Fast broadband DOA (direction of arrival) estimation method for signal frequency band heterogeneity based on joint space-frequency base selection

Abstract

The invention provides a fast broadband DOA (direction of arrival) estimation method for signal frequency band isomerism based on joint space-frequency base selection, which comprises the following steps: a space-frequency base set is obtained based on a set angle range and set frequency band division, and each space-frequency base comprises an angle and a section of frequency sub-band; calculating a precision parameter estimation value and a discrimination parameter of each space-frequency base based on the array receiving signal so as to calculate a corresponding likelihood function added value; taking the space-frequency base corresponding to the maximum likelihood function added value as a target space-frequency base, and updating the space-frequency base activation set and the inference parameter based on the discrimination parameter of the target space-frequency base; if the updating requirement is met, the precision parameter estimation value and the discrimination parameter of each space-frequency base are updated based on the inference parameter until the updating requirement is not met so as to obtain a target space-frequency base activation set; combining the space-frequency bases in the target space-frequency base activation set to obtain DOA preliminary estimation results and frequency band information of the plurality of broadband signals; and a final DOA estimation result is obtained by using a maximum likelihood criterion.

Description

Technical Field

[0001] The present invention relates to the technical field of direction of arrival (DOA) estimation, and in particular to a fast broadband DOA estimation method for signal frequency band heterogeneity based on joint space-frequency basis selection. Background Art

[0002] With the continuous development of array signal processing technology, Direction of Arrival (DOA) estimation for broadband signals has gradually become a key technology in the field of array signal processing. Due to its important applications in practical needs such as target detection and resource exploration, it has been widely used in radar, sonar, communications and other fields.

[0003] Traditional DOA estimation techniques for broadband signals primarily include beamforming and subspace methods. Both methods are widely used in various fields, each with its own advantages and disadvantages. Beamforming methods are simple to implement and have low computational complexity, making them the most widely used methods in engineering applications. However, their spatial resolution is limited by the array aperture, making it difficult to achieve high-resolution DOA estimation. Subspace methods can overcome the problem of insufficient spatial resolution caused by array aperture limitations, but they often require prior knowledge of the approximate number of signal sources and often suffer from performance degradation under conditions of low signal-to-noise ratio and low snapshot count.

[0004] In recent years, wideband DOA estimation techniques based on sparse representation have garnered widespread attention due to their advantages, including robustness to noise, high resolution, and the absence of prior knowledge of the number of signal sources. Sparse Bayesian learning (SBL)-based wideband DOA estimation is a representative example. However, the traditional SBL framework requires an inverse matrix transformation for each solution, resulting in excessive time complexity. Therefore, designing corresponding real-time DOA estimation methods is a promising approach.

[0005] More importantly, all of the aforementioned wideband DOA estimation methods assume that the target signals share the same frequency band, an assumption that is often difficult to satisfy. In reality, different wideband signals often occupy different frequency bands, which often leads to a significant degradation in DOA estimation performance for algorithms that ignore this situation. Summary of the Invention

[0006] The present invention aims to solve one of the technical problems in the related art at least to a certain extent.

[0007] To this end, the first object of the present invention is to propose a fast broadband DOA estimation method for signal frequency band heterogeneity based on joint space-frequency basis selection to achieve more accurate DOA estimation.

[0008] The second object of the present invention is to propose a fast broadband DOA estimation system for signal frequency band heterogeneity based on joint space-frequency basis selection.

[0009] A third object of the present invention is to provide an electronic device.

[0010] A fourth object of the present invention is to provide a computer-readable storage medium.

[0011] To achieve the above objectives, the first aspect of the present invention proposes a fast broadband DOA estimation method for signal frequency band heterogeneity based on joint space-frequency basis selection, comprising:

[0012] A space-frequency basis set is obtained based on the set angle range and the set frequency band, where each space-frequency basis in the space-frequency basis set includes an angle and a frequency sub-band;

[0013] Calculating the precision parameter estimation value and discrimination parameter of each space-frequency basis based on the array received signal, and calculating the likelihood function increase value of each space-frequency basis based on the precision parameter estimation value; taking the space-frequency basis corresponding to the maximum likelihood function increase value as the target space-frequency basis, and updating the space-frequency basis activation set and the corresponding inference parameters based on the discrimination parameter of the target space-frequency basis;

[0014] If the update requirement is met, the precision parameter estimation value and the discrimination parameter of each space-frequency basis are updated based on the inferred parameter, and the space-frequency basis activation set and the inferred parameter are further updated until the update requirement is no longer met, so as to obtain a target space-frequency basis activation set;

[0015] Combining the space-frequency bases in the target space-frequency base activation set to obtain preliminary DOA estimation results and frequency band information of multiple wideband signals;

[0016] A maximum likelihood criterion is adopted to obtain a final DOA estimation result of the corresponding wideband signal based on the preliminary DOA estimation result and frequency band information of each wideband signal.

[0017] In the method of the first aspect of the present invention, the increase in the likelihood function of each space-frequency basis based on the precision parameter estimate includes: for any space-frequency basis, calculating the corresponding likelihood function based on the precision parameter estimate; if the discrimination parameter of the space-frequency basis is greater than zero and the space-frequency basis belongs to the space-frequency basis activation set, the increase in the likelihood function is the difference between the likelihood function at the current update number and the historical likelihood function; if the discrimination parameter of the space-frequency basis is greater than zero and the space-frequency basis does not belong to the space-frequency basis activation set, the increase in the likelihood function is the likelihood function at the current update number; if the discrimination parameter of the space-frequency basis is less than zero and the space-frequency basis belongs to the space-frequency basis activation set, the increase in the likelihood function is the inverse of the likelihood function at the current update number.

