A coherent interference suppression method based on virtual source loading and null broadening

Abstract

The application discloses a coherent interference suppression method based on virtual source loading and null broadening, which comprises the following steps: 1, array structure and array signal modeling; 2, calculating the covariance matrix of the array signal; 3, estimating the coherent interference direction of the array signal and calculating the covariance matrix of the virtual interference of the array signal; 4, performing coning processing on the covariance matrix of the virtual interference to broaden the virtual interference angle; 5, solving the covariance matrix of the array signal after the virtual interference with the broadened angle is loaded and solving the MVDR weight vector; 6, obtaining the output by weighting the array signal with the obtained weight vector; and the application can realize the robust suppression of coherent interference.

Description

Technical Field

[0001] This invention belongs to the field of adaptive beamforming, specifically relating to a coherent interference suppression method based on virtual source loading and null broadening. Background Technology

[0002] Adaptive beamforming is a technique that weights array signals to simultaneously suppress interfering signals while allowing the desired signal to pass. This technique is widely used in radar, communications, and sonar. Traditional adaptive beamforming methods, such as MVDR, cannot form nulls in the direction of coherent interference, and may even cause signal cancellation. When multipath propagation or "smart" interference exists, the interference becomes coherent with the desired signal.

[0003] Most algorithms for resolving coherent interference utilize spatial smoothing. This method divides the array into several subarrays and resolves coherence by weighting the covariance matrices of these subarrays. However, the improvement in coherent interference reduction through spatial smoothing depends on the spatial separation of the subarrays. Therefore, spatial smoothing algorithms require a large number of subarrays for coherence resolution, which sacrifices the array aperture. Besides spatial smoothing algorithms, there are other algorithms such as MCMV, SPT, CTMV, and quartic constraint algorithms. These methods add constraints in the direction of coherent interference, thus requiring angular information about the interference. However, the angular estimation of coherent interference often contains errors, leading to performance degradation in these algorithms. Furthermore, there are blind beamforming algorithms based on higher-order cumulants. These methods have high computational complexity and are not suitable for real-time adaptive beamforming. Summary of the Invention

[0004] In view of this, the present invention provides a coherent interference suppression method based on virtual source loading and zero trap broadening, which can achieve robust suppression of coherent interference.

[0005] The technical solution for implementing the present invention is as follows:

[0006] A coherent interference suppression method based on virtual source loading and zero-depression broadening includes the following steps:

[0007] Step 1: Modeling the array structure and array signals;

[0008] Step 2: Calculate the covariance matrix of the array signals;

[0009] Step 3: Estimate the direction of coherent interference of the array signal and calculate the covariance matrix of the virtual interference of the array signal;

[0010] Step 4: Taper the covariance matrix of the virtual interference to broaden the virtual interference angle;

[0011] Step 5: Solve for the covariance matrix of the virtual interference after the expansion angle is applied to the array signal and solve for the MVDR weight vector;

[0012] Step 6: Use the obtained weight vector to weight the array signal to obtain the output.

[0013] Furthermore, in step three, the direction of coherent interference is estimated using an arbitrary DOA estimation algorithm to obtain the angle of coherent interference.

[0014] Furthermore, the covariance matrix of the array signal after loading virtual interference in step five is:

[0015]

[0016] Among them, R x Let be the covariance matrix of the array signal, I be the total number of interferences, J be the number of coherent interferences, and l be the number of interferences. i Let a(θ) be the loading amount of the virtual source for the i-th virtual interference. i ) represents the guiding vector, and H represents the conjugate transpose of the matrix.

[0017] Beneficial effects:

[0018] (1) The method of the present invention can achieve robust suppression of coherent interference and has certain application value in the field of adaptive beamforming.

[0019] (2) The non-coherent virtual interference loading proposed in steps four and five of this invention can effectively restore the rank of the signal covariance matrix, so that the beamformer can form nulls at the coherent interference, thereby achieving the suppression of coherent interference. Attached Figure Description

[0020] Figure 1 This is a flowchart of the algorithm of the present invention.

[0021] Figure 2 This is the array structure used in this invention.

[0022] Figure 3 This is a schematic diagram of the virtual interference source loading of the present invention.

[0023] Figure 4 This is a comparison of the orientation patterns of the method of this invention with those of other algorithms.

[0024] Figure 5 This is a comparison of the output signal-to-interference-plus-noise ratio (SINR) of the method of this invention with that of other algorithms.

