A sky wave radar method for suppressing narrowband interference based on distance frequency domain sampling

Abstract

The application discloses a sky wave radar narrow-band interference suppression method based on distance frequency domain sampling, which comprises the following steps: performing Fourier transform on three-dimensional array element level baseband echoes in a radar coherent processing interval along a distance dimension to obtain three-dimensional distance frequency domain echoes; taking three-dimensional distance frequency domain echo envelopes to perform two-dimensional integration to obtain one-dimensional distance frequency domain echoes; taking one-dimensional distance frequency domain echoes to perform maximum value outlier point detection to estimate interference frequency points; according to the estimated interference frequency points, interference distance frequency domain samples are estimated from the three-dimensional distance frequency domain echoes, and a sample covariance matrix is calculated; the covariance matrix is inversed by using a sample matrix inversion (SMI) algorithm, an optimal weight vector is calculated, and narrow-band interference is suppressed by using an adaptive beam forming technology. The algorithm has small calculation amount and is convenient for engineering implementation. Meanwhile, the application inherits all advantages of the SMI algorithm, uses the frequency domain sparse characteristics of the narrow-band interference, obtains purer interference samples than traditional time domain sampling, and significantly improves interference suppression efficiency.

Description

Technical Field

[0001] This application relates to the technical field of interference filtering, and in particular to a method for suppressing narrowband interference in skywave radar based on range-frequency domain sampling. Background Technology

[0002] Skywave over-the-horizon (OTHR) radar utilizes ionospheric reflection for target detection, overcoming the detection blind spots caused by the Earth's curvature and offering advantages such as long detection range and protection against low-altitude penetration. However, OTHR operates at a low frequency (less than 30MHz), making it susceptible to unintentional interference from radio frequency and transient sources, primarily narrowband interference with concentrated power on a limited number of frequencies. Simultaneously, narrowband active jamming (such as communication jamming) launched by the enemy also poses a significant threat to skywave radar.

[0003] Currently, there is a wealth of research on radar jamming suppression, but relatively little research specifically on narrowband jamming suppression. Generally speaking, existing methods can be categorized into five main types: sidelobe masking (SLB), blind source separation (BSS), frequency domain narrowband filtering, sidelobe cancellation (SLC), and adaptive beamforming (ADBF). Among these methods, SLB-based methods can suppress false target interference from sidelobes, but narrowband interference typically exhibits suppression characteristics in the echo range dimension, and using SLB can lead to excessive concealment, resulting in target loss. BSS-based methods can suppress narrowband interference from sidelobes and near the main lobe; however, BSS inherently has uncertainties in signal amplitude and sequence after separation, making accurate target angle measurement difficult after interference suppression. Frequency domain narrowband filtering is an effective way to combat narrowband interference; however, unlike other radar systems, OTHR radar echoes contain a large amount of strong clutter, and frequency domain filtering changes the clutter characteristics. After pulse Doppler (PD) processing, the clutter exhibits frequency domain broadening, affecting the detection of low-speed targets. SLC and ADBF-based methods can suppress narrowband interference from sidelobes and near the main lobe; however, OTHR radar is a continuous wave system, and the target echo is also a continuous wave. The proportion of narrowband interference within a pulse repetition period is usually very high, making it inevitable that traditional time domain sampling methods will also sample the target echo, resulting in a significant loss of signal-to-noise ratio after interference suppression. In addition, the number of range cells in one pulse repetition cycle of OTHR radar is usually small, and the number of time-domain samples required to estimate the covariance matrix is ​​usually insufficient to meet the target 3dB loss requirement.

[0004] Based on this, this application provides a method for suppressing narrowband interference in skywave radar based on range-frequency domain sampling to solve the above-mentioned technical problems. Summary of the Invention

[0005] To help solve the above-mentioned technical problems, this application provides a method for suppressing narrowband interference in skywave radar based on range-frequency domain sampling.

[0006] This application provides a method for suppressing narrowband interference in skywave radar based on range-frequency domain sampling, which employs the following technical solution:

[0007] First aspect

[0008] A method for suppressing narrowband interference in skywave radar based on range-frequency domain sampling includes the following steps:

[0009] S1. Perform Fourier transform along the range dimension on the three-dimensional array element-level baseband echo within the coherent processing interval (CPI) of the low-frequency radar to obtain the three-dimensional range frequency domain echo.

[0010] S2. Take the three-dimensional range frequency domain echo envelope, and then perform two-dimensional integration along the slow time and spatial domains respectively to obtain the one-dimensional range frequency domain echo.