[0018] In the method of the first aspect of the present invention, the updating of the space-frequency base activation set based on the discrimination parameter of the target space-frequency base includes: if the discrimination parameter of the target space-frequency base is greater than zero and the target space-frequency base belongs to the space-frequency base activation set, re-estimating the precision parameter estimation value of the target space-frequency base; if the discrimination parameter of the target space-frequency base is greater than zero and the target space-frequency base does not belong to the space-frequency base activation set, updating the target space-frequency base to the space-frequency base activation set; if the discrimination parameter of the target space-frequency base is less than zero and the target space-frequency base belongs to the space-frequency base activation set, deleting the target space-frequency base from the space-frequency base activation set.

[0019] In the method of the first aspect of the present invention, satisfying the update requirement means that the number of updates is equal to 1, or when the number of updates is not equal to 1, the convergence iteration error is greater than or equal to the error threshold and the number of updates is less than or equal to the maximum number of updates.

[0020] In the method of the first aspect of the present invention, the inference parameters include the covariance matrix and mean of the signal distribution of each activated space-frequency base in the space-frequency base activation set at different frequency points, and multiple defined variables corresponding to each space-frequency base in the space-frequency base set; each defined variable is determined based on the array steering vector and the array received signal.

[0021] In the method of the first aspect of the present invention, the maximum likelihood criterion is adopted to obtain the final DOA estimation result of the corresponding broadband signal based on the preliminary DOA estimation result and frequency band information of each broadband signal, including: for any broadband signal, determining the target angle range based on the preliminary DOA estimation result of the broadband signal, the angle step of the target angle range is smaller than the angle step of the set angle range; for each target angle in the target angle range of the broadband signal, obtaining the corresponding target likelihood function based on the frequency band information, and adopting the maximum likelihood criterion to obtain the final DOA estimation result of the broadband signal based on the target likelihood function of all target angles.

[0022] To achieve the above objectives, the second aspect of the present invention proposes a fast broadband DOA estimation system for signal frequency band heterogeneity based on joint space-frequency basis selection, comprising:

[0023] A division module, configured to obtain a space-frequency basis set based on a set angle range and a set frequency band, wherein each space-frequency basis in the space-frequency basis set includes an angle and a frequency sub-band;

[0024] an updating module, configured to calculate a precision parameter estimate and a discrimination parameter for each space-frequency basis based on the array received signal, and calculate a likelihood function increase value for each space-frequency basis based on the precision parameter estimate value; take the space-frequency basis corresponding to the maximum likelihood function increase value as a target space-frequency basis, and update a space-frequency basis activation set and corresponding inference parameters based on the discrimination parameter of the target space-frequency basis; if an update requirement is met, update the precision parameter estimate and the discrimination parameter of each space-frequency basis based on the inference parameter, and further update the space-frequency basis activation set and the inference parameters, until the update requirement is no longer met, thereby obtaining a target space-frequency basis activation set;

[0025] a preliminary estimation module, configured to combine the space-frequency bases in the target space-frequency base activation set to obtain preliminary DOA estimation results and frequency band information of multiple broadband signals;

[0026] The final estimation module is used to obtain the final DOA estimation result of the corresponding wideband signal based on the DOA preliminary estimation result and frequency band information of each wideband signal by adopting the maximum likelihood criterion.

[0027] In the system of the second aspect of the present invention, when the update module is used to update the space-frequency base activation set based on the discrimination parameter of the target space-frequency base, it is specifically used to: if the discrimination parameter of the target space-frequency base is greater than zero and the target space-frequency base belongs to the space-frequency base activation set, then re-estimate the precision parameter estimation value of the target space-frequency base; if the discrimination parameter of the target space-frequency base is greater than zero and the target space-frequency base does not belong to the space-frequency base activation set, then update the target space-frequency base to the space-frequency base activation set; if the discrimination parameter of the target space-frequency base is less than zero and the target space-frequency base belongs to the space-frequency base activation set, then delete the target space-frequency base from the space-frequency base activation set.

[0028] To achieve the above-mentioned purpose, the third aspect of the present invention proposes an electronic device, comprising: a processor, and a memory communicatively connected to the processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory to implement the method proposed in the first aspect of the present invention.

[0029] To achieve the above-mentioned purpose, the fourth aspect of the present invention proposes a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the method proposed in the first aspect of the present invention.

[0030] The present invention provides a fast broadband DOA estimation method, system, electronic device and storage medium for signal frequency band heterogeneity based on joint space-frequency basis selection. The method obtains a space-frequency basis set based on a set angle range and a set frequency band, and each space-frequency basis in the space-frequency basis set includes an angle and a frequency sub-band. The precision parameter estimation value and discrimination parameter of each space-frequency basis are calculated based on the array received signal, and the likelihood function increase value of each space-frequency basis is calculated based on the precision parameter estimation value. The space-frequency basis corresponding to the maximum likelihood function increase value is used as the target space-frequency basis, and the discrimination parameter of the target space-frequency basis is calculated based on the discrimination parameter of the target space-frequency basis. The parameters are updated by updating the space-frequency basis activation set and the corresponding inferred parameters. If the update requirements are met, the precision parameter estimates and discriminant parameters of each space-frequency basis are updated based on the inferred parameters. The space-frequency basis activation set and inferred parameters are then updated again until the update requirements are no longer met, thereby obtaining a target space-frequency basis activation set. The space-frequency basis in the target space-frequency basis activation set are combined to obtain preliminary DOA estimates and frequency band information for multiple wideband signals. Based on the preliminary DOA estimates and frequency band information for each wideband signal, a final DOA estimate for the corresponding wideband signal is obtained using the maximum likelihood criterion. In this case, a space-frequency basis set is partitioned and, based on the precision parameter estimates, discriminant parameters, likelihood function increase, and inferred parameters of each space-frequency basis, all space-frequency basis occupied by wideband signals are activated through iterative updating to obtain the target space-frequency basis activation set. By combining the space-frequency basis in the target space-frequency basis activation set, the frequency band information of each wideband signal is estimated. Finally, based on the preliminary DOA estimates and frequency band information, a final DOA estimate for the corresponding wideband signal is obtained using the maximum likelihood criterion. This achieves more accurate DOA estimation.

