Multi-peak searching method based on two-dimensional adaptive threshold and related equipment

Abstract

The invention provides a multi-peak searching method based on a two-dimensional adaptive threshold and related equipment, and relates to the technical field of wireless communication. Calculating a sample covariance matrix; performing eigendecomposition on the sample covariance matrix to obtain a plurality of eigenvalues and a plurality of eigenvectors, and constructing a noise subspace matrix according to each eigenvector; calculating a noise power estimation value according to each characteristic value, and calculating an adaptive threshold according to the noise power estimation value; constructing a two-dimensional MUSIC spatial spectrum function according to the noise subspace matrix, and performing twice search according to the spatial spectrum function and an adaptive threshold to obtain a plurality of second candidate peaks; determining a mirror image peak corresponding to each second candidate peak, and screening each second candidate peak according to each mirror image peak to obtain a plurality of third candidate peaks; and searching each third candidate peak for the third time, and determining the third candidate peak meeting a second preset condition as a target peak. By adopting the method, efficient and accurate multi-peak detection can be realized in a complex environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless communication, and in particular to a multi-peak search method based on a two-dimensional adaptive threshold and related equipment. BACKGROUND

[0002] In the field of wireless communication, with the wide application of multi-antenna array systems in the fields of automatic driving, 5G communication, etc., the direction of arrival (DOA) estimation technology faces the dual challenges of stable detection and real-time processing in complex electromagnetic environments. The multiple signal classification (MUSIC) algorithm has become the mainstream technology for two-dimensional DOA estimation due to its super-resolution capability.

[0003] The traditional MUSIC algorithm usually adopts a peak detection process based on offline calibration. First, a fixed threshold value is determined as a system parameter through repeated tests in a specific environment, and then a global search strategy based on a uniform grid is used for peak identification. In specific implementation, the azimuth angle domain and the elevation angle domain are equally divided into regular grids according to the same angle step, the spatial spectrum function value of each grid node is calculated one by one, and the spectrum value exceeding the preset fixed threshold is determined as a candidate peak. Through this combination of "fixed threshold + global search", the spatial positioning of the signal source is realized.

[0004] However, the above-mentioned "fixed threshold + global search" technical solution exposes systematic defects when facing complex and variable actual environments. The fixed threshold based on offline calibration cannot adapt to the real-time changes of noise level and signal strength, resulting in a decrease in detection accuracy. In order to compensate for the detection errors caused by improper threshold setting, existing methods often need to use more intensive grid search to improve detection reliability, which further aggravates the computational burden. However, the high computational complexity of global exhaustive search limits the system's ability to adjust and optimize detection parameters in real time, forming a technical bottleneck of "insufficient detection accuracy requiring intensive search, and intensive search cannot support parameter optimization", making it difficult for the traditional scheme to achieve efficient and accurate detection of target peaks in complex environments. SUMMARY

[0005] The present application provides a multi-peak search method based on a two-dimensional adaptive threshold and related equipment, which can achieve efficient and accurate detection of target peaks in complex environments.

[0006] In a first aspect of the present application, a multi-peak search method based on a two-dimensional adaptive threshold is provided, which specifically includes: receiving a signal data matrix about a target sent by a multi-antenna array at the current time, and calculating a sample covariance matrix according to the signal data matrix; performing eigen-decomposition on the sample covariance matrix to obtain a plurality of eigenvalues and a plurality of eigenvectors, and constructing a noise subspace matrix according to the eigenvectors; calculating noise power estimation values according to the eigenvalues, and calculating an adaptive threshold according to the noise power estimation values; constructing a two-dimensional MUSIC spatial spectrum function according to the noise subspace matrix, performing a first search on spatial spectrum values of the two-dimensional MUSIC spatial spectrum function, and taking spatial spectrum values greater than the adaptive threshold as first candidate peaks, the two-dimensional MUSIC spatial spectrum function being a corresponding relationship between each spatial spectrum value and an angle combination of an azimuth and an elevation corresponding to the spatial spectrum value; performing a second search within a preset neighborhood range of each of the first candidate peaks to obtain a plurality of second candidate peaks, a step of the second search being smaller than a step of the first search; determining mirror peaks corresponding to the second candidate peaks according to a geometric structure of the multi-antenna array, and screening the second candidate peaks according to the mirror peaks and a first preset condition to obtain a plurality of third candidate peaks; performing a third search on each of the third candidate peaks within a preset angle range of the third candidate peak, and determining a third candidate peak satisfying a second preset condition as a target peak.

[0007] By using the above technical solution, eigenvalues are obtained by performing eigen-decomposition on the sample covariance matrix, and noise power estimation values and an adaptive threshold are calculated in real time based on the eigenvalues, so that the detection threshold can be dynamically adjusted according to the actual situation of the current signal environment, and the technical defect that a fixed threshold cannot adapt to a complex electromagnetic environment is effectively overcome. At the same time, the method uses a three-level progressive search strategy, first uses the adaptive threshold to perform rough search to obtain first candidate peaks, then uses a smaller step to perform fine search in the neighborhood of the candidate peaks to obtain second candidate peaks, and finally analyzes and screens mirror peaks in combination with the geometric structure characteristics of the multi-antenna array and completes the final search, so as to significantly reduce the computational complexity while ensuring the detection accuracy. This technical solution based on the adaptive threshold and the multi-level search realizes real-time optimization of detection parameters and efficient use of computing resources, breaks through the performance bottleneck of the traditional "fixed threshold + global search" scheme in a complex environment, and realizes efficient and accurate detection of target peaks in a complex environment.

[0008] Optionally, the calculating noise power estimation values according to the eigenvalues, and calculating an adaptive threshold according to the noise power estimation values, comprises: calculating noise power estimation values based on a first formula; the first formula is: ; wherein, Here, N is the noise power estimate, N is the number of elements in the multi-antenna array, and M is the amount of signal data. Let be the i-th eigenvalue in the noise subspace; The adaptive threshold is calculated based on the second formula; The second formula is: ; Where T is the adaptive threshold. K is the noise power estimate, and K is the total number of search points, which is determined by the search step size of the azimuth angle and the pitch angle.

