A Deconvolution Method Based on Horizontal Linear Array of MVDR
By using the MVDR deconvolution method based on a horizontal linear array, the problems of wide main lobe and high sidelobe in beamforming are solved, achieving signal detection with narrower beam and lower sidelobe, thus improving underwater signal resolution and anti-interference capability.
Patent Information
- Application Number
- CN202411045338.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-31
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2044-07-31
Smart Images

Figure CN119149884B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of underwater acoustic signal processing, and more specifically to an MVDR deconvolution method based on a horizontal linear array. Background Technology
[0002] Horizontal linear arrays are typically located in underwater environments and are used to detect and separate weak signals, and to estimate their direction of arrival from strong interference sources and environmental noise. Conventional beamforming (CBF) algorithms are highly robust, but their wide main lobe makes it difficult to detect multiple targets within the same beam, and their high sidelobe level makes it difficult to detect weak signals under strong interference.
[0003] High-resolution beamforming based on the inversion of the signal covariance matrix, such as the Minimum Variance Distortionless Response (MVDR) algorithm, produces narrow beamwidths and low sidelobes, but is sensitive to signal mismatch. Assuming the signal is stationary, the sample covariance matrix requires a large number of snapshots to ensure fast convergence and minimal power loss in the beam output. For large-aperture arrays employing multiple elements, high-resolution algorithms typically perform poorly when the number of snapshots is small. Summary of the Invention
[0004] The purpose of this invention is to provide a MVDR deconvolution method based on a horizontal linear array to solve the problems of wide main lobe and high side lobe order in existing beamforming methods in the background art.
[0005] To achieve the above objectives, the present invention provides the following technical solution:
[0006] A MVDR deconvolution method based on a horizontal linear array, the method comprising the following steps:
[0007] Step 1: Set the parameters of the horizontal uniform linear array and the sound source signal parameters;
[0008] Step 2: Construct the sound source signal s(t) and noise n(t), obtain the array data received by the horizontal linear array through simulation calculation, and calculate the array data covariance matrix;
[0009] Step 3: Perform MVDR beamforming on the array data received by the horizontal linear array to obtain the MVDR output power;
[0010] Step 4: Solve for the beam pattern of the horizontal linear array;
[0011] Step 5: Obtain the original orientation distribution of the signal by deconvolution using the RL algorithm.
[0012] Preferably, in step 1, the element spacing of the horizontal linear array is set. λ is the wavelength, which sets the length, frequency, orientation, and signal-to-noise ratio of the sound source signal.
[0013] Preferably, in step 2, the signal s(t-τ) arriving at the horizontal linear array under far-field conditions is... m Characterized as Calculate the received data x(t) from the array as x(t) = a(θ0)s(t) + n(t). a(θ0) is the array steering vector, θ0 is the location of the sound source, j is an imaginary number, ω is the angular frequency, and T... M The time delay is the time delay between the sound source signal received by the Mth array element and the first array element.
[0014] Preferably, in step 3, the MVDR output power B MVDR (sinθ)=w H Rw, where w is the MVDR beamforming weighting vector. a(θ0) is the array steering vector, R is the array data covariance matrix, and H is the transpose.
[0015] Preferably, in step 4, the beam pattern of the horizontal linear array is as follows: Where c is the speed of sound in water, M is the number of elements in the horizontal linear array, λ is the wavelength, and θ is the horizontal angle.
[0016] Preferably, in step 5, based on the MVDR output power B MVDR (sinθ) and beam pattern B p (sinθ), obtained through the RL algorithm The energy distribution function of the sound source is obtained by solving, where, This refers to the horizontal scanning angle.
[0017] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0018] Compared with existing conventional methods, the beamwidth generated by the deconvolution beamforming in this invention is narrow, the sidelobe level is low, and the calculation is simple, with a computational load comparable to the CBF algorithm. Attached Figure Description
[0019] Figure 1 This is a flowchart of the MVDR deconvolution method based on a horizontal linear array according to the present invention.
[0020] Figure 2 This is the beam pattern of the horizontal linear array at 0° azimuth.
[0021] Figure 3 The orientation spectrum is for MVDR and the method of this invention.
[0022] Figure 4 for Figure 3 A magnified view of a portion of the image. Detailed Implementation
[0023] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.
[0024] Reference Figure 1 As shown, this MVDR deconvolution method based on a horizontal linear array includes the following steps:
[0025] Step 1: Set the relevant parameters for the horizontal uniform linear array and the sound source signal.
[0026] In this embodiment, the number of elements in the horizontal linear array is set to 90, the element spacing is set to 1m, the signal length of sound sources 1 and 2 is set to 1000, the frequency is set to 1500Hz, the signal-to-noise ratio is set to 15dB, and the azimuths are set to -35° and -45° respectively. The underwater sound speed is set to 1500m / s.
[0027] Step 2: Construct the sound source signal s(t) and noise n(t), obtain the array data received by the horizontal linear array through simulation calculation, and calculate the array data covariance matrix.