[0031] Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description which follows, or may be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0033] Figure 1 A flowchart of a fast broadband DOA estimation method for signal frequency band heterogeneity based on joint space-frequency basis selection provided by an embodiment of the present invention;

[0034] Figure 2 A schematic diagram of a specific process for updating the space-frequency base activation set provided by an embodiment of the present invention;

[0035] Figure 3 A block diagram of a fast broadband DOA estimation system for signal frequency band heterogeneity based on joint space-frequency basis selection provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0036] The following describes embodiments of the present invention in detail, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and are not to be construed as limiting the present invention.

[0037] The following describes a fast broadband DOA estimation method and system for signal frequency band heterogeneity based on joint space-frequency basis selection according to an embodiment of the present invention with reference to the accompanying drawings.

[0038] The embodiment of the present invention provides a fast broadband DOA estimation method for signal frequency band heterogeneity based on joint space-frequency basis selection to achieve more accurate DOA estimation.

[0039] Figure 1 A flowchart of a fast broadband DOA estimation method for signal frequency band heterogeneity based on joint space-frequency basis selection provided by an embodiment of the present invention.

[0040] like Figure 1 As shown, the fast broadband DOA estimation method for signal frequency band heterogeneity based on joint space-frequency basis selection includes the following steps:

[0041] Step S101 : obtaining a space-frequency basis set based on a set angle range and a set frequency band, wherein each space-frequency basis in the space-frequency basis set includes an angle and a frequency sub-band.

[0042] In step S101 , the set angle range may be the full angle range of the linear array (ie, 0°-180°), or may be an angle range of interest, such as 30°-90°.

[0043] In step S101 , the set frequency band may be the entire frequency band or the frequency band of interest.

[0044] In step S101, the set angle range is divided according to a certain angle step size to obtain an angle set. This angle set is also called a discrete angle grid point set. The discrete angle grid point set can be expressed as [θ1, θ2, ..., θ N ]. Where N is the number of discrete angle grid points. The mth discrete angle grid point in the discrete angle grid point set can be expressed as θ m m is the discrete angle grid point index, and m is 1,…,N.

[0045] In step S101, the set frequency band is divided according to a certain frequency step size to obtain multiple non-overlapping frequency sub-bands. All frequency sub-bands constitute a frequency sub-band set [Ψ1, Ψ2, ..., Ψ G] where G is the total number of frequency subbands. The g-th frequency subband in the frequency subband set can be expressed as Ψ g g is the frequency subband index. g takes the value of 1,…,G.

[0046] In step S101, a discrete angle grid point (i.e., an angle) is randomly combined with a frequency subband to obtain N*G space-frequency bases. All space-frequency bases constitute a space-frequency base set. Any space-frequency base can be expressed as {θ m ,Ψ g}. m=1,…,N, g=1,…,G.

[0047] Step S102: Calculate the precision parameter estimation value and discrimination parameter of each space-frequency basis based on the array received signal, and calculate the likelihood function increase value of each space-frequency basis based on the precision parameter estimation value; take the space-frequency basis corresponding to the maximum likelihood function increase value as the target space-frequency basis, update the space-frequency basis activation set based on the discrimination parameter of the target space-frequency basis, and update the corresponding inference parameters.

[0048] In step S102, the inferred parameters include the covariance matrix and mean of the signal distribution of each activated space-frequency basis in the space-frequency basis activation set at different frequency points, for example, ∑ f is the covariance matrix of the currently activated space-frequency basis at the fth frequency point, with a mean value of u f (t) is the mean value of the signal distribution of the activated space-frequency basis at the fth frequency point, t = 1, ..., T. The inference parameters also include multiple definition variables for each space-frequency basis in the space-frequency basis set; each definition variable is determined based on the array steering vector and the array received signal. The multiple definition variables include the first definition variable S f,m and second define the variable Q f,m (t).

[0049] In step S102, before each space-frequency base is activated, the corresponding precision parameter is an initial value. The initial value is, for example, infinite. m ,Ψ g} is not activated, α g,m =∞, m=1,…,N, g=1,…,G.

[0050] In step S102, the precision parameter estimation value and discrimination parameter of each space-frequency basis are calculated based on the array received signal, including: for any space-frequency basis, the precision parameter estimation value and discrimination parameter of the space-frequency basis are calculated based on the first intermediate variable and the second intermediate variable; wherein the first intermediate variable and the second intermediate variable are determined based on the first defined variable and the second defined variable of the corresponding space-frequency basis, and the first defined variable and the second defined variable of each space-frequency basis are calculated based on the array received signal, the array steering vector, and the array manifold matrix.

[0051] Take any space-frequency basis {θ m ,Ψ g} as an example, the space-frequency basis {θ m ,Ψ g The precision parameter estimates and discriminant parameters of} satisfy:

[0052]

[0053] Where, is the space-frequency basis {θ m ,Ψ g}The precision parameter estimate of δ g,m is the space-frequency basis {θ m ,Ψ g}discriminant parameter. F g is the frequency subband Ψ g The total number of frequency points included. T is the number of snapshots of the array receiving signal at a single frequency point. f is the frequency index. t is the snapshot number index. (·) H Represents the conjugate transpose of a matrix or the conjugate value of a number. g=1,…,G;m=1,…,N.s f,m is the first intermediate variable, q f,m (t) is the second intermediate variable.

[0054] Among them, the first intermediate variable s f,m and the second intermediate variable q f,m (t) Satisfy:

[0055]

[0056] Where, α g,m is the space-frequency basis {θ m ,Ψ g It should be noted that the space-frequency basis {θ m ,Ψ g} is not activated, the precision parameter α g,m Always equal to the initial value. Space frequency basis {θ m ,Ψ g} is activated, the accuracy parameter α g,m is equal to the estimated value of the precision parameter under the last update number. When the update number is 1, all space-frequency bases are not activated, so all space-frequency base precision parameters are initial values. If the initial value is infinite, formula (3) degenerates to s f,m =S f,m , formula (4) degenerates into q f,m (t) = Q f,m (t).

