CLEAN single-channel external radiation source bistatic radar direct wave suppression method based on principal component analysis

Abstract

The invention discloses a CLEAN single-channel external radiation source bistatic radar direct wave suppression method based on principal component analysis, and belongs to the field of signal processing. After echo channel signals are acquired, the distance of each frame of signal is compressed, and the distance offset and Doppler broadening are compensated according to an equivalent distance model; the method comprises the following steps: firstly, carrying out azimuth compression on a distance Doppler domain, then carrying out accumulation average processing on each frame in the distance Doppler domain after azimuth compression, then carrying out PCA principal component analysis to reconstruct a complete direct wave as a reference interference signal of a single-channel echo signal, and carrying out CLEAN correlation coefficient calculation in the distance Doppler domain to complete suppression and cancellation of the direct wave in a single channel. According to the method, strong direct wave interference in a satellite-borne passive radar system under single-channel receiving is suppressed, so that a target with relatively weak energy is not easily covered by strong direct waves or other background interference, and a foundation is laid for subsequent target detection and tracking processing.

Description

Technical Field

[0001] This invention belongs to the field of signal processing and relates to a method for suppressing direct waves from external radiation sources under single-channel data reception. Specifically, it relates to a CLEAN single-channel external radiation source bistatic radar direct wave suppression method based on principal component analysis. Background Technology

[0002] External radiation source radar does not actively transmit signals; instead, it utilizes existing signals in space to receive the reflected echoes from third-party radiation sources after they have passed through the target for target detection. This offers significant advantages in terms of concealment, anti-stealth capabilities, and anti-jamming performance. Space-based external radiation sources are an important signal source for external radiation source radar. These sources are launched from high / medium orbit satellites and received by low orbit satellites, maritime platforms, or ground-based platforms. Compared to traditional external radiation sources, their location is farther from the Earth's surface, resulting in less multipath interference and clutter. Multiple radiation sources can be used in conjunction to detect the same target, improving the signal-to-noise ratio. Furthermore, the types of space-based external radiation sources available are constantly evolving, including GNSS and GSM, making research on space-based external radiation sources indispensable.

[0003] The study of space-based external radiation source radar first requires direct wave suppression. In external radiation source radar, the bistatic configuration of a single-channel space-based external radiation source radar system means that the echo signal received by the radar system includes not only the target echo signal but also the direct wave signal from the space-based radiation source. Meanwhile, considering that the direct wave signal does not undergo target backscattering and its transmission distance is shorter than the target scattered echo signal, the strength of the direct wave signal is much greater than the target echo signal, thus causing effective interference to the target. Therefore, direct wave reconstruction is a crucial step in direct wave suppression. Furthermore, the relative motion between the radiation source and the receiving platform causes an approximately linear translation of the direct wave signal along the range dimension, resulting in range migration (RM) and Doppler spread (DM) in the direct wave signal.

[0004] The CLEAN algorithm was first proposed for application in astronomical image observation, used to image weak radio sources in the presence of strong nearby emission sources. Its basic idea is that the image to be processed can be considered as a convolution of the observation system's response and brightness distribution function; the part to be removed is the system-induced response, and the CLEAN algorithm is used to eliminate these effects. This idea has been introduced into direct wave suppression. In recent years, the CLEAN method has been widely used in radar signal processing. However, existing direct wave suppression methods mainly rely on two receiving channels: one direct wave channel and one reflected wave channel. While single-channel data reception simplifies system structure and reduces costs, effectively extracting and suppressing direct wave information remains a challenging problem in single-channel data reception mode. Summary of the Invention

[0005] To address the aforementioned problems, this invention proposes a CLEAN single-channel external radiation source bistatic radar direct wave suppression method based on principal component analysis. In the case of only a single receiving channel, this method can effectively extract and reconstruct the direct wave, thereby canceling and suppressing the direct wave in the reflected wave channel, laying the foundation for subsequent target detection and tracking processing.

[0006] This invention proposes a CLEAN single-channel external radiation source bistatic radar direct wave suppression method based on principal component analysis, which specifically includes the following steps:

[0007] Step 1: Read in multiple consecutive frames of single-channel echo data, transmitter position and velocity parameters, receiver position and velocity parameters, and radar system parameters.

