance resolution ΔR, as Fig.11 shown.

[0008] The distance unit size is equal to the distance resolution, which is related to the transmission signal bandwidth B, that is:

[0009]

[0010] The azimuth unit size is the azimuth resolution of the radar system parameters. When the radar azimuth angle range is Then the number of azimuth divisions is

[0011] Step 2: Calculate the amplitude multiplication coefficient K of each unit according to the radar equation m .

[0012]

[0013] Where λ is the radar operating wavelength; p t is the radar transmit peak power; s m The area of ​​the mth clutter unit; G(θ) is the radar antenna gain; r(m) is the distance between the mth clutter unit and the radar; L is the radar comprehensive loss; σ m is the scattering area of ​​the mth clutter unit.

[0014] σ m =σ 0 ·s m

[0015] In the formula, is the area of ​​the mth clutter unit; σ 0 The average backscatter coefficient of the clutter area is obtained by selecting the Ulaby ground scattering coefficient model according to the detection environment and radar system parameters, consulting the model parameter values, and calculating the average backscatter coefficient σ of the clutter area. 0 The Ulaby model can be expressed as:

[0016] σ 0 (dB) = P 1 +P 2 exp(-P 3 θ)+P 4 cos(P 5 θ+P 6 )

[0017] Where θ is the incident angle of the radar to the ground (0°~80°); parameter P 1 ~P 6 It is related to the type of ground feature, frequency band and polarization. Under the grassland Ulaby model, when the radar signal uses the L band, the polarization mode is vertical polarization, and the signal incident angle is in the range of 0° to 80°, then P 1 =-29.235; P 2 =-37.55; P3 =2.332; P 4 =-2.615; P 5 =5.0;P 6 =-1.616.

[0018] Step 3: Generate clutter unit echo signals and complete the superposition of echoes of various clutter units.

[0019] The present invention uses LFM signals to illuminate the ground, so the duration of any clutter unit echo is the duration of a transmitted signal. Since the distance between adjacent clutter units in a certain direction differs by a distance resolution, it is necessary to delay the clutter unit echoes by one distance unit before superimposing them. For the same distance unit, the echoes in different directions can be directly added together, such as Fig.12 As shown. Under a pulse irradiation, the echo of this area is expressed as:

[0020]

[0021] Where N e N is the number of azimuth resolution units for regional division; r K is the number of distance units divided under a certain orientation unit; i,j Represents the multiplicative coefficient of the clutter unit at the i-th direction and the j-th distance; τ j represents the echo delay of the jth clutter range unit; T represents the signal duration; f 0 is the signal center frequency; μ is the frequency modulation slope, and t represents the time variable.

[0022] Step 4: Generate a clutter data modulation coefficient matrix.

[0023] When generating the modulation coefficient matrix, the dimension of the matrix is ​​equal to the amount of data in one frame of echo signal, that is, the number of rows is equal to the number of repeated pulses in one frame of signal, and the number of columns is equal to the number of sampling points in the fast time dimension of the pulse. It is assumed that the amplitude of the multi-pulse clutter echo must satisfy the Rayleigh distribution, the power spectrum is the Gaussian spectrum, and the time correlation contains two components, fast fluctuation and slow fluctuation. The modulation matrix required to be generated needs to meet the above three characteristics.

[0024] First, generate a set of Gaussian white noise complex sequence g with the dimension of the number of repeated pulses N N×1 , N is determined according to the actual situation; the Gaussian power spectrum density function is:

[0025]

[0026] In the formula, f τ is the spectrum center frequency; σ f is the 3dB bandwidth of the spectrum, f α is the frequency independent variable.