[0057] Among them, the first definition variable S f,mand second define the variable Q f,m (t) Satisfy:

[0058]

[0059] Where, Φ f,·m is the array steering vector associated with the mth discrete angle grid point at the fth frequency point. f (t) is the array received signal of the t-th snapshot at the f-th frequency point. is the array flow matrix at the fth frequency point associated with the angle of the currently activated space-frequency basis. α0 is the accuracy parameter of the noise. It can be empirically initialized as F is the total number of frequency points. var(·) is the vector variance operator. It should be noted that when the frequency subband Ψ where the fth frequency point is located g There is no activated empty frequency basis, that is and ∑ f is an empty matrix, then formula (5) degenerates into Formula (6) degenerates into

[0060] In step S102, the likelihood function increase value of each space-frequency basis is calculated based on the precision parameter estimate value, including: for any space-frequency basis, the corresponding likelihood function is calculated based on the precision parameter estimate value; if the discrimination parameter of the space-frequency basis is greater than zero and the space-frequency basis belongs to the space-frequency basis activation set, then the likelihood function increase value is the difference between the likelihood function under the current update number and the historical likelihood function, and the historical likelihood function is calculated based on the historical precision parameter estimate value after the space-frequency basis is activated, wherein the historical precision parameter estimate value is the precision parameter estimate value under the previous update number; if the discrimination parameter of the space-frequency basis is greater than zero and the space-frequency basis does not belong to the space-frequency basis activation set, then the likelihood function increase value is the likelihood function under the current update number; if the discrimination parameter of the space-frequency basis is less than zero and the space-frequency basis belongs to the space-frequency basis activation set, then the likelihood function increase value is the inverse of the likelihood function under the current update number.

[0061] Take any space-frequency basis {θ m ,Ψ g} as an example, based on the corresponding accuracy parameter α g,m Calculate the space-frequency basis {θ m ,Ψ g}likelihood function, the space-frequency basis {θ m ,Ψ g The likelihood function of} satisfies:

[0062]

[0063] It should be noted that, referring to formula (7), the accuracy parameter α g,m Replaced with precision parameter estimates When , the likelihood function under the current number of updates can be calculated The precision parameter α g,m Replaced with historically accurate parameter estimates The historical likelihood function can be calculated

[0064] If δ g,m >0 and {θ m ,Ψ g}∈Ω, then Where Ω is the space-frequency base activation set. The space-frequency base activation set Ω is defined as the set containing the space-frequency bases that have been activated. ΔL g,m Add value to the likelihood function. If δ g,m >0 and but If δ g,m <0 and {θ m ,Ψ g}∈Ω, then

[0065] In step S102, the maximum value is selected from the likelihood function increase values of each space-frequency basis, and the space-frequency basis corresponding to the maximum likelihood function increase value is used as the target space-frequency basis.

[0066] Figure 2 This is a schematic diagram of a specific process for updating the space-frequency base activation set provided by an embodiment of the present invention. Figure 2 As shown, in step S102, the space-frequency base activation set is updated based on the discrimination parameter of the target space-frequency base, including: if the discrimination parameter of the target space-frequency base is greater than zero and the target space-frequency base belongs to the space-frequency base activation set, the precision parameter estimation value of the target space-frequency base is re-estimated (step S1021); if the discrimination parameter of the target space-frequency base is greater than zero and the target space-frequency base does not belong to the space-frequency base activation set, the target space-frequency base is updated to the space-frequency base activation set (step S1022); if the discrimination parameter of the target space-frequency base is less than zero and the target space-frequency base belongs to the space-frequency base activation set, the target space-frequency base is deleted from the space-frequency base activation set (step S1023).

[0067] In step S102, after the space-frequency base activation set is updated, the corresponding inference parameters also need to be updated.

[0068] Step S103: If the update requirement is met, the precision parameter estimation value and the discrimination parameter of each space-frequency basis are updated based on the inferred parameters, and then the space-frequency basis activation set and the inferred parameters are updated until the update requirement is no longer met, so as to obtain the target space-frequency basis activation set.

[0069] In step S103, satisfying the update requirement means that the number of updates is equal to 1, or when the number of updates is not equal to 1, the convergence iteration error is greater than or equal to the error threshold and the number of updates is less than or equal to the maximum number of updates. err =|ΔL (iter) -ΔL (iter-1) | / |ΔL (iter) -ΔL (1) |. Among them, ΔL (iter) The increase value of the likelihood function of the target space-frequency basis at the iter-th iteration (i.e., the current number of updates). ΔL (iter-1) Increase the likelihood function of the target space-frequency basis for the iter-1th iteration by ΔL (1) is the value added to the likelihood function of the target space-frequency basis in the first iteration. It should be noted that the number of updates in this embodiment is also called the number of iterations.

[0070] In step S103, the precision parameter estimation value and discrimination parameter of each space-frequency basis are updated based on the inference parameters, including: updating the current update number, substituting the updated value of the inference parameter of the previous update number into equations (3) and (4) to update the first intermediate variable and the second intermediate variable; updating the precision parameter estimation value and discrimination parameter of each space-frequency basis based on the updated first intermediate variable and the updated second intermediate variable.

[0071] Specifically, the specific implementation steps of obtaining the target space-frequency base activation set in step S102 to step S103 in the embodiment of the present invention are as follows:

[0072] 1) Parameter initialization. The initialization parameters include the precision parameter α g,m , noise accuracy parameter α0, maximum update times N maxiter , error threshold thr, update count index iter, and empty frequency base activation set Ω. The update count index iter has an initial value of 0. The empty frequency base activation set Ω is initially empty. The empty frequency base activation set Ω is defined as the set containing activated empty frequency bases.

[0073] 2) Use the update formula of the update number index to enter the current update number, that is, iter=iter+1.