[0025] Figure 6 This invention compares the output signal-to-interference-plus-noise ratio (SINR) of the method described in this invention with that of other algorithms under different DOA errors. Detailed Implementation

[0026] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0027] This invention provides a coherent interference suppression method based on virtual source loading and zero-traps broadening, the process of which is as follows: Figure 1 As shown.

[0028] Step 1: Modeling the array structure and array signals;

[0029] The array structure in step 1 is a uniform linear array with a spacing of d and N array elements, operating at a wavelength of λ. The array receives signals including the desired signal, coherent interference, and incoherent interference, with a total of I interferences, of which J are coherent interferences. The desired signal angle is θ0, and the coherent interference angle is {θ0}. I-J+1 ,…,θ I The angles of the incoherent interference are {θ1, θ2, ..., θ}. I-J The interference angles are all unknown, and the array model is as follows: Figure 2 As shown, the array receives the signal as follows:

[0030]

[0031] in

[0032] A = [a(θ0), a(θ1), ..., a(θ)] I (2)

[0033]

[0034] s(t)=[s0(t),s1(t),…,s I (t)] (4)

[0035] n(t) = [n0(t), n1(t), ..., n N-1 (t)] T (5)

[0036] T represents the transpose, λ is the wavelength, and s i (t) is the complex envelope of the signal, n k (t){k=0,1,…,N-1} is Gaussian noise and n i (t) and n j (t) When i≠j, it is unrelated, and J represents the number of coherent interferences among the I interferences.

[0037] Where s j (t)=k j s0(t){j=I-J+1,…,I},k j Since is a complex constant, the above equation can be expressed as

[0038]

[0039] Step 2: Calculate the covariance matrix of the array signals;

[0040] Step 2 calculates the covariance matrix as follows:

[0041]

[0042] in The power of the desired signal. Represents the interference power, σ 2 Representing noise power, R is calculated using a finite number of snapshots in practical applications. x That's sufficient; there's no need to calculate the power. This explanation is provided for easier derivation and proof later.

[0043] Step 3: Estimate the direction of coherent interference and calculate the covariance matrix of the virtual interference;

[0044] Step 3, estimating the direction of coherent interference, can be achieved using any DOA estimation algorithm, such as spatial smoothing spectrum estimation, to obtain the angle {θ} of the coherent interference. I-J+1 ,…,θ I The covariance matrix of the virtual interference is first derived as follows:

[0045] Let J be the number of incoherent virtual interferences, s′. i (t), loaded into the array output to obtain

[0046]

[0047] Where f i (t)=s j (t)+s′ j (t), and f j (t)≠ks0(t), meaning that adding virtual interference removes the linear relationship between the interference and the desired signal. Furthermore, the signal subspace of the covariance matrix after adding virtual interference is span{a(θ0),a(θ1),a(θ2),…,a(θ...} I-J ),…,a(θ I That is, the rank of the covariance matrix is ​​I+1, and the signal subspace of the covariance matrix before loading is span{a}. s ,a(θ1),a(θ2),…,a(θ I-J The rank of the covariance matrix is ​​I-J+1. The covariance matrix after loading virtual interference is...

[0048]

[0049] Because of s′ j (t) is uncorrelated with x(t), therefore the covariance matrix after adding virtual disturbance is:

[0050]

[0051] Where R v The virtual interference covariance matrix is ​​the covariance matrix of the virtual interference described in step 3. i The amount of virtual signal source loading is used. Therefore, the covariance matrix after loading virtual interference is the original covariance matrix plus the covariance matrix of virtual interference. Here, there is no need to construct incoherent interference signal. Solving the loaded covariance matrix only requires adding the covariance matrix of virtual incoherent interference to the original covariance matrix as shown in formula (10). This solution method will be used to solve the covariance matrix in step 5.

[0052] Step 4: Taper the virtual interference covariance matrix to broaden the virtual interference angle;

[0053] Step 4, widening the interference angle, aims to mitigate the performance degradation caused by DOA estimation errors. When DOA errors exist, inaccurate null directions can still lead to the reception of coherent interference, resulting in the cancellation of the desired signal and severely impacting beamforming performance. Furthermore, when the source is coherent, poor DOA estimation performance inevitably leads to errors. Therefore, widening the virtual interference angle is necessary to increase the algorithm's robustness, such as... Figure 3 As shown.