[0009] By adopting the above technical solution, the noise power estimate is calculated based on the first formula using the eigenvalues ​​in the noise subspace, fully utilizing the statistical characteristics of the noise subspace. Here, the denominator (NM) represents the dimension of the noise subspace, and the numerator sums and averages the eigenvalues ​​from the (M+1)th to the Nth eigenvalue in the noise subspace, effectively improving the accuracy and stability of the noise power estimate and avoiding errors that may arise from estimating a single eigenvalue. Furthermore, by calculating an adaptive threshold based on the noise power estimate and the total number of search points K using the second formula, a truly data-driven threshold calculation is achieved. The total number of search points K is determined by the search step size of the azimuth and elevation angles, allowing the adaptive threshold to be dynamically adjusted according to the actual search density, ensuring appropriate detection sensitivity under different search accuracies. This adaptive threshold calculation method based on the statistical distribution of noise subspace eigenvalues, compared to traditional fixed threshold settings, better adapts to changes in signal-to-noise ratios and noise environments, avoiding false peak detection caused by excessively low thresholds and preventing the omission of true targets caused by excessively high thresholds.

[0010] Optionally, the first search of the spatial spectral values ​​of the two-dimensional MUSIC spatial spectral function, selecting spatial spectral values ​​greater than the adaptive threshold as first candidate peaks, includes: Search for the spatial spectrum value in the two-dimensional MUSIC spatial spectrum function at each of the first angle steps according to the preset first angle step size; The spatial spectrum value in each of the first angle steps that is greater than the adaptive threshold and is a maximum value is taken as the first candidate peak. The azimuth and elevation angles in the angle combination corresponding to the first candidate peak are respectively located in the corresponding preset azimuth and preset elevation angle ranges.

[0011] By employing the aforementioned technical solution, the two-dimensional MUSIC spatial spectral function is searched according to a preset first angle step size, achieving efficient coverage in the coarse search stage. Compared to traditional methods that use a fixed fine step size to fully traverse the entire angle space, this reduces computational complexity and improves search efficiency. Simultaneously, by setting dual screening conditions—requiring the spatial spectral value to be both greater than an adaptive threshold and a maximum—the quality and reliability of the first candidate peak are improved. The condition of being greater than the adaptive threshold ensures that the candidate peak has sufficient signal strength to distinguish it from the noise background, while the maximum condition guarantees that the candidate peak has the largest spatial spectral response in a local region, avoiding misclassification of non-peak points as candidate targets. Furthermore, by restricting the azimuth and elevation angles in the angle combination corresponding to the first candidate peak to fall within their respective preset azimuth and elevation angle ranges, invalid searches in physically impossible or uninteresting angle regions are avoided, further improving the algorithm's practicality and relevance. This first search strategy, which combines coarse search step size, dual screening conditions, and angle range constraints, not only ensures the effective identification of potential target regions but also provides a high-quality set of candidate peaks for subsequent fine searches, thereby improving the overall algorithm's operating efficiency while ensuring detection performance.

[0012] Optionally, the second search within a preset neighborhood of each of the first candidate peaks to obtain multiple second candidate peaks includes: The azimuth and elevation angles corresponding to each of the first candidate peaks are used as centers to determine the azimuth and elevation angle neighborhood ranges, respectively. Based on the second angle step size, each azimuth angle is traversed within the azimuth angle neighborhood, and each pitch angle is traversed within the pitch angle neighborhood, to obtain the set of fine search angles corresponding to each first candidate peak. The second angle step size is smaller than the first angle step size. Extract the spatial spectrum value corresponding to each combination of fine search angles in each set of fine search angles from the two-dimensional MUSIC spatial spectrum function, and determine the largest spatial spectrum value in each set of fine search angles as the second candidate peak.

[0013] By employing the aforementioned technical solution, the azimuth and elevation neighborhoods are determined centered on the azimuth and elevation angles corresponding to each first candidate peak, respectively. This achieves an effective transition from global coarse search to local fine search, avoiding the enormous computational overhead of high-precision traversal of the entire angle space. Simultaneously, by traversing each azimuth and elevation angle within the neighborhood using a second angle step size smaller than the first angle step size, a fine search angle set for each first candidate peak is constructed. This fine-step local search strategy enables more accurate angle positioning within the vicinity of the first candidate peak, effectively improving the accuracy and resolution of target angle estimation. Furthermore, by extracting the spatial spectral values ​​corresponding to each fine search angle combination from the two-dimensional MUSIC spatial spectral function, and determining the largest spatial spectral value in each fine search angle set as the second candidate peak, the true peak position within each neighborhood is accurately identified and located, avoiding the problems of missing the true peak due to an excessively large search step size or generating excessive redundant computation due to an excessively small step size. This neighborhood-based hierarchical fine-grained search mechanism not only improves the accuracy of angle estimation, but also effectively balances the relationship between algorithm accuracy and computational efficiency by limiting the high-precision search to a limited neighborhood range, providing a more reliable set of candidate peaks for subsequent mirror peak screening and target confirmation.

[0014] Optionally, the step of determining the mirror peak corresponding to each second candidate peak based on the geometry of the multi-antenna array, and filtering each second candidate peak based on each mirror peak and a first preset condition to obtain multiple third candidate peaks includes: The axis of symmetry is determined based on the geometry of the multi-antenna array, and the mirror angle position corresponding to each second candidate peak is calculated based on the axis of symmetry. Extract the mirror space spectrum value corresponding to each mirror angle position from the two-dimensional MUSIC space spectrum function, and determine each mirror space spectrum value as a mirror peak; Calculate the average spatial spectrum value of each second candidate peak and the average mirror spatial spectrum value of the mirror peak corresponding to each second candidate peak within a preset time range after the current time. The second candidate peak whose average spatial spectral value is greater than the average mirror spatial spectral value is determined as the third candidate peak.