[0074] 3) Use equations (1) and (2) to calculate any space-frequency basis {θ m ,Ψ g It should be noted that when the space-frequency base is not activated, the precision parameter α g,m Always equal to the initial value. m ,Ψ g} is activated, the accuracy parameter α g,mEqual to the accuracy parameter estimate at the last update number. g = 1,…,G; m = 1,…,N.

[0075] 4) Under the current update number, calculate any space-frequency basis {θ m ,Ψ g The likelihood function of the increase value ΔL g,m If δ g,m >0 and {θ m ,Ψ g}∈Ω, then If δ g,m >0 and but If δ g,m <0 and {θ m ,Ψ g}∈Ω, then

[0076] 5) Select the maximum likelihood function increase value from the likelihood function increase values corresponding to all space-frequency bases, and use the space-frequency base corresponding to the maximum likelihood function increase value as the target space-frequency base to be operated under the current number of updates. For example, the target space frequency base is the nth discrete angle grid point and the gth * The target space-frequency basis is composed of frequency sub-bands. The discriminant parameter can be expressed as

[0077] 6) If and Then, a re-estimation operation is performed on the target space-frequency basis and the corresponding inference parameters are updated.

[0078] 7) If and Then, an activation operation is performed on the target space-frequency basis and the corresponding inference parameters are updated.

[0079] 8) If and Then the target empty frequency basis is deleted and the corresponding inference parameters are updated.

[0080] 9) When iter=1, directly return to step 2) to enter the next update; when iter≥2, calculate the convergence iterative error E err . E err =|ΔL (iter) -ΔL (iter-1) | / |ΔL (iter) -ΔL (1) |. Among them, ΔL (iter)The increase value of the likelihood function of the target space-frequency basis at the iter-th iteration (i.e., the current number of updates). ΔL (iter-1) Increase the likelihood function of the target space-frequency basis for the iter-1th iteration by ΔL (1) Add values to the likelihood function of the target space-frequency basis for the first iteration.

[0081] 10) Determine whether E is satisfied err <thr or iter> N maxiter If it is satisfied, it indicates that the update iterative process converges. At this time, the update requirements are not met, the update is terminated, and the space-frequency base activation set Ω at this time is output. The output space-frequency base activation set Ω is the target space-frequency base activation set that only includes the activated space-frequency bases. If it is not satisfied, it indicates that the update iterative process does not converge. At this time, the update requirements are met, and the update requirements are met. Return to step 2) to update the current update number, substitute the updated inference parameters into equations (3) and (4), and re-execute steps 3)-9) until E is satisfied. err <thr or iter> N maxiter .

[0082] Considering that the activation, re-estimation, and deletion operations performed in each update iteration will change the space-frequency basis contained in the set Ω and its corresponding parameters (including the precision parameter, the discrimination parameter, the first intermediate variable, and the second intermediate variable), the changes in these corresponding parameters are related to the inference parameters involved in the three operations 6)-8). The process of updating the inference parameters in the three operations is as follows:

[0083] The process of updating the corresponding inference parameters when performing the activation operation includes:

[0084] First, calculate the covariance matrix ∑ of all target space-frequency bases in the target space-frequency base activation set at the fth frequency point f Updated value of

[0085]

[0086] Where, When m is equal to n corresponding to the target space-frequency basis, S can be calculated by referring to the formula of the first defined variable. f,n , is the estimated value of the precision parameter of the target space-frequency basis. -1 Represents the inverse of a matrix or the reciprocal of a number. Φ f,·n is the array steering vector associated with the nth discrete angle grid point at the fth frequency point.

[0087] Secondly, the mean u f Updated value of (t) satisfy:

[0088]

[0089] In the formula, let m be equal to n corresponding to the target space-frequency basis and refer to the formula of the second defined variable to calculate Q f,n (t).

[0090] Finally, for any space-frequency basis {θ m ,Ψ g}, first define the variable S f,m Updated value of and second define the variable Q f,m Updated value of (t) satisfy:

[0091]

[0092] Where,

[0093] The process of updating the corresponding inference parameters when performing the re-estimation operation includes:

[0094] First, calculate the covariance matrix ∑ of all target space-frequency bases in the target space-frequency base activation set at the fth frequency point f Updated value of

[0095]

[0096] Where, ∑ f.n ,∑ f,n ,∑ f (n,n) are ∑ f The nth column, nth row, and (n, n)th element of . is the estimated value of the precision parameter under the last update number after the target space-frequency base is activated. f Updated value of (t) satisfy:

[0097]

[0098] Where u f,n (t) is the mean u f The nth element of (t).

[0099] Finally, for any space-frequency basis {θ m ,Ψ g}, first define the variable S f,m Updated value of and second define the variable Q f,m Updated value of (t) satisfy:

[0100]

[0101] The process of updating the corresponding inference parameters when performing a delete operation includes:

[0102] First, calculate the covariance matrix ∑ of all target space-frequency bases in the target space-frequency base activation set at the fth frequency point f Updated value of

[0103]

[0104] Secondly, the mean u f Updated value of (t) satisfy:

[0105]

[0106] Finally, for any space-frequency basis {θ m ,Ψ g}, first define the variable S f,m Updated value of and second define the variable Q f,m Updated value of (t) satisfy:

[0107]

[0108] Step S104: combining the space-frequency bases in the target space-frequency base activation set to obtain preliminary DOA estimation results and frequency band information of multiple wideband signals.

[0109] In step S104, combining the space-frequency bases in the target space-frequency base activation set specifically refers to combining the frequency subbands of the space-frequency bases with the same angle in all the space-frequency bases. Different angles correspond to different broadband signals. The preliminary DOA estimation results of different broadband signals are the corresponding angles, and the frequency band information of different broadband signals is the combined frequency subbands. For example, the preliminary DOA estimation result of the k-th broadband signal is the angle θ k , all angles θ in the target space-frequency basis activation set k The frequency band Ψ is obtained by combining the frequency sub-bands of the space-frequency basis k Band Ψ k is the frequency band occupancy of the k-th broadband signal.