[0011] S3. Take the one-dimensional range frequency domain echo within the radar signal bandwidth to detect maximum outliers and estimate the interference frequency.

[0012] S4. Based on the estimated interference frequency, estimate the interference range frequency domain samples from the three-dimensional range frequency domain echo, and calculate the covariance matrix of the interference range frequency domain samples.

[0013] S5. The covariance matrix is ​​inverted using the Sampling Matrix Inversion (SMI) algorithm to calculate the optimal weight vector, and narrowband interference is suppressed by adaptive beamforming (ADBF) technology.

[0014] Optionally, step S1 includes:

[0015] S11. Let x be the baseband echo of the array element level within a coherent processing interval (CPI). r (t k ,t f ,t m ), t k For spatial time, t f For fast time, t m For slow time, the discrete forms are as follows:

[0016]

[0017] Where N is the number of array elements, L is the number of echo range gates, and M is the number of accumulated pulses; d is the array element interval, t is the number of echo range gates, and t is the number of echo range gates. s For echo sampling period, T r The radar pulse repetition period is c = f0λ, where c is the speed of light, f0 is the radar carrier frequency, and λ is the wavelength.

[0018] S12, along fast time t f Calculate x r (t k ,t f ,t mThe Fourier transform of ) yields the three-dimensional range-frequency domain echo as

[0019]

[0020] Where, f∈[-f s / 2,f s [ / 2] represents the distance frequency domain, f s The sampling frequency.

[0021] Optionally, step S2 includes:

[0022] S21. Obtain the three-dimensional distance-frequency domain echo x r (t k ,f,t m The envelope of ) along the spatial time t k Slow time t m Performing a two-dimensional integral yields a one-dimensional range-frequency domain echo.

[0023]

[0024] Optionally, step S3 includes:

[0025] S31. Obtain the one-dimensional range-frequency domain echo X within the radar signal bandwidth. r (f), f∈[-B / 2,B / 2], where B is the radar signal bandwidth. Then, maximum outlier detection is performed to estimate P interference frequency points. p = 0, 1, 2, ... P-1.

[0026] Optionally, step S4 includes:

[0027] S41. Based on the estimated interference frequency Estimating interference range-frequency samples from three-dimensional range-frequency echoes

[0028]

[0029] The estimated sample for the nth array element is:

[0030]

[0031] in, The range-frequency domain of the echo received by the nth array element during the m-th pulse repetition period is: The amplitude, then The second line is placed to the right of the first line, the third line is placed to the right of the second line, and so on. Rewritten as a 1×PM vector:

[0032]

[0033] Where i = 0, 1, 2, ... PM-1. The estimated samples of all array elements (N) are obtained as follows:

[0034]

[0035] S42. Calculate the estimated samples of all array elements. covariance matrix

[0036]

[0037] Optionally, step S5 includes:

[0038] S51. Use the Sample Matrix Inversion (SMI) algorithm to invert the covariance matrix and calculate the optimal weight vector.

[0039]

[0040] Where a(θ) is the receiving guidance vector;

[0041] S52, using the optimal weight vector W opt Adaptive beamforming is used for the baseband echo at the epoch level to suppress narrowband interference.

[0042] Optionally, the Fourier transform in step S1 is a Fast Fourier Transform (FFT).

[0043] Optionally, the frequency range of the one-dimensional range frequency domain echo in step S3 is the same as the radar signal bandwidth.

[0044] Optionally, the criterion for detecting maximum outliers in step S3 is one of the following: 3 times the standard deviation criterion, 1.5 times the interquartile range criterion, or 2.5 times the absolute deviation of the median criterion.

[0045] Optionally, the low-frequency radar includes a skywave radar.

[0046] In summary, this application includes at least one of the following beneficial technical effects:

[0047] 1. This invention uses Fourier transform to convert the baseband echo to the range frequency domain and estimates the interference frequency using outlier detection. The algorithm has low computational complexity and is easy to implement in engineering. Furthermore, this invention inherits all the advantages of the SMI algorithm, utilizing the sparsity characteristics of narrowband interference in the frequency domain to obtain cleaner interference samples than traditional time-domain sampling, significantly improving suppression efficiency.

[0048] 2. This invention is not only applicable to OTHR radar, but also to other low-frequency array radar systems used for narrowband interference suppression.