[0015] By employing the aforementioned technical solution, the symmetry axis is determined based on the geometric structure of the multi-antenna array, and the mirror angle position corresponding to each second candidate peak is calculated based on this symmetry axis. This fully utilizes the physical symmetry characteristics of the antenna array, providing a theoretical basis for accurately identifying false mirror peaks caused by array geometric asymmetry or signal correlation. Simultaneously, by extracting the mirror spatial spectrum value corresponding to each mirror angle position from the two-dimensional MUSIC spatial spectrum function and identifying it as a mirror peak, a direct correspondence between the real target peak and its corresponding mirror peak is established, making subsequent authenticity determination possible. Furthermore, by calculating the average spatial spectrum value of each second candidate peak within a preset time range after the current moment, and the average mirror spatial spectrum value of the corresponding mirror peak of each second candidate peak, the statistical characteristics of the time dimension are effectively utilized to improve the stability and reliability of mirror peak discrimination, avoiding the impact of random errors that may exist in single-moment measurements on the screening results. Finally, by identifying the second candidate peak with an average spatial spectrum value greater than the average mirror spatial spectrum value as the third candidate peak, intelligent screening based on statistical characteristics is achieved. The real target, due to its physical stability, usually has a spatial spectrum value that is consistently higher than the spatial spectrum value of its mirror position, while false mirror peaks often exhibit the opposite characteristics or random fluctuations. This mirror peak screening mechanism, which combines array geometric symmetry, time statistical averaging, and spatial spectral value comparison, significantly improves the algorithm's ability to identify real targets in complex multi-target environments and effectively reduces the incidence of false detections.

[0016] Optionally, the step of performing a third search on each of the third candidate peaks within a preset angle range, and determining the third candidate peak that satisfies the second preset condition as the target peak, includes: Centered on the azimuth and elevation angles corresponding to each of the third candidate peaks, each of the fine search angle combinations is traversed within the preset azimuth and elevation angle superposition ranges. Extract the spatial spectrum values ​​corresponding to each of the fine search angle combinations from the two-dimensional MUSIC spatial spectrum function, and superimpose the spatial spectrum values ​​to obtain the total power value corresponding to each of the third candidate peaks; Based on the total power values, a maximum peak value search is performed on each of the third candidate peaks, and the third candidate peak that meets the second preset condition is determined as the fourth candidate peak. The second preset condition is that the spatial spectral value corresponding to the third candidate peak is the maximum value. The trajectory smoothing process is performed on each of the fourth candidate peaks to obtain the target peak.

[0017] By employing the aforementioned technical solution, and traversing each fine-search angle combination within a preset azimuth and elevation angle superposition range centered on the azimuth and elevation angles corresponding to each third candidate peak, local neighborhood power accumulation analysis for each candidate target is achieved. This spatial neighborhood superposition strategy effectively improves the power concentration of the real target signal while smoothing and suppressing random noise and spurious peaks. Simultaneously, by extracting the spatial spectrum values ​​corresponding to each fine-search angle combination from the two-dimensional MUSIC spatial spectrum function and superimposing them, the total power value corresponding to each third candidate peak is obtained. This fully utilizes the characteristic that the real target has strong spatial spectrum response continuity within its angular neighborhood, making the total power value of the real target significantly higher than the noise background and spurious peaks, thus providing a more reliable basis for subsequent target confirmation. Furthermore, by searching for maximum peak points of each third candidate peak based on each total power value, and determining the third candidate peak that meets the second preset condition (i.e., the spatial spectrum value corresponding to the third candidate peak is a maximum value) as the fourth candidate peak, a dual verification mechanism based on power concentration and extreme value characteristics is realized, significantly improving the accuracy and reliability of target confirmation. Finally, the target peak was obtained by smoothing the trajectory of each fourth candidate peak, which effectively eliminated the random fluctuations in angle estimation caused by factors such as measurement noise and channel changes, and ensured the continuity and stability of target angle tracking.

[0018] Optionally, the step of superimposing the spatial spectral values ​​to obtain the total power value corresponding to each of the third candidate peaks includes: The total power value corresponding to each of the third candidate peaks is calculated based on the third formula. The third formula is: ; in, This represents the total power value corresponding to the third candidate peak. The azimuth angle within the i-th angle combination in the preset angular neighborhood of the third candidate peak. The pitch angle within the i-th angle combination in the preset angle neighborhood of the third candidate peak is determined. Let be the spatial spectrum value corresponding to the i-th angle combination within the preset angle neighborhood of the third candidate peak, k and K are the preset angle neighborhood parameters of the third candidate peak in azimuth and elevation angles, respectively, m and M are the neighborhood range parameters of the mirror peak in azimuth and elevation angles, and j is the index of the angle combination within the preset angle neighborhood of the mirror peak.

[0019] By adopting the above technical solution, this application achieves a quantitative assessment of the target peak power concentration by accurately calculating the total power value corresponding to each third candidate peak based on the third formula. This avoids the uncertainty and inconsistency that may arise from relying on empirical thresholds or subjective judgments in traditional methods. Furthermore, the third formula introduces the azimuth, elevation, and corresponding spatial spectrum values ​​corresponding to each angle combination within the preset angular neighborhood of the third candidate peak as calculation parameters. This fully considers the spatial spectrum response distribution characteristics of the target within its angular neighborhood, ensuring that the real target exhibits strong power concentration within the neighborhood due to the stability of its physical existence. This allows for effective differentiation between the real target and false peaks based on the magnitude of the total power value. Moreover, by setting independent neighborhood range parameters k and K (preset angular neighborhoods for the third candidate peak at azimuth and elevation angles) and m and M (neighborhood ranges for mirror peaks at azimuth and elevation angles), a flexible processing strategy of differentiated neighborhood ranges for different types of peaks is achieved. This allows for dynamic adjustment of neighborhood parameters according to actual application scenarios and target characteristics, further enhancing the algorithm's adaptability and robustness. In addition, by introducing angle combination indices i and j corresponding to the angle combinations within the preset angle neighborhood of the third candidate peak and the mirror peak, the integrity and accuracy of the superposition calculation process are ensured, and the problems of omission or repeated calculation are avoided.

[0020] In a second aspect, this application provides a multi-peak search device based on a two-dimensional adaptive threshold, the multi-peak search device based on a two-dimensional adaptive threshold includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the multi-peak search device based on a two-dimensional adaptive threshold to perform the method as described in the first aspect and any possible implementation thereof.