[0027] Transform the power spectrum of the white noise complex sequence into a Gaussian power spectrum density:

[0028] G N×1 =ifft(fft(g N×1 )·P(f α ))

[0029] According to the above process in step 4, the cycle is repeated M times, and each modulation vector G is generated. N×1 As each column of the modulation matrix data, M is the number of time sampling points of the clutter signal, which is determined according to the actual situation to obtain the modulation matrix A N×M :

[0030]

[0031] Modulation Matrix A N×M Perform spatial correlation processing to obtain a modulation coefficient matrix that satisfies three statistical characteristics:

[0032]

[0033] In the formula, fft() row Indicates that the matrix is ​​Fourier transformed row by row; ifft() row represents the inverse Fourier transform of the matrix row by row; Γ 1×M (f β ) is the power spectral density function corresponding to the spatial correlation function, f β is the frequency independent variable of the power spectrum function, and the spatial correlation function is defined as:

[0034]

[0035] Γ 1×M (f β ) = fft(Γ(d))

[0036] In the formula, 0≤a≤1 represents the proportion of the fast fluctuation component in the total energy in the spatial correlation function, b and c are the decorrelation distances of the fast and slow fluctuation components, respectively, and d represents the correlation distance. According to the radar incident angle and working band, the values ​​of a, b, and c can be determined by consulting the data. In a grassland environment, when the radar signal uses the L band and the signal incident angle is 60°, a=0.545cm; b=5.29cm; c=26.71cm.

[0037] Step 5: Generate N clutter echo signals within one coherent processing time.

[0038] The pulse echo signal obtained in step 3 is multiplied row by row by the modulation matrix in step 4, so that N clutter echo signals are obtained, and the three statistical characteristics proposed in step 4 are satisfied between the N echoes:

[0039]

[0040] Step 6: cyclically generate a frame of clutter echo signal within each coherence time.

[0041] During operation, the radar will continuously receive multiple frames of echo signals in the clutter area, and N clutter echoes received within a coherent time constitute a frame of echo signals. In actual environments, the center position and spectrum width of the power spectrum between each frame of clutter echoes will change in a small range, and the power spectrum function corresponding to the spatial correlation will also show this characteristic. Based on the one-frame echo data generated in step 5, the Doppler spectrum defocusing characteristics of the moving target are introduced to achieve different time delays for each echo in a frame of signal, which can quickly achieve the corresponding changes in the power spectrum.

[0042] According to the power spectrum corresponding to the clutter time correlation, the first defocus factor is defined as:

[0043]

[0044] In the formula, f χ is the frequency point corresponding to the fast time-dimensional Fourier transform of a clutter; τ p The delay for the pth clutter in a frame is set according to the spectrum width; p = 1, 2…, N, where N is the number of clutter echoes within the coherence time.

[0045] The frequency domain data of each clutter echo in a frame of signal calculated in step 5 are respectively compared with a set of defocus factors w p Point multiplication is performed to obtain the clutter data of the power spectrum expansion corresponding to the time correlation:

[0046]

[0047] In the formula, fft() row Indicates Fourier operation on the row dimension of the data matrix; ifft() row Represents the inverse Fourier transform along the row dimension.

[0048] According to the power spectrum corresponding to the spatial correlation of clutter, the second defocus factor is defined as:

[0049]

[0050] In the formula, f ε is the frequency point corresponding to the slow-time Fourier transform at the same distance unit in a frame of clutter signal; τ q The delay for the slow time dimension data under the qth distance unit is set according to the spectrum width requirements; q = 1, 2…, M, where M is the number of fast time dimension distance units of the clutter signal.

[0051] The above process obtains a frame The signal matrix combines the slow time data under each distance unit with a set of defocus factors z q Point product to obtain the clutter data of the power spectrum expansion corresponding to the spatial correlation:

[0052]

[0053] In the formula, fft() col Indicates that Fourier operation is performed on the column dimension of the data matrix; []′ indicates transposing the matrix; ifft() col Indicates the inverse Fourier transform in column dimension; U N×M That is, a new frame of clutter data is generated based on the first frame of clutter data by utilizing the defocusing characteristics.