[0054] Assume the DOA estimation error is Δθ i It satisfies a uniform distribution within a certain interval (-δ, δ), (where Δθ) i ,δ,θ i (All are normalized angles), then Δθ i The probability density function is

[0055]

[0056] We can obtain sinθ i Approximately obeys (sinθ) i -δcosθ i sinθ i +δcosθ i A uniform distribution on ) is denoted as sinθ. i For v i The covariance matrix of the broadened virtual interference can be obtained as follows:

[0057]

[0058] R′ v The (k,l)th element is

[0059]

[0060] Therefore R′ v It has the following form

[0061]

[0062] in Represents element-wise multiplication, T(θ) i The (k,l)th element of ) is

[0063]

[0064] T(θ i The function of the matrix is ​​to broaden the width of the virtual interference. Observing formula (14), it can be found that the broadened virtual interference covariance matrix is ​​the sum of the terms of the original virtual interference covariance matrix and the matrix T(θ). i The element product of ).

[0065] Step 5: Solve for the covariance matrix of the virtual interference after it is applied to the array signal and solve for the MVDR weight vector;

[0066] Step 5 involves solving for the loaded covariance matrix by adding the broadened virtual interference covariance matrix described in Step 4 to the covariance matrix obtained in Step 2.

[0067]

[0068] According to the Minimum Variance Distortionless Ratio (MVDR) criterion, the weight vector can be obtained as follows:

[0069]

[0070] Step 6: Use the obtained weight vector to weight the array signal to obtain the output.

[0071] The weighted array signal obtained in step 6 is the complex conjugate multiplication array signal of the weight vector obtained in step 5, which is the final output of beamforming.

[0072]

[0073] s(t) is the output of beamforming, where interference has been suppressed.

[0074] The experimental scenarios are shown in the table below.

[0075]

[0076]

[0077] from Figure 4As can be seen, all methods except MVDR and CTMV can form nulls at incoherent interference locations. MVDR forms a peak in the direction of coherent interference, receiving the coherent interference and causing the desired signal to cancel out. Spatial smoothing algorithms can suppress coherent interference to some extent and eliminate the coherence between coherent interference and the desired signal, but they cannot form deep nulls and will still receive some of the coherent interference. Due to errors in DOA estimation, the MCMV method forms nulls at incorrect angles, causing beam distortion. The CTMV algorithm's radiation pattern cannot form nulls in the direction of interference, but instead receives coherent interference similar to the MVDR algorithm. Only the proposed method can effectively form nulls at coherent interference locations.

[0078] from Figure 5 It can be seen that the proposed method has the highest output signal-to-interference-plus-noise ratio.

[0079] from Figure 6 It can be seen that since the spatial smoothing algorithm and the MVDR algorithm do not require prior information about the DOA, the DOA estimation error has no impact on the spatial smoothing algorithm and the MVDR algorithm. When there is no angle estimation error, the proposed algorithm's output signal-to-interference-plus-noise ratio is close to that of the MCMV algorithm but lower than that of the CTMV algorithm. When there is a certain error in the DOA estimation, the proposed algorithm has the highest output signal-to-interference-plus-noise ratio.

[0080] In summary, the method described in this invention provides a novel approach for suppressing coherent interference in the field of adaptive beamforming.

[0081] In summary, the above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for suppressing coherent interference based on virtual source loading and null-traps broadening, characterized in that, Includes the following steps: Step 1: Modeling the array structure and array signals; Step 2: Calculate the covariance matrix of the array signals; Step 3: Estimate the coherent interference direction of the array signal and calculate the covariance matrix of the virtual interference of the array signal; whereby the covariance matrix of the virtual interference is... in, Let covariance be the array signal. The total number of interferences. The number of coherent interferences. Let be the loading amount of the virtual source for the i-th virtual interference. H represents the guiding vector, and H represents the conjugate transpose of the matrix; Step 4: Taper the covariance matrix of the virtual interference to broaden the virtual interference angle; Step 5: Solve for the covariance matrix of the array signal after the virtual interference is applied by the widened angle, and solve for the MVDR weight vector; whereby the covariance matrix of the array signal after the virtual interference is applied by the widened angle is: in, Represents element-wise multiplication; The The elements are Normalized estimation error for DOA The distribution range, i.e. In the interval Uniformly distributed within; Step 6: Use the obtained weight vector to weight the array signal to obtain the output.

2. The coherent interference suppression method based on virtual source loading and null broadening as described in claim 1, characterized in that, In step three, the direction of coherent interference is estimated using an arbitrary DOA estimation algorithm to obtain the angle of coherent interference.

Images