[0110] Step S105 : Using the maximum likelihood criterion, a final DOA estimation result of the corresponding wideband signal is obtained based on the preliminary DOA estimation result and the frequency band information of each wideband signal.

[0111] In step S105, the maximum likelihood criterion is adopted to obtain the final DOA estimation result of the corresponding wideband signal based on the preliminary DOA estimation result of each wideband signal, including: for any wideband signal, determining the target angle range based on the preliminary DOA estimation result of the wideband signal, and the angle step of the target angle range is smaller than the angle step of the set angle range; for each target angle in the target angle range of the wideband signal, obtaining the corresponding target likelihood function based on the frequency band information, and adopting the maximum likelihood criterion to obtain the final DOA estimation result of the wideband signal based on the target likelihood function of all target angles.

[0112] Taking the kth broadband signal as an example, assuming that the angle and frequency band of the kth broadband signal are θ k and Ψ k For angle θ k First define a target angle range The target angle range is in θ k A denser set of angle grid points nearby. N k is the number of target angles in the target angle range, that is, the angle step of the target angle range is smaller than the angle step of the set angle range. Any target angle in the target angle range can be Indicates that i takes 1,…,N k .

[0113] For the k-th broadband signal, the accuracy parameter of any target angle can be re-estimated as:

[0114]

[0115] in, F k is the frequency band of the kth broadband signal Ψ k The total number of frequency points included, T is the number of snapshots of the array receiving signal at a single frequency point. f,i and q f,i (t) is defined as follows:

[0116]

[0117] Where (21)-(22) use the array flow pattern vector associated with the newly defined target angle range in for ). At the same time, C f,-k C f Remove the submatrix with the coarse estimate term associated with the k-th wideband signal.

[0118] Then refer to (7) to calculate the likelihood function corresponding to each target angle The DOA search strategy under the maximum likelihood criterion of formula (23) is used to achieve more accurate DOA estimation:

[0119]

[0120] Based on formula (23), the target angle corresponding to the maximum likelihood function can be obtained. The target angle is the angle corresponding to the accurately estimated k-th broadband signal.

[0121] In order to implement the above embodiment, the present invention further proposes a fast broadband DOA estimation system for signal frequency band heterogeneity based on joint space-frequency basis selection.

[0122] Figure 3 A block diagram of a fast broadband DOA estimation system for signal frequency band heterogeneity based on joint space-frequency basis selection provided by an embodiment of the present invention.

[0123] like Figure 3 As shown, the fast broadband DOA estimation system for signal frequency band heterogeneity based on joint space-frequency basis selection includes a division module 11, an update module 12, a preliminary estimation module 13 and a final estimation module 14, wherein:

[0124] A division module 11 is configured to obtain a space-frequency basis set based on a set angle range and a set frequency band, wherein each space-frequency basis in the space-frequency basis set includes an angle and a frequency sub-band;

[0125] An updating module 12 is configured to calculate a precision parameter estimate and a discrimination parameter for each space-frequency basis based on the array received signal, and calculate a likelihood function increase value for each space-frequency basis based on the precision parameter estimate value; use the space-frequency basis corresponding to the maximum likelihood function increase value as a target space-frequency basis, update the space-frequency basis activation set and the corresponding inference parameters based on the discrimination parameter of the target space-frequency basis; if the update requirements are met, update the precision parameter estimate and the discrimination parameter of each space-frequency basis based on the inference parameters, and further update the space-frequency basis activation set and the inference parameters, until the update requirements are no longer met, thereby obtaining a target space-frequency basis activation set;

[0126] A preliminary estimation module 13 is configured to combine the space-frequency bases in the target space-frequency base activation set to obtain preliminary DOA estimation results and frequency band information of multiple broadband signals;

[0127] The final estimation module 14 is configured to obtain a final DOA estimation result of the corresponding wideband signal based on the preliminary DOA estimation result and frequency band information of each wideband signal using a maximum likelihood criterion.

[0128] Furthermore, in a possible implementation of an embodiment of the present invention, the update module 12 calculates the likelihood function increase value of each space-frequency basis based on the precision parameter estimate value, including: for any space-frequency basis, calculating the corresponding likelihood function based on the precision parameter estimate value; if the discrimination parameter of the space-frequency basis is greater than zero and the space-frequency basis belongs to the space-frequency basis activation set, then the likelihood function increase value is the difference between the likelihood function under the current update number and the historical likelihood function; if the discrimination parameter of the space-frequency basis is greater than zero and the space-frequency basis does not belong to the space-frequency basis activation set, then the likelihood function increase value is the likelihood function under the current update number; if the discrimination parameter of the space-frequency basis is less than zero and the space-frequency basis belongs to the space-frequency basis activation set, then the likelihood function increase value is the inverse of the likelihood function under the current update number.

[0129] Furthermore, in a possible implementation of an embodiment of the present invention, in the update module 12, the inference parameters include the covariance matrix and mean of the signal distribution of each activated space-frequency basis in the space-frequency basis activation set at different frequency points, and multiple defined variables corresponding to each space-frequency basis in the space-frequency basis set; each defined variable is determined based on the array steering vector and the array received signal.

[0130] Furthermore, in a possible implementation manner of an embodiment of the present invention, when the update module 12 is used to update the space-frequency base activation set based on the discrimination parameter of the target space-frequency base, it is specifically used to: if the discrimination parameter of the target space-frequency base is greater than zero and the target space-frequency base belongs to the space-frequency base activation set, then re-estimate the precision parameter estimation value of the target space-frequency base; if the discrimination parameter of the target space-frequency base is greater than zero and the target space-frequency base does not belong to the space-frequency base activation set, then update the target space-frequency base to the space-frequency base activation set; if the discrimination parameter of the target space-frequency base is less than zero and the target space-frequency base belongs to the space-frequency base activation set, then delete the target space-frequency base from the space-frequency base activation set.