[0054] The beneficial effect of the present invention is: a clutter echo signal of a certain area is generated by combining a grid division method in a statistical manner, that is, the amplitude probability density function, the power spectrum density function and the spatial correlation function are used for characterization. According to the definition of the clutter signal amplitude probability density function and the power spectrum density function, the number of slow time dimension sampling points on a certain distance unit needs to be predicted in advance when generating a modulation matrix. At the same time, in the process of grid division of a certain area, according to the clutter reflection cross-sectional area, distance and azimuth of the area, each unit divided will have its own reflection intensity. Therefore, the process of generating multiple groups of clutter signals in each frame is the result of independent generation under each clutter unit and then superposition. If the simulated radar works for a long time and generates multiple frames of echoes, due to the small range fluctuations in the power spectrum position and spectrum width and spatial correlation between each frame of clutter echoes, if the clutter data is generated in a repeated cycle, the signal system processing time will be significantly increased. The present invention introduces the Doppler spectrum defocusing characteristics of moving targets on the basis of generating the first frame of clutter data and on the premise that the clutter signal satisfies the corresponding amplitude probability density function, power spectrum density function and spatial correlation function, so as to ensure that the power spectrum and spatial correlation of the clutter change dynamically with time, thereby greatly reducing the system processing time required for generating multiple frames of clutter data. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 The figure is a flow chart of clutter signal generation according to the present invention.

[0056] Figure 2 This is the probability density map of the clutter envelope in the first distance unit in the first frame.

[0057] Figure 3 This is the clutter power spectrum at the first distance unit in the first frame.

[0058] Figure 4 This is the spatial correlation of the first clutter signal and its corresponding power spectrum in the first frame.

[0059] Figure 5 This is the probability density map of the clutter envelope in the first distance unit in the second frame.

[0060] Figure 6 This is the clutter power spectrum at the first distance unit in the second frame.

[0061] Figure 7 This is the spatial correlation of the first clutter signal and its corresponding power spectrum in the second frame.

[0062] Figure 8 It is the system time taken to run the first frame of clutter data in MATLAB environment.

[0063] Fig. 9 It is the system time taken to run the second frame of clutter data based on the existing first frame of data in the MATLAB environment.

[0064] Fig.10 It is the system time taken to simulate 50 circles of radar movement and generate 50 frames of clutter data based on the first frame of data in MATLAB environment.

[0065] Fig.11 Schematic diagram of ground clutter unit division.

[0066] Fig.12 The process of signal superposition of multiple clutter units in the range direction. DETAILED DESCRIPTION

[0067] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments.

[0068] The present invention provides a method for efficiently generating clutter echoes, which ensures that clutter signals meet statistical characteristics and realizes dynamic changes in statistical characteristics between frame echoes, thereby avoiding system calculation pressure caused by repeated generation of clutter data.

[0069] like Figure 1 As shown, the steps of generating clutter signals of the present invention include: determining clutter units based on the "grid division method"; calculating the regional average backscattering coefficient, and calculating the multiplicative coefficient of each clutter unit according to the radar equation; completing the superposition of echoes of each clutter unit; generating a clutter data modulation coefficient matrix; generating a frame of clutter echo signals; and cyclically generating a frame of clutter echo signals within each coherence time.

[0070] Step 1: Determine the clutter area and divide the clutter units:

[0071] According to the radar transmission PRF and the clutter area beam dwell time T beam , obtain the total number of pulses N in the clutter area; according to the signal sampling frequency and echo duration, obtain the number of signal sampling points M.

[0072] In this embodiment, the sampling amplitude obeys the Rayleigh distribution, the power spectrum corresponding to the time correlation obeys the Gaussian spectrum, and the spatial correlation contains fast fluctuation and slow fluctuation components. The azimuth width of the clutter area is set to The azimuth resolution is The distance range is set to [R min =100m R max =150km], the distance resolution is ΔR=75m, and the clutter environment is selected as grassland environment.

[0073] Number of distance units:

[0074]

[0075] The number of azimuth units is:

[0076]

[0077] Step 2: Calculate the amplitude multiplication coefficient K of each clutter unit according to the radar equation m .

[0078] Check the Ulaby model parameters to determine the P 1 ~P 6 Model coefficients, calculate the average backscatter coefficient σ 0 (dB):

[0079] σ 0 (dB) = P 1 +P 2 exp(-P 3 θ)+P 4 cos(P 5 θ+P 6 )

[0080] Where θ is the incident angle of the radar onto the ground; parameter P 1 ~P 6 Depends on the type of terrain, frequency band and polarization.

[0081] Calculate the area of ​​each clutter cell:

[0082]

[0083] In the formula, r m is the distance between the mth scattering unit and the radar.