[0021] Thirdly, this application provides a computer program product containing instructions that, when run on a two-dimensional adaptive threshold multi-peak search device, cause the two-dimensional adaptive threshold multi-peak search device to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, this application provides a computer-readable storage medium including instructions that, when executed on a two-dimensional adaptive threshold-based multi-peak search device, cause the two-dimensional adaptive threshold-based multi-peak search device to perform the method described in the first aspect and any possible implementation thereof. Attached Figure Description

[0023] Figure 1This is a flowchart illustrating a multi-peak search method based on a two-dimensional adaptive threshold provided in an embodiment of this application; Figure 2 This is an example diagram of the first search provided in the embodiments of this application; Figure 3 This is an example diagram of the second search provided in the embodiments of this application; Figure 4 This is an exemplary hardware structure diagram of a multi-peak search device based on two-dimensional adaptive threshold provided in an embodiment of this application. Detailed Implementation

[0024] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0025] In the description of the embodiments of this application, the words "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design that is described as "for example" or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design options. Rather, the use of the words "for example" or "for instance" is intended to present the relevant concepts in a specific manner.

[0026] In the description of the embodiments of this application, the term "multiple" means two or more. For example, multiple systems means two or more systems, and multiple screen terminals means two or more screen terminals. 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 indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0027] This application provides a multi-peak search method based on two-dimensional adaptive threshold, referencing... Figure 1 , Figure 1 This is a flowchart illustrating a multi-peak search method based on a two-dimensional adaptive threshold provided in an embodiment of this application, including steps S101 to S107, as follows: S101: Receive the signal data matrix about the target transmitted by the multi-antenna array at the current moment, and calculate the sample covariance matrix based on the signal data matrix.

[0028] In this embodiment, a multi-antenna array refers to an array antenna system composed of multiple antenna elements arranged according to a specific geometric structure, such as a circular array (CRA). The spacing between each element is typically set to half a wavelength to achieve spatial sampling and direction-of-arrival estimation of the target signal. The multi-antenna array transmits a signal data matrix about the target. This signal data matrix is ​​a complex data matrix formed after analog-to-digital conversion and preprocessing of the target signal received by each element in the multi-antenna array at the current moment. It contains spatial information such as the target's azimuth and elevation angles. A sample covariance matrix can be calculated based on the signal data matrix. This sample covariance matrix is ​​a statistical characteristic matrix calculated based on the signal data matrix, used to describe the correlation and power distribution characteristics between the signals received by each element, providing a mathematical basis for subsequent signal subspace decomposition and MUSIC spatial spectrum function construction.

[0029] Specifically, the electronic device synchronously receives electromagnetic signals from the target at the current moment through a multi-antenna array system. Each antenna processes the received analog signals through preprocessing steps such as RF front-end processing, analog-to-digital conversion, and digital down-conversion, forming a signal data matrix containing target spatial information. The number of rows in this signal data matrix corresponds to the number of array elements, and the number of columns corresponds to the number of sampling snapshots. Subsequently, the electronic device, based on the received signal data matrix, calculates the covariance matrix using the formula... Matrix operations are performed to calculate the sample covariance matrix between the signals of each array element. In this calculation, the signal data matrix X is first subjected to a conjugate transpose operation to obtain... Then X and Perform matrix multiplication, and finally divide by the number of sampling snapshots N to normalize the power, thus obtaining the covariance matrix.

[0030] S102: Perform eigenvalue decomposition on the sample covariance matrix to obtain multiple eigenvalues ​​and multiple eigenvectors, and construct the noise subspace matrix based on each eigenvector.

[0031] In this embodiment, eigenvalues ​​represent scalar values ​​obtained after eigenvalue decomposition of the sample covariance matrix, reflecting the magnitude of the variance of the covariance matrix in the direction of the corresponding eigenvector. Larger eigenvalues ​​correspond to the power contribution of the target signal, while smaller eigenvalues ​​correspond to the power contribution of noise. Eigenvectors are vectors corresponding to each eigenvalue, representing the main direction of change of the covariance matrix, and are used to construct the basic vectors of the signal subspace and noise subspace.

[0032] Specifically, first, the number of array elements N and the amount of signal data M of the multi-antenna array are obtained. Then, eigenvalue decomposition is performed on the covariance matrix of the received signal to obtain all eigenvalues, which are then sorted in descending order. Next, select the (M+1)th to the Nth eigenvalues, which correspond to the eigenvalues ​​of the noise subspace. The eigenvectors corresponding to these eigenvalues ​​form the noise subspace matrix.

[0033] S103: Calculate the noise power estimate based on each characteristic value, and calculate the adaptive threshold based on the noise power estimate.

[0034] In the embodiments of this application, the adaptive threshold represents a threshold used to distinguish between effective signal peaks and noise peaks, which can yield candidate angles, and the candidate angles refer to angle combinations that may correspond to the actual signal directions.

[0035] Specifically, these characteristic values ​​are substituted into the first formula to obtain the noise power estimate.

[0036] The first formula is: ,in, Here, N is the noise power estimate, N is the number of elements in the multi-antenna array, and M is the amount of signal data. Let be the i-th eigenvalue in the noise subspace. Then, substitute the obtained noise power estimate into the second formula to obtain the adaptive threshold.

[0037] The second formula is: Where T is the adaptive threshold. K is the noise power estimate, and K is the total number of search points, which is determined by the search step size of the azimuth angle and the pitch angle.

[0038] S104: Construct a two-dimensional MUSIC spatial spectrum function based on the noise subspace matrix. Perform a first search on the spatial spectrum values ​​of the two-dimensional MUSIC spatial spectrum function. Select the spatial spectrum values ​​that are greater than the adaptive threshold as the first candidate peaks. The two-dimensional MUSIC spatial spectrum function is the correspondence between each spatial spectrum value and the azimuth and elevation angle combinations corresponding to each spatial spectrum value.

[0039] In this embodiment, the two-dimensional MUSIC spatial spectrum function refers to a spatial power spectrum function for target angle estimation constructed based on the noise subspace matrix. Utilizing the orthogonality between the signal steering vector and the noise subspace, high-resolution two-dimensional angle search is achieved by calculating the projection of the steering vector onto the noise subspace. The two-dimensional MUSIC spatial spectrum function contains spatial spectrum values, which represent the power spectral density calculated by the two-dimensional MUSIC spatial spectrum function under a specific combination of azimuth and elevation angles. These values ​​reflect the probability of a target signal existing at that angular location; a larger value indicates a higher probability of a real target existing in that angular direction. To quickly locate potential target angle regions and reduce computational complexity, a first search is required. The first search refers to a coarse search process that traverses the entire angle search space using a large angle step size. Spatial spectrum values ​​that satisfy the threshold conditions and are local maxima identified during the first search correspond to the first candidate peaks representing the angle locations where real targets may exist.