[0008] Step 2: Perform range pulse compression, range migration compensation, and Doppler phase compensation on each frame of signal to complete direct wave synchronization.

[0009] Step 3: After the range compression and compensation of the range error Doppler phase error in Step 2, perform azimuth fast Fourier transform on the echo signals of each frame; further, perform cumulative averaging on the two-dimensional compressed first (k-1) frame signals in the range Doppler domain.

[0010] Step 4: Perform PCA principal component analysis on the accumulated averaged echo signal, and select the direct wave corresponding to the principal component with the maximum energy as the reconstructed direct wave signal.

[0011] Step 5: Using the direct wave reference signal after two-dimensional compression and reconstruction as the reference signal for the CLEAN direct wave suppression algorithm, calculate the CLEAN correlation coefficient CR; further, use the calculated correlation coefficient... Multiplying the signal by the reconstructed reference signal yields the direct wave interference component to be removed by the CLEAN algorithm. Further subtraction removes the interference, ultimately yielding the CLEAN-processed reflected wave signal.

[0012] The advantages of this invention are:

[0013] 1. This invention performs direct wave reconstruction on the echo of the reflected channel in single-channel receiving mode, without the need to use the reference channel in dual-channel receiving mode, reducing the system design cost and overcoming the limitation of the difficulty in obtaining the analytical expression of the reference signal of the space-based external radiation source system. It also eliminates the need to estimate and reconstruct the reference signal in advance, and directly derives the algorithm from the existing reflected channel data, which greatly reduces the difficulty of algorithm implementation.

[0014] 2. This invention is applicable to various receiving scenarios. It is suitable for receiver modes without a reference channel and for situations with multiple receivers receiving through a single channel. It can reconstruct and cancel interference in the direct wave signals received by each receiver, thereby achieving effective direct wave suppression. This method can be applied not only to single-base systems with single-channel reception but also to multi-base radar systems composed of multiple single-channel receiving platforms. Attached Figure Description

[0015] Figure 1 This is a flowchart of the CLEAN single-channel external radiation source bistatic radar direct wave suppression method based on principal component analysis according to the present invention.

[0016] Figure 2 This is a schematic diagram of a single-channel dual-base external radiation source radar system provided in an embodiment of the present invention.

[0017] Figure 3 This is the distance-direction coherent detection result after distance migration compensation processing.

[0018] Figure 4 This represents the processing result in the range-Doppler domain after phase compensation processing.

[0019] Figure 5 The result is the range-Doppler domain processing result of the reconstructed direct wave.

[0020] Figure 6 This represents the processing results in the range-Doppler domain after direct wave suppression.

[0021] Figure 7 A comparison of cross-sections in the azimuth Doppler domain before and after direct wave suppression.

[0022] Figure 8 The curves showing the similarity between direct wave interference and suppressed echo as a function of INR. Detailed Implementation

[0023] The present invention will now be described in further detail with reference to the accompanying drawings.

[0024] This invention relates to a CLEAN single-channel external radiation source bistatic radar direct wave suppression method based on principal component analysis, such as... Figure 1 As shown, the specific steps are as follows:

[0025] Step 1: Read in multiple consecutive frames of single-channel echo data, transmitter position and velocity parameters, receiver position and velocity parameters, and radar system parameters.

[0026] The radar system parameters include PRF, carrier frequency, Kr, sampling rate, pulse width, azimuth accumulation time, and bandwidth. The input echo data for each frame of the reflection channel is as follows:

[0027]

[0028]

[0029] In the formula, Indicates the direct amplitude value; Indicates the amplitude of the reflected wave; Indicates the reflected echo signal. Indicates a direct wave signal; Indicates the target echo signal; It is Gaussian white noise. , For the real and imaginary parts of the complex Gaussian noise; Indicates distance and time; Indicates location and time. The speed of light; , and These represent the distances from the transmitter to the receiver, from the transmitter to the target, and from the target to the receiver over time, respectively.

[0030] Step 2: Perform range pulse compression, range migration compensation, and Doppler phase compensation on each frame of signal to complete direct wave synchronization;

[0031] a. Based on the ephemeris parameters of the radiation source platform and the motion information of the receiving platform, an equivalent distance model is used to characterize the relative distance variation between the radiation source and the receiving platform, including the equivalent distance. Doppler center frequency Doppler frequency modulation .