[0084] Find the reflection cross-sectional area of ​​each clutter unit:

[0085] σ m =σ 0 ·s m

[0086] Substitute the azimuth and elevation angle, distance, and reflection cross-sectional area data of each clutter unit relative to the radar into the radar equation to obtain the echo multiplication coefficient of each unit:

[0087]

[0088] Where λ is the radar operating wavelength; p t is the radar transmit peak power; s m The area of ​​the mth clutter unit; is the radar antenna gain, θ are the azimuth and elevation angles of the mth clutter unit relative to the radar; r m is the distance between the clutter unit and the radar; L is the radar comprehensive loss; σ m is the scattering area of ​​the mth clutter unit.

[0089] Step 3: Complete the echo superposition of each clutter unit.

[0090] When adding up the echoes of the distance clutter unit, it is necessary to delay one sampling point respectively before adding them up. For the same distance unit, the echoes in different directions can be directly added together; multiply the clutter unit multiplication coefficient obtained in step 2 by each echo signal, and then perform the superposition processing.

[0091]

[0092] Where N e N is the number of azimuth resolution units for regional division; r K is the number of distance units divided under a certain orientation unit; i,j The amplitude multiplicative coefficient of the clutter unit at the i-th direction and the j-th distance; τ j represents the echo delay of the jth clutter range unit; T represents the signal duration; f 0 is the signal center frequency; μ is the frequency modulation slope.

[0093] Step 4: Generate a clutter data modulation coefficient matrix.

[0094] According to the requirement that the generated amplitude obeys the Rayleigh distribution, the power spectrum corresponding to the time correlation obeys the Gaussian spectrum, and the spatial correlation contains clutter data of fast fluctuation and slow fluctuation components, the modulation matrix must satisfy the above three statistical characteristics.

[0095] First, generate a set of white noise complex sequence with a data volume of N, and transform the power spectrum of the white noise complex sequence into a Gaussian power spectrum:

[0096] G N×1 =ifft(fft(g N×1 )·P(f α ))

[0097] In the formula, g N×1 is N white noise complex data that obeys Gaussian distribution; P(f α )Gaussian power spectral density function.

[0098] The modulation vector G is generated cyclically N×1 As each column data of the modulation matrix, the modulation matrix A is obtained N×M :

[0099]

[0100] Modulation Matrix A N×M Perform spatial correlation processing:

[0101]

[0102] In the formula, fft() row Indicates that the matrix is ​​Fourier transformed row by row; ifft() row represents the inverse Fourier transform of the matrix row by row; Γ 1×M (f β ) is the power spectral density function corresponding to the spatial correlation function.

[0103] Step 5: Generate N clutter echo signals within one coherent processing time.

[0104] The pulse echo signal obtained in step 3 is respectively combined with the modulation matrix in step 4 Perform element-wise multiplication on each row of , and get N clutter echo signals:

[0105]

[0106] Step 6: cyclically generate a frame of clutter echo signal within each coherence time.

[0107] By introducing the Doppler spectrum defocusing characteristics of moving targets and realizing different time delays for the fast time dimension and slow time dimension in a frame of signal, the corresponding changes of the power spectrum corresponding to the time correlation and spatial correlation can be quickly realized. Based on this, the clutter echoes continuously received by the radar under long-term operation can be simulated.

[0108] Generate a defocus factor that causes a change in the power spectrum corresponding to the time correlation:

[0109]

[0110] In the formula, f χ is the frequency point corresponding to the fast time-dimensional Fourier transform of a clutter; τ pThe time delay for the pth clutter is set according to the spectrum width; p = 1, 2…, N, where N is the number of clutter echoes within the first coherence time.

[0111] Generate a defocus factor that causes a change in the power spectrum corresponding to the spatial correlation:

[0112]

[0113] In the formula, f ε is the frequency point corresponding to the slow-time Fourier transform at the same distance unit in a frame of clutter signal; τ q The delay for the slow time dimension data under the qth distance unit is set according to the spectrum width requirements; q = 1, 2…, M, where M is the number of fast time dimension distance units of the clutter signal.