[0040] Specifically, the direction response vector corresponding to each angle combination Based on the fundamental principles of the MUSIC algorithm, the MUSIC spatial spectrum function values ​​corresponding to various angle combinations are calculated. The specific calculation formula is as follows: ;in, The values ​​of the MUSIC spatial spectrum function corresponding to each angle combination are: Let be the eigenvector matrix of the noise subspace. This is the conjugate transpose of the directional response vector. This refers to the direction response vector. When the direction response vector is orthogonal to the noise subspace, the MUSIC spatial spectrum function values ​​corresponding to each angle combination will exhibit peak values. When the direction response vector is not orthogonal to the noise subspace, no peak values ​​will be formed. The electronic device operates within a preset azimuth range according to a preset first angular step size. A two-dimensional traversal search is performed on the angle ∈ [0°, 360°] and the preset pitch angle range θ∈ [0°, 90°] to calculate the spatial spectrum value corresponding to each angle combination, forming a power spectrum distribution map covering the entire angle space. During the search process, the electronic device compares the calculated spatial spectrum values ​​with an adaptive threshold and simultaneously checks whether the spatial spectrum value is a local maximum, i.e., if the condition is met. ,in and The first angular step size is defined as follows. For spatial spectral values ​​that simultaneously satisfy the conditions of being greater than the adaptive threshold and being local maxima, the electronic device marks them as the first candidate peak.

[0041] like Figure 2 As shown, Figure 2 This is an example diagram of the first search provided in the embodiments of this application. The diagram statistically analyzes the results of a preset azimuth range with a first angular step size. When performing a two-dimensional traversal search on angles ∈ [0°, 360°] and the preset pitch angle range θ∈ [0°, 90°], the spatial spectral values ​​corresponding to each angle combination are... Figure 2 The mid-plane axis represents the pitch and azimuth angles, and the vertical axis power spectrum represents the spatial spectrum values ​​corresponding to each angle combination.

[0042] Based on the above embodiments, as an optional embodiment, S104: the step of performing a first search on the spatial spectral values ​​of the two-dimensional MUSIC spatial spectral function and taking spatial spectral values ​​greater than the adaptive threshold as the first candidate peak may specifically include the following steps: S301: Search for spatial spectrum values ​​in the two-dimensional MUSIC spatial spectrum function at each first angle step according to the preset first angle step.

[0043] In this embodiment, the preset first angle step size represents the angle traversal interval parameter set in the coarse search stage, which is used to determine the sampling density when performing spatial spectrum search in the two dimensions of azimuth and elevation. It is usually set to 5° to 10° to achieve rapid coverage of the entire angle search space and reduce computational complexity.

[0044] Specifically, the electronic device, according to a preset first angular step size, with azimuth = 0°, Δ , 2Δ The sequence of , ..., 360° and pitch angle θ=0°, Δθ, 2Δθ, ..., 90° generates various angle combinations, where Δ Δθ is the first angular step size. For each generated angular combination... The electronic device directly substitutes this angle combination into the two-dimensional MUSIC spatial spectrum function constructed in the previous steps, and obtains the corresponding spatial spectrum value through function calculation. .

[0045] S302: The spatial spectrum value in each first angle step that is greater than the adaptive threshold and is a maximum value is taken as the first candidate peak. The azimuth and elevation angles in the angle combination corresponding to the first candidate peak are located in the corresponding preset azimuth and preset elevation angle ranges, respectively.

[0046] Specifically, the system first determines whether each spatial spectral value in each first angular step is greater than the currently set adaptive threshold, filtering out spatial spectral values ​​that exceed the threshold. Then, the electronic device performs local maxima detection on these spatial spectral values ​​that exceed the threshold, comparing them with the spatial spectral values ​​at adjacent angular positions within the first angular step's neighborhood to determine if they are local maxima. For spatial spectral values ​​that simultaneously satisfy both the conditions of being greater than the adaptive threshold and being local maxima, the electronic device marks them as first candidate peaks.

[0047] S105: Perform a second search within the preset neighborhood of each first candidate peak to obtain multiple second candidate peaks. The step size of the second search is smaller than the step size of the first search.

[0048] In this embodiment, the preset neighborhood range represents a local search area centered on the angular position corresponding to the first candidate peak. This includes an azimuth neighborhood range and a pitch neighborhood range, used to define the angle range for fine-grained searching. It is typically set to several times the size of the first angle step to ensure coverage of the possible angle range of the real target. To improve angle estimation accuracy, a second search is required. The second search refers to a refined angle search process based on the first coarse search. A smaller angle step is used to sample the neighborhood of the first candidate peak, resulting in a second candidate peak. The second candidate peak represents the angular position of the candidate target obtained in the fine-grained search stage, corresponding to the maximum value of the spatial spectrum within the neighborhood of each first candidate peak, thus achieving higher angle estimation accuracy.

[0049] Specifically, for each first candidate peak obtained in the aforementioned steps, the electronic device determines a corresponding neighborhood search range centered on its corresponding azimuth and elevation angles, forming an azimuth neighborhood range and an elevation neighborhood range. Subsequently, the electronic device performs a traversal search within each neighborhood range using a preset second angle step size, where the second angle step size is smaller than the first angle step size, typically set to 1 / 2 to 1 / 5 of the first angle step size. The electronic device generates a set of refined search angles corresponding to each first candidate peak according to the second angle step size, covering all angle combinations within the neighborhood range. For each set of refined search angles, the electronic device substitutes each angle combination into a two-dimensional MUSIC spatial spectrum function to calculate the corresponding spatial spectrum value, and selects the largest spatial spectrum value and its corresponding angle combination as the second candidate peak for that first candidate peak. By performing the above refined search process on all first candidate peaks, the electronic device obtains multiple second candidate peaks.

[0050] like Figure 3 As shown, Figure 3 This is an example diagram of the second search provided in an embodiment of this application. The diagram statistically analyzes the spatial spectral values ​​corresponding to each set of refined search angles when performing a two-dimensional traversal search of the corresponding neighborhood search range centered on the first candidate peak with a second angle step size. Figure 3 The mid-plane axis represents the pitch and azimuth angles, and the vertical axis represents the power spectrum, which represents the spatial spectrum value corresponding to each angle combination.