[0049] 3. The sampling matrix inversion (SMI) algorithm used in this invention is a mature technology that only requires adjustment from time-domain sampling to range-frequency domain sampling. It can provide technical improvements for radar equipment that already uses adaptive beamforming technology.

[0050] 4. The method of this invention only involves the signal processing flow, that is, it only requires upgrading the processing system and software, without changing other system structures, and has the value of promotion and application. Attached Figure Description

[0051] Figure 1 This is a schematic diagram of the method for suppressing narrowband interference according to the present invention.

[0052] Figure 2 This is a flowchart of the method for suppressing narrowband interference according to the present invention.

[0053] Figure 3 It is a distance Doppler image before interference suppression.

[0054] Figure 4 It is a range Doppler image after interference suppression using traditional methods.

[0055] Figure 5 It is a range Doppler image after interference suppression using the method of this invention.

[0056] Figure 6 This is the receiver radiation pattern (displayed in polar coordinates).

[0057] Figure 7 This is the receiver pattern (partial view). Detailed Implementation

[0058] The following is in conjunction with the appendix Figure 1-7 This application will be described in further detail.

[0059] Reference Figure 1 The embodiments of the present invention consist of the following steps: array element-level baseband echo 1, range-frequency domain echo 2, interference frequency estimation 3, interference range-frequency domain sample 4, sample covariance matrix 5, optimal weight vector 6, and adaptive beamforming 7. In this embodiment, array element-level baseband echo 1 refers to the OTHR radar array element-level baseband echo; range-frequency domain echo 2 is obtained by calculating the Fourier transform of the array element-level baseband echo along the fast time (range dimension); interference frequency estimation 3 uses the maximum outlier detection method to estimate the frequency of narrowband interference from the range-frequency domain echo; interference range-frequency domain sample 4 is extracted from the range-frequency domain echo based on the estimated interference frequency; sample covariance matrix 5 is the covariance matrix of the interference sample; optimal weight vector 6 is calculated using the sampling matrix inversion (SMI) algorithm; and adaptive beamforming 7 uses the optimal weight vector to perform adaptive beamforming, thereby suppressing narrowband interference.

[0060] This application discloses a method for suppressing narrowband interference in skywave radar based on range-frequency domain sampling. (Refer to...) Figure 2 This includes the following steps:

[0061] S1. Perform Fourier transform along the range dimension on the three-dimensional array element-level baseband echo within the coherent processing interval (CPI) of the low-frequency radar to obtain the three-dimensional range frequency domain echo.

[0062] The low-frequency radar includes skywave radar or other low-frequency array radar systems. In this embodiment, the low-frequency range of the low-frequency radar is less than 30MHz.

[0063] Specifically, S1 includes:

[0064] S11. Let x be the baseband echo of the array element level within a coherent processing interval (CPI). r (t k ,t f ,t m ), t k For spatial time, t f For fast time, t m For slow time, the discrete forms are as follows:

[0065]

[0066] Where N is the number of array elements, L is the number of echo range gates, and M is the number of accumulated pulses; d is the array element interval, t is the number of echo range gates, and t is the number of echo range gates. s For echo sampling period, T r The radar pulse repetition period is c = f0λ, where c is the speed of light, f0 is the radar carrier frequency, and λ is the wavelength.

[0067] In this embodiment, the number of array elements is set to 64, the number of echo range gates is set to 800, and the number of accumulated pulses is set to 128. r (t k ,t f ,t m The dimension is 64×800×128.

[0068] S12, along fast time t f Calculate x r (t k ,t f ,t m The Fast Fourier Transform (FFT) of the signal yields the three-dimensional range-frequency echo.

[0069]

[0070] Where, f∈[-f s / 2,f s [ / 2] represents the distance frequency domain, f sThe sampling frequency.

[0071] In this embodiment, the number of array elements N is set to 64, the number of echo range gates L is set to 800, and the number of accumulated pulses M is set to 128. r (t k ,f,t m The dimension is also 64×800×128.

[0072] S2. Take the three-dimensional distance-frequency domain echo envelope, and then perform two-dimensional integration along the slow time and spatial domains respectively to obtain the one-dimensional distance-frequency domain echo.

[0073] Specifically, S2 includes:

[0074] S21. Obtain the three-dimensional distance-frequency domain echo x r (t k ,f,t m The envelope of ) along the spatial time t k Slow time t m Performing a two-dimensional integral yields a one-dimensional range-frequency domain echo.