[0114] For the N clutter echo signals within one coherent processing time obtained in step 5, defocusing processing is performed successively, and the processing result is used as a frame of clutter echo signals within the next coherent processing time, so as to realize the change of the statistical characteristics of each frame of clutter signal:

[0115]

[0116] In the formula, fft() row Indicates that the matrix is ​​Fourier transformed row by row; ifft() row It means that the matrix is ​​inverse Fourier transformed row by row;

[0117] The beneficial effects of the present invention are:

[0118] In step 1, considering that the radar illuminates the ground, the clutter reflection area presents a distributed scattering phenomenon, that is, the ground is no longer a point target, but a distribution of multiple scattering sources, and the echo signal is the result of the coherent superposition of the echoes of various scattering sources on the ground. Therefore, the azimuth-range grid division method is used to simulate the ground distributed clutter signal, which conforms to the form of coherent superposition of echoes.

[0119] In step 4, considering that the clutter signal is a type of random process, it is convenient to describe the clutter from a statistical point of view, and the process can be characterized by a one-dimensional probability density function and its autocorrelation function. The statistical modeling method is adopted, that is, data is generated according to the amplitude probability density function, time autocorrelation function and spatial autocorrelation function characteristics of the clutter. Since the power spectrum corresponding to the time autocorrelation function is related to the slow time dimension data of the clutter, and the generated modulation coefficient matrix is ​​multiplied with the signal to realize the clutter statistical characteristics of the signal, the modulation data that meets the three statistical characteristics is generated separately in step 4, combined with the signal after the coherent echo is superimposed in step 3, a frame of clutter data can be generated in the subsequent step 5.

[0120] When the radar is working continuously for a long time, the statistical characteristics of different frames of echo signals received in the same clutter area will change: the change of envelope probability distribution, the spectral center and spectral width of the power spectrum corresponding to time correlation and space correlation may also shift and broaden to a certain extent. If the change characteristics of the statistical model are given, the clutter can be more comprehensively characterized, and the effectiveness of the clutter suppression filter in the radar simulator can be effectively verified. The change of the clutter statistical model means that each frame of clutter signal will be continuously regenerated from step 4, which seriously affects the efficiency of radar simulator data generation. Therefore, in step 6, the first frame of clutter signal generated in step 5 is used to introduce defocusing factors in the fast time dimension and slow time dimension of the echo, respectively, to achieve the shift and broadening of the original clutter power spectrum, and the subsequent generation of each frame of clutter only needs to use the clutter data generated in the previous frame to execute step 6 cyclically, so as to achieve the rapid generation of clutter signals.

[0121] The present invention is beneficial for quickly generating clutter data based on the statistical characteristics of clutter and the introduced defocusing characteristics of moving target echoes. Figure 3 The clutter power spectrum at the first distance unit in the first frame is shown. Figure 6 The clutter power spectrum at the first distance unit in the second frame is shown. By comparing the two figures, it can be seen that the introduction of the defocus factor in the process of generating the second frame of data causes a small range change in the power spectrum position and spectrum width, which is consistent with the change of clutter echoes in a real environment. Figure 4 The spatial correlation of the first clutter signal and the corresponding power spectrum in the first frame are shown. Figure 7 The spatial correlation of the first clutter signal and the corresponding power spectrum in the second frame are shown. By comparing the two figures, it can be seen that the power spectrum position and spectrum width have also undergone small changes. Figure 8 It is the system time spent running the first frame of clutter data in the MATLAB environment. Since it is necessary to consider the division of clutter units, the calculation of the corresponding multiplication coefficients of the radar equation, and the superposition of clutter unit echoes, the time required for the first frame of clutter generation is relatively long. Save the first frame of generated data, and based on the defocusing characteristics, cyclically generate multiple frames of echo signals. Fig.10 In the MATLAB environment, based on the existing first frame data (56.861 seconds), the radar is simulated to use 50 circles. The system time spent on the clutter data from the 2nd to the 50th frame (19.103 seconds) shows that the operating efficiency is significantly improved.