[0051] Based on the above embodiments, as an optional embodiment, S105: the step of performing a second search within a preset neighborhood of each first candidate peak to obtain multiple second candidate peaks may specifically include the following steps: S401: Determine the azimuth and elevation neighborhoods, respectively, with the azimuth and elevation angles corresponding to each first candidate peak as the center.

[0052] Specifically, for each first candidate peak obtained in the aforementioned steps, the electronic device extracts its corresponding azimuth and elevation angles as neighborhood centers. Based on a preset neighborhood radius parameter, the electronic device calculates the azimuth neighborhood range of the first candidate peak. The pitch angle neighborhood range is .

[0053] S402: Based on the second angle step size, traverse each azimuth angle in the azimuth angle neighborhood and traverse each pitch angle in the pitch angle neighborhood, to obtain the set of fine search angles corresponding to each first candidate peak. The second angle step size is smaller than the first angle step size.

[0054] In this embodiment, the fine search angle set represents the set of all angle combinations generated within the neighborhood of each first candidate peak according to the second angle step size. Each set contains all possible combinations of azimuth and elevation angles within the neighborhood of the first candidate peak. The second angle step size represents the uniform angle sampling interval used in the fine search phase. Its value is significantly smaller than the first angle step size to achieve higher angular resolution, and is typically set to 1 / 2 to 1 / 5 of the first angle step size. For example, when the first angle step size is 5°, the second angle step size can be set to 1° or 2°.

[0055] Specifically, the electronic device performs high-density angle sampling within the azimuth and pitch neighborhoods of each first candidate peak determined in the aforementioned steps, using a preset second angle step size. For each first candidate peak, the electronic device performs high-density angle sampling within its azimuth neighborhood. The inner traversal is performed according to the second angle step size to generate the azimuth sequence. +Second angle step size, ... Meanwhile, in the pitch angle neighborhood The system iterates through the interior using the second angle step size to generate a sequence of pitch angles. +Second angle step size, ... The electronic equipment combines all the obtained azimuth and elevation angles using Cartesian products to form the set of fine search angles corresponding to the first candidate peak. ∈ Azimuth sequence, ∈ Pitch angle sequence}.

[0056] S403: Extract the spatial spectrum value corresponding to each combination of fine search angles in each set of fine search angles from the two-dimensional MUSIC spatial spectrum function, and determine the largest spatial spectrum value in each set of fine search angles as the second candidate peak.

[0057] Specifically, the electronic device iterates through each set of refined search angles generated in the preceding steps. For each combination of refined search angles in each set, the electronic device substitutes this angle combination into the two-dimensional MUSIC spatial spectrum function to calculate the corresponding spatial spectrum value. After calculating the spatial spectrum values ​​of all angle combinations within a set of refined search angles, the electronic device selects the spatial spectrum value with the largest value as the representative result of that set, and determines this largest spatial spectrum value and its corresponding angle combination as the second candidate peak. By performing the above operation on each set of refined search angles corresponding to the first candidate peak, the electronic device finally obtains a set of second candidate peaks corresponding to the number of first candidate peaks.

[0058] S106: Determine the mirror peak corresponding to each second candidate peak according to the geometry of the multi-antenna array, and filter each second candidate peak according to each mirror peak and the first preset condition to obtain multiple third candidate peaks.

[0059] In this embodiment, a mirror peak refers to a false peak generated in the two-dimensional MUSIC spatial spectrum function due to the symmetry of the multi-antenna array geometry. These false peaks appear as high-spectral-value points similar to the real target peaks in the spatial spectrum, but there is no real signal source at the corresponding angular position. A first preset condition is used to distinguish between the real peaks and the mirror peaks. The first preset condition indicates that the average spatial spectrum value of the second candidate peak within a preset time range is greater than the average mirror spatial spectrum value of its corresponding mirror peak. After filtering by the first preset condition, a third candidate peak is obtained. The third candidate peak represents the real target peak retained after mirror peak suppression processing. These real target peaks have passed the mirror peak discrimination test and have higher reliability and accuracy.

[0060] Specifically, the electronic device first analyzes the symmetry characteristics of the multi-antenna array based on its geometric structure, determines the main symmetry axis direction of the multi-antenna array, and establishes a mirror mapping relationship based on the symmetry axis. For the angular position of each second candidate peak... Calculate its mirror angular position on the other side of the axis of symmetry. The electronic device substitutes the position of each mirror angle into the two-dimensional MUSIC spatial spectrum function to calculate the corresponding mirror spatial spectrum value, forming a set of mirror peaks that correspond one-to-one with the second candidate peak. To utilize the signal's temporal stability characteristics, the electronic device statistically analyzes spatial spectrum data at multiple time points within a preset time range after the current moment, calculating the average spatial spectrum value of each second candidate peak within that time range, as well as the average mirror spatial spectrum value of each mirror peak within the same time range. The electronic device then filters the second candidate peaks according to a first preset condition: it retains the second candidate peaks whose average spatial spectrum value is greater than the corresponding average mirror spatial spectrum value, identifying them as the third candidate peak, while excluding second candidate peaks that do not meet the condition as mirror peaks.

[0061] Based on the above embodiments, as an optional embodiment, S106: determining the mirror peak corresponding to each second candidate peak according to the geometry of the multi-antenna array, and filtering each second candidate peak according to each mirror peak and the first preset condition to obtain multiple third candidate peaks, may specifically include the following steps: S501: Determine the axis of symmetry based on the geometry of the multi-antenna array, and calculate the mirror angle position corresponding to each second candidate peak based on the axis of symmetry.

[0062] Specifically, the electronic device determines its multiple axes of symmetry based on the geometric characteristics of the circular array, including an axis of symmetry passing through the center of the array and parallel to the x-axis and y-axis, and a diagonal axis of symmetry at a 45° angle to the coordinate axes. Based on the geometric relationships of these axes of symmetry, the electronic device selects each second candidate peak... Establish a complete set of mirror angle constraints. The electronic device calculates all possible mirror angle positions according to the mirror transformation formula: azimuth mirrors include... (Symmetric about a 90° axis) (Symmetric about 0° axis) (Symmetric about the origin) etc.; Pitch angle mirroring includes , , The electronic device calculates all mirror angle positions of each second candidate peak under each axis of symmetry, forming a candidate set of mirror angle positions for that second candidate peak.