[0075]

[0076] In this embodiment, the number of array elements N is set to 64, the number of echo range gates L is set to 800, the number of accumulated pulses M is set to 128, and X... r (f) The dimension is 1×800.

[0077] S3. Take the one-dimensional range frequency domain echo within the radar signal bandwidth to detect maximum outliers and estimate the interference frequency.

[0078] Among them, the frequency domain range of the one-dimensional range frequency domain echo is the same as the radar signal bandwidth; the criterion for detecting maximum outliers is one of the following: 3 times the standard deviation criterion, 1.5 times the interquartile range criterion, or 2.5 times the median absolute deviation criterion.

[0079] Specifically, S3 includes:

[0080] S31. Obtain the one-dimensional range-frequency domain echo X within the radar signal bandwidth. r (f), f∈[-B / 2,B / 2], where B is the radar signal bandwidth. Then, maximum outlier detection is performed to estimate P interference frequency points. p = 0, 1, 2, ... P-1.

[0081] In this embodiment, the number of array elements N is set to 64, the number of echo range gates L is set to 800, the number of accumulated pulses M is set to 128, and the sampling frequency f is set to... s =2B, taking the one-dimensional range-frequency domain echo X within the radar signal bandwidth. r (f), f∈[-B / 2,B / 2],X rThe dimension of (f) is 1 × 400, where 400 = 800 × B / f s .

[0082] S4. Based on the estimated interference frequency, estimate the interference range frequency domain samples from the three-dimensional range frequency domain echo, and calculate the covariance matrix of the interference range frequency domain samples.

[0083] Specifically, S4 includes:

[0084] S41. Based on the estimated interference frequency Estimating interference range-frequency samples from three-dimensional range-frequency echoes

[0085]

[0086] The estimated sample for the nth array element is:

[0087]

[0088] in, The range-frequency domain of the echo received by the nth array element during the m-th pulse repetition period is: The amplitude, then The second line is placed to the right of the first line, the third line is placed to the right of the second line, and so on. Rewritten as a 1×PM vector:

[0089]

[0090] Where i = 0, 1, 2, ... PM-1. The estimated samples of all array elements (N) are obtained as follows:

[0091]

[0092] In this embodiment, the number of array elements N is set to 64, the number of echo range gates L is set to 800, the number of accumulated pulses M is set to 128, and the number of interference frequency points P is set to 10. Therefore, the interference range frequency domain sample... The dimension is 64×10×128, before rearrangement The dimension is 128×10, and after rearrangement it is 1×1280. The dimension is 64×1280.

[0093] S42. Calculate the estimated samples of all array elements. covariance matrix

[0094]

[0095] S5. The covariance matrix is ​​inverted using the Sampling Matrix Inversion (SMI) algorithm to calculate the optimal weight vector, and narrowband interference is suppressed by adaptive beamforming (ADBF) technology.

[0096] Specifically, S5 includes:

[0097] S51. Use the Sample Matrix Inversion (SMI) algorithm to invert the covariance matrix and calculate the optimal weight vector.

[0098]

[0099] Where a(θ) is the receiving guidance vector;

[0100] In this embodiment, the number of array elements N is set to 64, the number of echo range gates L is set to 800, the number of accumulated pulses M is set to 128, and the number of interference frequency points is set to 10. Therefore, the interference range frequency domain sample... The dimension is 64×10×128, before rearrangement The dimension is 128×10, and after rearrangement it is 1×1280. The dimension is 64×1280. The dimension is 64×64, W opt The dimension is 64×1.

[0101] S52, using the optimal weight vector W opt Adaptive beamforming is performed on the baseband echo at the phasor level to suppress narrowband interference. Then, pulse compression, coherent accumulation, and target detection are applied to the beam echo.

[0102] This invention uses Fast Fourier Transform (FFT) to convert the baseband echo to the range frequency domain and estimates the interference frequency using outlier detection. The algorithm has low computational cost and is easy to implement in engineering. Furthermore, this invention inherits all the advantages of the SMI algorithm, utilizing the sparsity characteristics of narrowband interference in the frequency domain to obtain cleaner interference samples than traditional time-domain sampling, significantly improving suppression performance.

[0103] Figures 3-5 This is the distance-Doppler plot (RD plot) of the measured data in this application.

[0104] Range-Doppler plots are a commonly used image representation method in radar signal processing. They can be used to show the variation of pulse echo signals received by radar in terms of range and Doppler frequency.

[0105] The measured data in the figure show that, compared with the traditional time-domain sampling beamforming method, optimizing the sample selection can significantly improve the echo signal-to-interference ratio and the reproduction of weak targets after interference suppression.