[0032] A Taylor expansion of the equivalent distance model is performed, and the second-order result is used for approximation. The expression for the equivalent distance model is:

[0033]

[0034] exist Taylor's expansion:

[0035]

[0036] in, For wavelength, This indicates the relative equivalent flight speed between the radiation source platform and the receiver platform; Indicates the equivalent oblique angle of view; Doppler center frequency. Doppler frequency modulation The relationship between the platform's motion characteristics and the following formula is shown:

[0037]

[0038]

[0039] In the formula, This represents the relative position vector between the radiation source platform and the receiver platform; This represents the relative velocity vector between the radiation source platform and the receiver platform.

[0040] b. Calculate the distance migration factor and phase Doppler factor It is used for distance and phase compensation.

[0041]

[0042] In the formula, and For distance delay error and Doppler phase error:

[0043]

[0044] ;

[0045] In the formula, It represents the complex unit and is used to construct phase compensation factors in complex exponential form.

[0046] c. Perform range-direction matched filtering, range migration compensation, and Doppler phase compensation on each frame of the single-channel signal.

[0047]

[0048]

[0049] in, The frequency domain response of the matched filter is used for range compression. For the range-directed frequency variable, It is a distance-to-fast Fourier transform. It is the inverse fast Fourier transform of the distance direction.

[0050] Step 3: After range compression and Doppler phase error compensation in Step 2, perform azimuth-directed Fast Fourier Transform on each frame of echo signal:

[0051]

[0052] In the formula, For azimuth to Fast Fourier Transform, For azimuth frequency variables.

[0053] By employing range-azimuth two-dimensional compression, the energy of the direct wave signal is made more prominent; furthermore, the compressed first (k-1) frames of signal are accumulated and averaged in the range-Doppler domain.

[0054] .

[0055] Step 4: PCA direct wave reconstruction.

[0056] PCA principal component analysis was performed on the cumulatively averaged echo signal. The method is as follows:

[0057] The accumulated average echo signal Represented in matrix form as ;in This represents the number of sampling points in the azimuth-time dimension. This represents the number of sampling points over time.

[0058] First, for the matrix To remove the mean: ;in, It represents the matrix mean.

[0059] Subsequently, after removing the mean... Performing singular value decomposition (SVD) yields:

[0060] in, It is an orthogonal matrix, and its column vectors are respectively the pairwise matrices. and eigenvectors. Let be a diagonal matrix composed of singular values, with each element being a singular value and arranged in ascending order. , It is the number of singular values; Principal component number;

[0061] The above Main component and corresponding characteristic signals The weighted sum can be expressed as:

[0062]

[0063] in, .

[0064] Since the direct wave has the highest energy in the signal, it usually corresponds to the maximal singular value term in singular value decomposition. Therefore, the principal component with the highest energy is chosen as the direct wave reference signal for reconstruction. This effectively preserves the main components of the direct wave and suppresses the effects of the target and noise.

[0065] Step 5: CLEAN direct wave cancellation.

[0066] 1) Using the direct wave reference signal after two-dimensional compression and reconstruction as the reference signal for the direct wave suppression CLEAN algorithm, calculate the CLEAN correlation coefficient CR.

[0067] Considering that the signal-to-noise ratio of the direct wave is higher after range-azimuth compression, CLEAN cancellation is performed in the range-Doppler domain to reconstruct the direct wave reference signal from principal component analysis. Combine the current frame echo signal to be processed, after undergoing range-azimuth two-dimensional compression, range migration, and phase compensation as described in steps 2 and 3 above. Calculate the weighted correlation coefficient of single-channel data .

[0068]

[0069] In the formula, express Complex conjugate; Indicates the azimuth direction of the reconstructed direct wave signal. The distance corresponding to each pulse

[0070] To signal data; Indicates the azimuth direction of the echo signal. The range signal data corresponding to each pulse.

[0071] 2) Calculate the correlation coefficient Multiplying the signal by the reconstructed reference signal yields the direct wave interference component to be removed by the CLEAN algorithm. Further subtraction removes the interference, ultimately resulting in the CLEAN-processed reflected wave signal. , is represented as:

[0072] . Example

[0073] To verify the effectiveness of the method of this invention, a system was constructed using MATLAB software as follows. Figure 2 The simulation platform shown simulates a real single-channel receiver with a dual-base external radiation source scenario. System simulation processing and performance analysis are performed. A northeast-sky coordinate system is established with the receiver's location as the origin. The target is set as a moving aircraft, and the radiation source is a high-orbit satellite. The system reads in the position and velocity coordinates of the transmitter and receiver, signal bandwidth, carrier frequency, PRF, etc. The system parameter settings are shown in Table 1.