[0063] S502: Extract the mirror space spectrum values ​​corresponding to each mirror angle position from the two-dimensional MUSIC space spectrum function, and determine each mirror space spectrum value as a mirror peak.

[0064] Specifically, the electronic device substitutes the calculated mirror angle positions one by one into the constructed two-dimensional MUSIC spatial spectrum function, and obtains the spatial spectrum value corresponding to each mirror angle position through function calculation. The electronic device extracts the mirror spatial spectrum values ​​at each position according to the coordinate order of the mirror angle positions, forming a sequence of mirror spatial spectrum values ​​that correspond one-to-one with the second candidate peak. For each possible mirror angle position corresponding to the second candidate peak, the electronic device calculates the spatial spectrum value for each mirror angle position, establishing a mapping relationship between the second candidate peak and all its possible mirror spatial spectrum values. The electronic device labels and classifies the extracted mirror spatial spectrum values, identifying the mirror spatial spectrum value with a maximum value as the mirror peak.

[0065] S503: Calculate the average spatial spectrum value of each second candidate peak within a preset time range after the current time, and the average mirror spatial spectrum value of the mirror peak corresponding to each second candidate peak.

[0066] Specifically, the electronic device sets a preset time range [t0, t0+ΔT] starting from the current time t0, where ΔT is the preset time window length. Within this time range, the electronic device performs multiple MUSIC spatial spectrum calculations at fixed sampling intervals to obtain the spatial spectrum function values ​​on the time series {t1, t2, ..., tn}. For each second candidate peak, the electronic device extracts its spatial spectrum value at each time point and calculates the average spatial spectrum value of the second candidate peak over the entire time range. Simultaneously, the electronic device performs the same time averaging calculation on the mirror peaks corresponding to each second candidate peak, extracts the mirror spatial spectrum values ​​of each mirror peak on the time series {t1, t2, ..., tn}, and calculates the average mirror spatial spectrum value of each mirror peak.

[0067] S504: The second candidate peak whose average spatial spectral value is greater than the average mirror spatial spectral value is determined as the third candidate peak.

[0068] Specifically, the electronic device performs a numerical comparison between the average spatial spectral value and the average mirror spatial spectral value for each second candidate peak. The electronic device extracts the average spatial spectral value of each second candidate peak one by one, and simultaneously obtains the average mirror spatial spectral value of all mirror peaks corresponding to that second candidate peak. The electronic device selects the maximum average mirror spatial spectral value corresponding to that second candidate peak as the comparison benchmark, and executes a judgment condition: if the average spatial spectral value of the second candidate peak is greater than its corresponding maximum average mirror spatial spectral value, the electronic device marks that second candidate peak and determines it as the third candidate peak.

[0069] S107: Perform a third search on each third candidate peak within the preset angle range of each third candidate peak, and determine the third candidate peak that meets the second preset condition as the target peak.

[0070] In this embodiment of the application, the preset angle range represents the angle search area that expands outward from the position of the third candidate peak, including the preset azimuth angle superposition range and the preset pitch angle superposition range.

[0071] Specifically, the electronic device uses the azimuth and elevation angles of each third candidate peak as its center coordinates, and traverses all fine-search angle combinations within a preset azimuth and elevation angle superposition range using a fine step size. For each fine-search angle combination, the electronic device extracts the corresponding spatial spectrum value from the two-dimensional MUSIC spatial spectrum function, and sums all spatial spectrum values ​​within the preset angle range of the third candidate peak to obtain its total power value. Utilizing the physical characteristic that real signal sources generally have high power distributions in the angular regions around them, while false peaks only have isolated high values, the electronic device judges the authenticity of the third candidate peak by the magnitude of the total power value. Based on the total power value of the third candidate peak, the electronic device identifies the third candidate peak that meets the second preset condition as the target peak.

[0072] Based on the above embodiments, as an optional embodiment, S107: the step of performing a third search on each third candidate peak within a preset angle range of each third candidate peak, and determining the third candidate peak that meets the second preset condition as the target peak, may specifically include the following steps: S601: Using the azimuth and elevation angles corresponding to each third candidate peak as the center, traverse each fine search angle combination within the preset azimuth and elevation angle superposition ranges.

[0073] Specifically, for each third candidate peak, the electronic device extracts its corresponding azimuth and elevation angles as the search center coordinates. Based on preset azimuth and elevation angle stacking ranges, the electronic device determines the angle search area for that third candidate peak as an azimuth and elevation angle range. Within the determined angle search area, the electronic device performs a gridded traversal according to a preset fine search step size, generating all possible fine search angle combinations. Starting from the beginning of the azimuth range, the electronic device increments the preset fine search step size to the end of the azimuth range; simultaneously, starting from the beginning of the elevation range, it increments the preset fine search step size to the end of the elevation range, combining each azimuth and elevation angle value using a Cartesian product to form a complete set of fine search angle combinations.

[0074] S602: Extract the spatial spectrum values ​​corresponding to each combination of fine search angles from the two-dimensional MUSIC spatial spectrum function, and superimpose the spatial spectrum values ​​to obtain the total power value corresponding to each third candidate peak.

[0075] Specifically, for each third candidate peak, the electronic device processes all precise search angle combinations one by one. By substituting the azimuth and elevation angles of each angle combination into the two-dimensional MUSIC spatial spectrum function, the corresponding spatial spectrum value is extracted. The electronic device then superimposes all spatial spectrum values ​​of the third candidate peak within a preset angle range according to the third formula: ,in, This represents the total power value corresponding to the third candidate peak. The azimuth angle within the i-th angle combination in the preset angular neighborhood of the third candidate peak. The pitch angle within the i-th angle combination in the preset angle neighborhood of the third candidate peak is determined. Let be the spatial spectrum value corresponding to the i-th angle combination within the preset angle neighborhood of the third candidate peak, k and K are the preset angle neighborhood parameters of the third candidate peak in azimuth and elevation angles, respectively, m and M are the neighborhood range parameters of the mirror peak in azimuth and elevation angles, and j is the index of the angle combination within the preset angle neighborhood of the mirror peak. The outer summation iterates through K search points in the azimuth direction, and the inner summation iterates through M search points in the elevation direction. The total power value of the third candidate peak is obtained by summing these parameters.