[0106] Reference Figure 6 and Figure 7As can be seen, the received pattern obtained by the sampling method of "interference pulse + range frequency domain" has a lower grating lobe and a deeper notch at the interference DOA.

[0107] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for suppressing narrowband interference in skywave radar based on range-frequency domain sampling, characterized in that, Includes the following steps: S1. Perform a Fourier transform along the range dimension on the three-dimensional array element-level baseband echo within the coherent processing interval (CPI) of the low-frequency radar to obtain the three-dimensional range frequency domain echo; S2. Take the envelope of the three-dimensional range frequency domain echo and then perform two-dimensional integration along the slow time and spatial domains respectively to obtain the one-dimensional range frequency domain echo; S3. Take the one-dimensional range frequency domain echo within the radar signal bandwidth and perform maximum outlier detection to estimate the interference frequency; S4. Based on the estimated interference frequency, estimate the interference range frequency domain samples from the three-dimensional range frequency domain echo and calculate the covariance matrix of the interference range frequency domain samples; S5. The covariance matrix is ​​inverted using the SMI algorithm to calculate the optimal weight vector, and narrowband interference is suppressed by adaptive beamforming (ADBF) technology. Step S1 includes: S11, setting the array element-level baseband echo within a coherent processing interval CPI as x. r (t k ,t f ,t m ), t k For spatial time, t f For fast time, t m For slow time, the discrete forms are as follows: Where N is the number of array elements, L is the number of echo range gates, and M is the number of accumulated pulses; d is the array element interval, t is the number of echo range gates, and t is the number of echo range gates. s For echo sampling period, T r The radar pulse repetition period; c = f0λ, where c is the speed of light, f0 is the radar carrier frequency, and λ is the wavelength; S12, along fast time t f Calculate x r (t k ,t f ,t m The Fourier transform of ) yields the three-dimensional range-frequency domain echo as Where, f∈[-f s / 2,f s [ / 2] represents the distance frequency domain, f s The sampling frequency; Step S2 includes: S21, obtaining the three-dimensional distance-frequency domain echo x r (t k ,f,t m The envelope of ) along the spatial time t k Slow time t m Performing a two-dimensional integral yields a one-dimensional range-frequency domain echo. ; Step S3 includes: S31, obtaining the one-dimensional range frequency domain echo X within the radar signal bandwidth. r (f), f∈[-B / 2,B / 2], where B is the radar signal bandwidth. Then, maximum outlier detection is performed to estimate P interference frequency points. p = 0, 1, 2, ..., P-1; Step S4 includes: S41, based on the estimated interference frequency point Estimating interference range-frequency samples from three-dimensional range-frequency echoes The estimated sample for the nth array element is: in, The range-frequency domain of the echo received by the nth array element during the m-th pulse repetition period is: The amplitude, then The second line is placed to the right of the first line, the third line is placed to the right of the second line, and so on. Rewritten as a 1×PM vector: Where i = 0, 1, 2, ... PM-1, from The N estimated samples of all array elements are obtained as follows: S42. Calculate the estimated samples of all array elements. covariance matrix 。 2. The method for suppressing narrowband interference in skywave radar based on range-frequency domain sampling according to claim 1, characterized in that, Step S5 includes: S51, inverting the covariance matrix using the SMI algorithm to calculate the optimal weight vector. Where a(θ) is the receiving guidance vector; S52, using the optimal weight vector W opt Adaptive beamforming is used for the baseband echo at the epoch level to suppress narrowband interference.

3. The method for suppressing narrowband interference in skywave radar based on range-frequency domain sampling according to claim 1, characterized in that, The Fourier transform in step S1 is the Fast Fourier Transform (FFT).

4. The method for suppressing narrowband interference in skywave radar based on range-frequency domain sampling according to claim 1, characterized in that, The frequency range of the one-dimensional range frequency domain echo in step S3 is the same as the bandwidth of the radar signal.

5. The method for suppressing narrowband interference in skywave radar based on range-frequency domain sampling according to claim 1, characterized in that, The criteria for detecting maximum outliers in step S3 are one of the following: 3 times the standard deviation criterion, 1.5 times the interquartile range criterion, and 2.5 times the absolute deviation of the median criterion.

6. The method for suppressing narrowband interference in skywave radar based on range-frequency domain sampling according to claim 1, characterized in that, The low-frequency radar includes skywave radar.

Images