[0074] Table 1 Simulation Parameters

[0075] carrier frequency 1.25GHz Target location [-5km, 8.66km, 100km] PRF 1500Hz Target speed [50m / s, 86.6m / s, 0] Kr 40GHz / s Sampling rate 10MHz Location accumulation time 1s Interference ratio -22db bandwidth 4MHz Pulse width 100μs Satellite position [-708km, -2138km, 36000km] Receiver position [0,0,0] satellite speed [3000m / s,0,0] Receiver speed [0,0,0]

[0076] Direct wave suppression is achieved using the method of this invention. During the process, the following steps are selected: The component is the reconstructed direct wave reference signal, i.e. Finally obtained And further draw the RD diagram, such as Figures 3-7 The results of direct wave suppression processing in a single-channel receiving radar system are systematically presented. Figure 3 and Figure 4 The results of range-direction coherent detection and the processing results in the range-Doppler domain after range migration compensation and phase compensation are presented. Compared with the Doppler spectral spread before compensation, the direct wave signal is effectively focused after compensation. Figure 5 The direct wave reconstructed from multiple frames of target echo signals using PCA is presented. After two-dimensional compression in range and azimuth, an improved CLEAN algorithm is used for direct wave cancellation. The processed direct wave suppression results in the range-Doppler domain are shown below. Figure 6 As shown. Figure 7 The diagram shows a comparison of the azimuth profile in the range-Doppler domain before and after direct wave suppression, demonstrating that the direct wave was effectively suppressed. The suppression ratio before and after suppression is:

[0077]

[0078] in, To suppress direct wave interference, the power of the direct wave before it arrives is reduced. The suppression ratio is calculated in this experiment to obtain the residual direct wave power after direct wave interference suppression. It is 32.1dB.

[0079] Similarity is a performance indicator for interference cancellation, and the similarity function is:

[0080]

[0081] in, For direct wave interference signals, It expands a matrix into vectors column by column. It is the Frobenius norm.

[0082] Calculate the similarity between interference and echo after interference cancellation. Under different INR conditions, similarity is as follows: Figure 8 As shown, with the increase of INR, the similarity between the interference and the echo after interference cancellation decreases. The INR gradually decreases. When the INR is large, the similarity is small, and the interference component in the echo after interference cancellation is small. The larger the INR, the better the extraction performance of the PCA algorithm, and the CLEAN algorithm can use the extracted principal components as a reference to eliminate interference. Simulation results show that the interference cancellation performance of the method proposed in this invention improves with the increase of INR.

[0083] This invention effectively suppresses strong direct wave interference in the reflected echo, making targets with lower signal power visible, thus laying the foundation for subsequent target detection and tracking.

Claims

1. A method for suppressing direct waves from a CLEAN single-channel external radiation source bistatic radar based on principal component analysis, characterized in that: Includes the following steps: Step 1: Read in multiple consecutive frames of single-channel echo data, transmitter position and velocity parameters, receiver position and velocity parameters, and radar system parameters; Step 2: Perform range pulse compression, range migration compensation, and Doppler phase compensation on each frame of signal to complete direct wave synchronization; Step 3: After the range compression and compensation for the range error Doppler phase error in Step 2, perform azimuth-directed fast Fourier transform on the echo signals of each frame; further, perform cumulative averaging on the two-dimensional compressed first (k-1) frame signals in the range Doppler domain. Step 4: Perform PCA principal component analysis on the cumulative averaged echo signal, and select the direct wave corresponding to the principal component with the maximum energy as the reconstructed direct wave signal; Step 5: Using the direct wave reference signal after two-dimensional compression and reconstruction as the reference signal for the CLEAN direct wave suppression algorithm, calculate the CLEAN correlation coefficient CR; further, use the calculated correlation coefficient... Multiplying the signal by the reconstructed reference signal yields the direct wave interference component to be removed by the CLEAN algorithm. Further subtraction removes the interference, ultimately yielding the CLEAN-processed reflected wave signal.