[0028] In step 2 of this invention, under far-field conditions, the sound source signal s(t-τα) reaching the acoustic array is approximately expressed as: That is to say in, Let represent the time delay between the sound source signal received by the α-th array element and the 1st array element, c be the underwater speed of sound, θ be the horizontal angle, j be the imaginary number, ω be the angular frequency, ω = 2πf, and f be the frequency; then, the received data x(t) of the horizontal array is calculated using the sound source signal arriving at the acoustic array:
[0029]
[0030] in Let θ0 be the orientation vector of the array, and θ0 be the location of the sound source; from this, the array data covariance matrix R = x(t)x is obtained. H (t), where H is the transpose.
[0031] Step 3: Perform MVDR beamforming on the received array data to obtain the MVDR output power.
[0032] In step 3 of this invention, the calculation formula for MVDR beamforming is shown below, and the MVDR output power B is calculated. MVDR ;
[0033] B MVDR (sinθ)=wH Rw
[0034] In the formula, w is the MVDR beamforming weighting vector. R is the array data covariance matrix obtained in step 2. θ is the array guide vector, θ is the horizontal angle, and H is the transpose.
[0035] In this embodiment, when the array element spacing is half a wavelength, the array steering vector a(θ) can be expressed as the following formula.
[0036]
[0037] Step 4: Calculate the beam pattern of the horizontal linear array using the following formula:
[0038]
[0039] Where c is the speed of sound in water, M is the number of elements in the horizontal linear array, λ is the wavelength, and θ is the horizontal angle.
[0040] In this embodiment, the beam pattern of the horizontal linear array is obtained by calculation as follows: Figure 2 As shown.
[0041] Step 5: Obtain the original directional distribution of the sound source signal by deconvolution using the RL algorithm.
[0042] Deconvolution is an ill-posed problem. For the special case where both the output power and the beam pattern are positive definite functions, the solution is unique and can be solved using the Richardson-Lucy (RL) algorithm.
[0043] In this invention, sinθ is used as a variable to ensure that the beam pattern is translated invariantly at the steering angle, so that the RL algorithm can be applied to beamforming.
[0044] In step 5 of this invention, the RL algorithm is used to... During deconvolution, the MVDR output power B MVDR (sinθ) and beam pattern B p Substituting (sinθ) into the equation, the energy distribution function of the sound source can be obtained. The sound source energy distribution function consists of multiple Dirac pulses representing the direction of the sound source:
[0045]
[0046] In the formula, For the horizontal scanning angle, S i Let θ be the signal strength, δ be the Dirac impulse function, and sinθ be the signal intensity. iK represents the actual signal angle and the total number of actual targets.
[0047] Combination Figure 3 and 4 It can be seen that, compared with MVDR, the method of the present invention can achieve a narrower beamwidth and a lower sidelobe.
[0048] This invention deconvolvees the output power of MVDRs, which can restore the original signal distribution, resulting in a narrower beamwidth and lower sidelobes. Deconvolution is often used to solve deblurring problems in image processing, and its effect is excellent when both input and output are positive numbers.
[0049] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any way. Other variations and modifications may be made without departing from the technical solutions described in the claims.
Claims
1. A MVDR deconvolution method based on a horizontal linear array, characterized in that, The method includes the following steps: Step 1: Set the parameters of the horizontal uniform linear array and the sound source signal parameters; Step 2, construct the sound source signal With noise The array data received by the horizontal linear array is obtained through simulation calculation, and the array data covariance matrix is calculated. Step 3: Perform MVDR beamforming on the array data received by the horizontal linear array to obtain the MVDR output power; in step 3, the MVDR output power... ,in For MVDR beamforming, weighting vectors are generated. , For array guide vector, The covariance matrix of the array data. For transpose; Step 4, solve for the beam pattern of the horizontal linear array; in step 4, the beam pattern of the horizontal linear array is... ,in The speed of sound in water, The number of elements in the horizontal linear array. For wavelength, The horizontal angle is defined as follows: the beam pattern is a theoretical beam pattern determined solely by the sound speed in water, the number of array elements, the wavelength, and the horizontal angle, and it remains fixed during the deconvolution iteration process without changing with the array data covariance matrix. Step 5: Based on MVDR output power and beam pattern Using RL algorithm The energy distribution function of the sound source is obtained by solving, where, The horizontal scanning angle is based on the theoretical beam pattern that is independent of the array received data. As a deconvolution kernel function, The variables are deconvolved to ensure that the beam pattern, which serves as the kernel function, is translationally invariant at the steering angle, thus enabling the RL algorithm to be directly applied to beamforming. The source energy distribution function obtained by deconvolution consists of multiple Dirac pulses representing the direction of the source.
2. The MVDR deconvolution method based on a horizontal linear array as described in claim 1, characterized in that, In step 1, the element spacing of the horizontal linear array is set. , The wavelength is used to set the length, frequency, orientation, and signal-to-noise ratio of the sound source signal.
3. The MVDR deconvolution method based on a horizontal linear array as described in claim 1, characterized in that, In step 2, the signal arriving at the horizontal linear array under far-field conditions... Characterization Receive data for computing array , , For array guide vector, The location of the sound source. It is an imaginary number. Angular frequency, For the first The time delay between each array element and the sound source signal received by the first array element.
Citation Information
Patent Citations
Deconvolution beamforming three-dimensional acoustic imaging method
CN109375227A