[0131] Furthermore, in a possible implementation of an embodiment of the present invention, in the update module 12, satisfying the update requirement means that the number of updates is equal to 1, or when the number of updates is not equal to 1, the convergence iteration error is greater than or equal to the error threshold and the number of updates is less than or equal to the maximum number of updates.

[0132] Furthermore, in a possible implementation of an embodiment of the present invention, the final estimation module 14 is specifically used to: adopt the maximum likelihood criterion to obtain the final DOA estimation result of the corresponding broadband signal and frequency band information based on the preliminary DOA estimation result of each broadband signal, including: for any broadband signal, determining the target angle range based on the preliminary DOA estimation result of the broadband signal, and the angle step of the target angle range is smaller than the angle step of the set angle range; for each target angle in the target angle range of the broadband signal, obtaining the corresponding target likelihood function based on the frequency band information, and adopting the maximum likelihood criterion to obtain the final DOA estimation result of the broadband signal based on the target likelihood function of all target angles.

[0133] It should be noted that the above explanation of the embodiment of the fast broadband DOA estimation method for signal frequency band heterogeneity based on joint space-frequency basis selection is also applicable to the fast broadband DOA estimation system for signal frequency band heterogeneity based on joint space-frequency basis selection in this embodiment, and will not be repeated here.

[0134] In an embodiment of the present invention, a space-frequency basis set is obtained by dividing based on a set angle range and a set frequency band, and each space-frequency basis in the space-frequency basis set includes an angle and a frequency sub-band; the precision parameter estimation value and discrimination parameter of each space-frequency basis are calculated based on the array received signal, and the likelihood function increase value of each space-frequency basis is calculated based on the precision parameter estimation value; the space-frequency basis corresponding to the maximum likelihood function increase value is used as the target space-frequency basis, and the space-frequency basis activation set is updated based on the discrimination parameter of the target space-frequency basis and the corresponding inference parameter is updated; if the update requirement is met, the precision parameter estimation value and discrimination parameter of each space-frequency basis are updated based on the inference parameter, and then the space-frequency basis activation set and the inference parameter are updated until the update requirement is no longer met, so as to obtain the target space-frequency basis activation set; the space-frequency basis in the target space-frequency basis activation set is combined to obtain the preliminary DOA estimation results and frequency band information of multiple wideband signals; and the maximum likelihood criterion is adopted to obtain the final DOA estimation result of the corresponding wideband signal based on the preliminary DOA estimation result and frequency band information of each wideband signal. In this case, a set of space-frequency bases is obtained by partitioning. Based on the precision parameter estimates, discrimination parameters, likelihood function increase values, and inference parameters of each space-frequency base, all space-frequency bases occupied by wideband signals are activated through update iterations to obtain a target space-frequency base activation set. By combining the space-frequency bases in the target space-frequency base activation set, the frequency band information of each wideband signal can be estimated. Then, using the maximum likelihood criterion, the final DOA estimation result of the corresponding wideband signal is obtained based on the preliminary DOA estimation result and the frequency band information. This achieves a more accurate DOA estimation.

[0135] The method and system of the present invention utilize a joint space-frequency basis selection algorithm: first, the entire space-frequency spectrum to be estimated is divided into multiple space-frequency basis points, each of which comprises a set of angle and frequency subbands. A maximum likelihood criterion is then constructed to sequentially select those space-frequency basis points that are actually occupied by the signal (i.e., each time the space-frequency basis points corresponding to the increment of the maximum likelihood function are updated, they are selected as target space-frequency basis points for re-estimation, activation, or deletion). Finally, by combining these activated space-frequency basis points, a rough DOA estimate and frequency band information of the signal source are obtained. During the precise DOA estimation process, the rough DOA estimate and frequency band information obtained by the joint space-frequency basis selection algorithm are used to design a corresponding precise DOA estimation process using the already acquired signal frequency band information. The method and system of the present invention, referring to the fast Bayesian theory, employs a point estimation strategy to activate only the space-frequency basis points occupied by wideband signals, thereby avoiding the matrix inversion operation in the SBL framework and improving the algorithm's computational efficiency. Furthermore, by analyzing all activated space-frequency basis points, the occupied frequency band of each wideband signal can be effectively inferred, and this frequency band information can be used to achieve more accurate DOA estimation.

[0136] In order to implement the above embodiments, the present invention also proposes an electronic device, comprising: a processor, and a memory communicatively connected to the processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory to implement the method provided by the above embodiments.

[0137] In order to implement the above embodiments, the present invention further provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the methods provided in the above embodiments.

[0138] In order to implement the above embodiments, the present invention further provides a computer program product, including a computer program, which implements the methods provided in the above embodiments when executed by a processor.

[0139] In the descriptions of the foregoing embodiments, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine different embodiments or examples described in this specification and features of different embodiments or examples, unless they are mutually inconsistent.

[0140] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0141] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention pertain.

[0142] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and a portable compact disc read-only memory (CDROM). Furthermore, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing it in another suitable manner if necessary, and then storing it in a computer memory.

[0143] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0144] Those skilled in the art will understand that all or part of the steps in the method of the above embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0145] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing module, or each unit may exist physically separately, or two or more units may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or in the form of software functional modules. If the integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may also be stored in a computer-readable storage medium.

[0146] The storage medium mentioned above may be a read-only memory, a magnetic disk, or an optical disk, etc. Although the embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and are not to be construed as limiting the present invention. Persons skilled in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A fast broadband DOA estimation method for signal frequency band heterogeneity based on joint space-frequency basis selection, characterized in that: include: A space-frequency basis set is obtained based on the set angle range and the set frequency band, where each space-frequency basis in the space-frequency basis set includes an angle and a frequency sub-band; Calculating the precision parameter estimation value and discrimination parameter of each space-frequency basis based on the array received signal, and calculating the likelihood function increase value of each space-frequency basis based on the precision parameter estimation value; taking the space-frequency basis corresponding to the maximum likelihood function increase value as the target space-frequency basis, and updating the space-frequency basis activation set and the corresponding inference parameters based on the discrimination parameter of the target space-frequency basis; If the update requirement is met, the precision parameter estimation value and the discrimination parameter of each space-frequency basis are updated based on the inferred parameter, and the space-frequency basis activation set and the inferred parameter are further updated until the update requirement is no longer met, so as to obtain a target space-frequency basis activation set; Combining the space-frequency bases in the target space-frequency base activation set to obtain preliminary DOA estimation results and frequency band information of multiple wideband signals; A maximum likelihood criterion is adopted to obtain a final DOA estimation result of the corresponding wideband signal based on the preliminary DOA estimation result and frequency band information of each wideband signal.