2. The method for suppressing direct waves from a CLEAN single-channel external radiation source bistatic radar based on principal component analysis as described in claim 1, characterized in that: In step one, the radar system parameters read in include PRF, carrier frequency, Kr, sampling rate, pulse width, azimuth accumulation time, and bandwidth.

3. The method for suppressing direct waves from a CLEAN single-channel external radiation source bistatic radar based on principal component analysis as described in claim 1, characterized in that: In step one, the echo data read in for each frame of the reflection channel is as follows: In the formula, Indicates the direct amplitude value; Indicates the amplitude of the reflected wave; Indicates the reflected echo signal. Indicates a direct wave signal; Indicates the target echo signal; It is Gaussian white noise. , For the real and imaginary parts of the complex Gaussian noise; Indicates distance and time; Indicates location and time. The speed of light; , and These represent the distances from the transmitter to the receiver, from the transmitter to the target, and from the target to the receiver over time, respectively.

4. The method for suppressing direct waves from a CLEAN single-channel external radiation source bistatic radar based on principal component analysis as described in claim 1, characterized in that: The specific method for step two is as follows: a. Based on the ephemeris parameters of the radiation source platform and the motion information of the receiving platform, an equivalent distance model is used to characterize the relative distance variation between the radiation source and the receiving platform, including the equivalent distance. Doppler center frequency Doppler frequency modulation ; A Taylor expansion of the equivalent distance model is performed, and the second-order result is used for approximation. The expression for the equivalent distance model is: exist Taylor's expansion: in, Indicates the center frequency of the Doppler signal; Indicates Doppler frequency modulation. For wavelength, This indicates the relative equivalent flight speed between the radiation source platform and the receiver platform; Indicates the equivalent oblique angle of view; Doppler center frequency. Doppler frequency modulation The relationship between the platform's motion characteristics and the following formula is shown: In the formula, This represents the relative position vector between the radiation source platform and the receiver platform; This represents the relative velocity vector between the radiation source platform and the receiver platform; b. Calculate the distance migration factor and phase Doppler factor , used for distance and phase compensation; In the formula, and For distance delay error and Doppler phase error: ； In the formula, Represents the complex unit, used to construct phase compensation factors in complex exponential form; c. Perform range-directed matched filtering, range migration compensation, and Doppler phase compensation on each frame of the single-channel signal. in, The frequency domain response of the matched filter is used for range compression. For the range-directed frequency variable, It is a distance-to-fast Fourier transform. It is the inverse fast Fourier transform of the distance direction.

5. The method for suppressing direct waves from a CLEAN single-channel external radiation source bistatic radar based on principal component analysis as described in claim 1, characterized in that: In step four, the PCA principal component analysis method is as follows: The accumulated average echo signal Represented in matrix form as ;in This represents the number of sampling points in the azimuth-time dimension. This represents the number of sampling points over time. First, for the matrix To remove the mean: ;in, The mean of the matrix; Subsequently, after removing the mean... Performing singular value decomposition (SVD) yields: in, It is an orthogonal matrix, and its column vectors are respectively the pairwise matrices. and eigenvectors; Let be a diagonal matrix composed of singular values, with each element being a singular value and arranged in ascending order. , It is the number of singular values; Principal component number; The above Main component and corresponding characteristic signals The weighted sum can be expressed as: in, .

6. The method for suppressing direct waves from a CLEAN single-channel external radiation source bistatic radar based on principal component analysis as described in claim 1, characterized in that: In step five, the CLEAN correlation coefficient (CR) is calculated as follows: The direct wave reference signal reconstructed by principal component analysis Combine the current frame echo signal to be processed, after undergoing range-azimuth two-dimensional compression, range migration, and phase compensation as described in steps 2 and 3 above. Calculate the weighted correlation coefficient of single-channel data : In the formula, express Complex conjugate; Indicates the azimuth direction of the reconstructed direct wave signal. The distance corresponding to each pulse To signal data; Indicates the azimuth direction of the echo signal. The range signal data corresponding to each pulse; The calculated correlation coefficient Multiplying the signal by the reconstructed reference signal yields the direct wave interference component to be removed by the CLEAN algorithm. Further subtraction removes the interference, ultimately resulting in the CLEAN-processed reflected wave signal. , represented as: 。

Images