Claims

1. A method for generating a clutter signal based on statistical characteristics, characterized in that: Step 1: determine the clutter area and divide the clutter units; Step 2: Calculate the amplitude multiplication coefficient K of each unit m , Where λ is the radar operating wavelength; p t is the radar transmit peak power; s m The area of ​​the mth clutter unit; G(θ) is the radar antenna gain; r(m) is the distance between the mth clutter unit and the radar; L is the radar comprehensive loss; σ m is the scattering area of ​​the mth clutter unit; Step 3: Generate a clutter unit echo signal, the echo is expressed as: Where N e N is the number of azimuth resolution units for regional division; r K is the number of distance units divided under a certain orientation unit; i,j Represents the multiplicative coefficient of the clutter unit at the i-th direction and the j-th distance; τ j represents the echo delay of the jth clutter range unit; T represents the signal duration; f0 is the signal center frequency; μ is the frequency modulation slope, and t represents the time variable; Step 4: Generate a clutter data modulation coefficient matrix, including the following steps: Steps to calculate the Gaussian power spectral density function: In the formula, f τ is the spectrum center frequency; σ f is the 3dB bandwidth of the spectrum, f α is the frequency independent variable; Transform the power spectrum of the white noise complex sequence into a Gaussian power spectrum density: G N×1 =ifft(fft(g N×1 )·P(f α )) According to the above process in step 4, the cycle is repeated M times, and each modulation vector G is generated. N×1 As each column of the modulation matrix data, M is the number of time sampling points of the clutter signal, which is determined according to the actual situation to obtain the modulation matrix A N×M : Modulation Matrix A N×M Perform spatial correlation processing to obtain the modulation coefficient matrix: In the formula, fft() row Indicates that the matrix is ​​Fourier transformed row by row; ifft() row represents the inverse Fourier transform of the matrix row by row; Γ 1×M (f β ) is the power spectral density function corresponding to the spatial correlation function; f β is the frequency independent variable of the power spectrum function, and the power spectrum density function Γ 1×M (f β ) = fft(Γ(d)), where the spatial correlation function is: Where, 0≤a≤1 represents the proportion of the fast fluctuation component in the total energy in the spatial correlation function, b and c are the decorrelation distances of the fast and slow fluctuation components, d represents the correlation distance, and e is the natural logarithm; Step 5, generating N clutter echo signals within one coherent processing time; The pulse echo signal is multiplied by the modulation matrix row by row to obtain N clutter echo signals: Clusig N×M is the N clutter echo signals after modulation; Step 6, cyclically generate a frame of clutter echo signal within each coherence time; Define the first defocus factor: In the formula, f χ is the frequency point corresponding to the fast time-dimensional Fourier transform of a clutter; τ p is the delay achieved for the pth clutter in a frame; p = 1,2L,N, where N represents the number of clutter echoes within one coherence time; Compute the power spectrum of the expanded clutter data: is the focal factor acting on the clutter signal Clusig N×M The signal autocorrelation matrix calculated later; In the formula, fft() row Indicates Fourier operation on the row dimension of the data matrix; ifft() row Indicates inverse Fourier transform in row dimension; According to the power spectrum corresponding to the spatial correlation of clutter, the second defocus factor is defined as: In the formula, f ε is the frequency point corresponding to the slow-time Fourier transform at the same distance unit in a frame of clutter signal; τ q is the delay achieved for the slow time dimension data at the qth range unit; q = 1,2L,M, where M is the number of fast time dimension range units of the clutter signal; Clutter data with power spectrum expansion corresponding to spatial correlation: In the formula, fft() col Indicates that Fourier operation is performed on the column dimension of the data matrix; []′ indicates transposing the matrix; ifft() col Indicates the inverse Fourier transform in column dimension; U N×M That is, a new frame of clutter data is generated based on the first frame of clutter data by utilizing the defocusing characteristics.

2. The method for generating a clutter signal based on statistical characteristics according to claim 1, characterized in that: The clutter unit includes the distance unit and the azimuth unit. The distance unit formula is: B is the bandwidth of the transmitted signal, c is the speed of light, and the azimuth unit is the azimuth resolution of the radar system parameters.

3. The method for generating a clutter signal based on statistical characteristics according to claim 1, characterized in that: The scattering area σ in step 2 m =σ0·s m , where is the area of ​​the mth clutter unit; σ0 represents the average backscattering coefficient of the clutter area; represents the azimuth resolution; ΔR represents the range resolution.

Images