2. The fast broadband DOA estimation method for signal frequency band heterogeneity based on joint space-frequency basis selection according to claim 1, characterized in that: The calculating the likelihood function increase value of each space-frequency basis based on the precision parameter estimation value includes: For any space-frequency basis, calculating a corresponding likelihood function based on the precision parameter estimate; If the discriminant parameter of the space-frequency basis is greater than zero and the space-frequency basis belongs to the space-frequency basis activation set, the likelihood function increment is the difference between the likelihood function at the current update number and the historical likelihood function; If the discriminant parameter of the space-frequency basis is greater than zero and the space-frequency basis does not belong to the space-frequency basis activation set, the likelihood function increment value is the likelihood function under the current update number; If the discrimination parameter of the space-frequency basis is less than zero and the space-frequency basis belongs to the space-frequency basis activation set, the increase value of the likelihood function is the opposite of the likelihood function under the current update number.

3. The fast broadband DOA estimation method for signal frequency band heterogeneity based on joint space-frequency basis selection according to claim 1, characterized in that: The updating of the space-frequency basis activation set based on the discrimination parameter of the target space-frequency basis includes: If the discrimination parameter of the target space-frequency basis is greater than zero and the target space-frequency basis belongs to the space-frequency basis activation set, re-estimating the precision parameter estimation value of the target space-frequency basis; If the discrimination parameter of the target space-frequency base is greater than zero and the target space-frequency base does not belong to the space-frequency base activation set, updating the target space-frequency base to the space-frequency base activation set; If the discrimination parameter of the target space-frequency base is less than zero and the target space-frequency base belongs to the space-frequency base activation set, the target space-frequency base is deleted from the space-frequency base activation set.

4. The fast broadband DOA estimation method for signal frequency band heterogeneity based on joint space-frequency basis selection according to claim 1, characterized in that: The satisfying the update requirement means that the number of updates is equal to 1, or when the number of updates is not equal to 1, the convergence iteration error is greater than or equal to the error threshold and the number of updates is less than or equal to the maximum number of updates.

5. The fast broadband DOA estimation method for signal frequency band heterogeneity based on joint space-frequency basis selection according to claim 1, characterized in that: The inference parameters include the covariance matrix and mean of the signal distribution of each activated space-frequency basis in the space-frequency basis activation set at different frequency points, as well as multiple defined variables corresponding to each space-frequency basis in the space-frequency basis set; each defined variable is determined based on the array steering vector and the array received signal.

6. The fast broadband DOA estimation method for signal frequency band heterogeneity based on joint space-frequency basis selection according to claim 1, characterized in that: The method of adopting the maximum likelihood criterion to obtain the final DOA estimation result of the corresponding wideband signal based on the preliminary DOA estimation result and frequency band information of each wideband signal includes: For any wideband signal, determining a target angle range based on a preliminary DOA estimation result of the wideband signal, wherein an angle step size of the target angle range is smaller than an angle step size of the set angle range; For each target angle in the target angle range of the broadband signal, the corresponding target likelihood function is obtained based on the frequency band information, and the maximum likelihood criterion is adopted to obtain the final DOA estimation result of the broadband signal based on the target likelihood functions of all target angles.

7. A fast broadband DOA estimation system for signal frequency band heterogeneity based on joint space-frequency basis selection, characterized in that: include: A division module, configured to obtain a space-frequency basis set based on a set angle range and a set frequency band, wherein each space-frequency basis in the space-frequency basis set includes an angle and a frequency sub-band; an updating module, configured to calculate a precision parameter estimate and a discrimination parameter for each space-frequency basis based on the array received signal, and calculate a likelihood function increase value for each space-frequency basis based on the precision parameter estimate value; take the space-frequency basis corresponding to the maximum likelihood function increase value as a target space-frequency basis, and update a space-frequency basis activation set and corresponding inference parameters based on the discrimination parameter of the target space-frequency basis; if an update requirement is met, update the precision parameter estimate and the discrimination parameter of each space-frequency basis based on the inference parameter, and further update the space-frequency basis activation set and the inference parameters, until the update requirement is no longer met, thereby obtaining a target space-frequency basis activation set; a preliminary estimation module, configured to combine the space-frequency bases in the target space-frequency base activation set to obtain preliminary DOA estimation results and frequency band information of multiple broadband signals; The final estimation module is used to obtain the final DOA estimation result of the corresponding wideband signal based on the DOA preliminary estimation result and frequency band information of each wideband signal by adopting the maximum likelihood criterion.

8. The fast broadband DOA estimation system for signal frequency band heterogeneity based on joint space-frequency basis selection according to claim 7, characterized in that: When the updating module is used to update the space-frequency base activation set based on the discrimination parameter of the target space-frequency base, the updating module is specifically used to: if the discrimination parameter of the target space-frequency base is greater than zero and the target space-frequency base belongs to the space-frequency base activation set, re-estimate the precision parameter estimation value of the target space-frequency base; if the discrimination parameter of the target space-frequency base is greater than zero and the target space-frequency base does not belong to the space-frequency base activation set, update the target space-frequency base to the space-frequency base activation set; If the discrimination parameter of the target space-frequency base is less than zero and the target space-frequency base belongs to the space-frequency base activation set, the target space-frequency base is deleted from the space-frequency base activation set.

9. An electronic device, characterized in that: include: a processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 6 when executed by a processor.

Images