[0076] S603: Based on the total power values, perform a maximum peak point search for each third candidate peak, and determine the third candidate peak that meets the second preset condition as the fourth candidate peak. The second preset condition is that the spatial spectral value corresponding to the third candidate peak is the maximum value.

[0077] Specifically, the electronic device performs a maximum search on each total power value, identifying local or global maxima by comparing the magnitudes of the total power values. The electronic device sets a maximum judgment criterion, marking the third candidate peak whose total power value reaches a local maximum among neighboring candidate peaks or a global maximum among all candidate peaks as satisfying a second preset condition. The electronic device traverses all third candidate peaks, checking whether their total power value satisfies the maximum condition for each peak. The candidate peaks that meet the criteria are extracted from the third candidate peak set and relabeled as the fourth candidate peaks.

[0078] S604: Perform trajectory smoothing on each fourth candidate peak to obtain the target peak.

[0079] Specifically, the electronic device collects the angular position sequences of each fourth candidate peak in multiple consecutive time snapshots, constructing the azimuth time series of each candidate peak. } and pitch angle time series { } where k represents the time index. The electronic device sets the sliding window length T, typically set to 10 time points, to ensure a balance between smoothing effect and response speed. For the current time t, the electronic device extracts T historical angle values ​​from time tT-1 to t, and calculates the arithmetic mean of the azimuth and pitch angles as the smoothed angle estimate. The electronic device follows... Calculate the smoothed azimuth angle according to Calculate the smoothed pitch angle. The electronic device will use the angle coordinates after trajectory smoothing ( , ( ) as the target peak.

[0080] The following describes an exemplary multi-peak search device based on two-dimensional adaptive threshold provided in the embodiments of this application. Figure 4 This is an exemplary hardware structure diagram of a multi-peak search device based on two-dimensional adaptive threshold provided in an embodiment of this application.

[0081] In some embodiments, the multi-peak search device based on a two-dimensional adaptive threshold is a computer device or includes a computer device. The computer device includes a processor, memory, and a network interface connected via a system bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device stores data. The network interface of the computer device is used to communicate with other external terminals or servers via a network connection. In some embodiments, the network interface can be a wired network interface; in some embodiments, the network interface can also be a wireless network interface. When the computer program is executed by the processor, it implements the methods in the embodiments of this application.

[0082] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0083] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0084] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0085] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.

[0086] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A multi-peak search method based on two-dimensional adaptive threshold, characterized in that, The method includes: Receive the signal data matrix about the target transmitted by the multi-antenna array at the current moment, and calculate the sample covariance matrix based on the signal data matrix; The sample covariance matrix is ​​decomposed into eigenvalues ​​and eigenvectors, and a noise subspace matrix is ​​constructed based on the eigenvectors. Calculate the noise power estimate based on each of the aforementioned eigenvalues, and calculate the adaptive threshold based on the noise power estimate; A two-dimensional MUSIC spatial spectrum function is constructed based on the noise subspace matrix. The spatial spectrum value of the two-dimensional MUSIC spatial spectrum function is searched for the first time. The spatial spectrum value that is greater than the adaptive threshold is taken as the first candidate peak. The two-dimensional MUSIC spatial spectrum function is the correspondence between each spatial spectrum value and the azimuth and elevation angle combination corresponding to each spatial spectrum value. A second search is performed within a preset neighborhood of each of the first candidate peaks to obtain multiple second candidate peaks. The step size of the second search is smaller than the step size of the first search. The mirror peaks corresponding to each second candidate peak are determined according to the geometric structure of the multi-antenna array, and each second candidate peak is filtered according to each mirror peak and the first preset condition to obtain multiple third candidate peaks; A third search is performed on each of the third candidate peaks within a preset angle range, and the third candidate peak that meets the second preset condition is determined as the target peak.

2. The multi-peak search method based on two-dimensional adaptive threshold according to claim 1, characterized in that, The step of calculating a noise power estimate based on each of the aforementioned feature values, and calculating an adaptive threshold based on the noise power estimate, includes: The noise power estimate is calculated based on the first formula; The first formula is: ; in, Here, N is the noise power estimate, N is the number of elements in the multi-antenna array, and M is the amount of signal data. Let be the i-th eigenvalue in the noise subspace; The adaptive threshold is calculated based on the second formula; The second formula is: ; Where T is the adaptive threshold. K is the noise power estimate, and K is the total number of search points, which is determined by the search step size of the azimuth angle and the pitch angle.

3. The multi-peak search method based on two-dimensional adaptive threshold according to claim 1, characterized in that, The first search of the spatial spectral values ​​of the two-dimensional MUSIC spatial spectral function, selecting spatial spectral values ​​greater than the adaptive threshold as first candidate peaks, includes: Search for the spatial spectrum value in the two-dimensional MUSIC spatial spectrum function at each of the first angle steps according to the preset first angle step size; The spatial spectrum value in each of the first angle steps that is greater than the adaptive threshold and is a maximum value is taken as the first candidate peak. The azimuth and elevation angles in the angle combination corresponding to the first candidate peak are respectively located in the corresponding preset azimuth and preset elevation angle ranges.

4. The multi-peak search method based on two-dimensional adaptive threshold according to claim 3, characterized in that, The second search is performed within a preset neighborhood of each of the first candidate peaks to obtain multiple second candidate peaks, including: The azimuth and elevation angles corresponding to each of the first candidate peaks are used as centers to determine the azimuth and elevation angle neighborhood ranges, respectively. Based on the second angle step size, each azimuth angle is traversed within the azimuth angle neighborhood, and each pitch angle is traversed within the pitch angle neighborhood, to obtain the set of fine search angles corresponding to each first candidate peak. The second angle step size is smaller than the first angle step size. Extract the spatial spectrum value corresponding to each combination of fine search angles in each set of fine search angles from the two-dimensional MUSIC spatial spectrum function, and determine the largest spatial spectrum value in each set of fine search angles as the second candidate peak.

5. The multi-peak search method based on two-dimensional adaptive threshold according to claim 1, characterized in that, The process involves determining the mirror peaks corresponding to each second candidate peak based on the geometry of the multi-antenna array, and then filtering each second candidate peak based on each mirror peak and a first preset condition to obtain multiple third candidate peaks, including: The axis of symmetry is determined based on the geometry of the multi-antenna array, and the mirror angle position corresponding to each second candidate peak is calculated based on the axis of symmetry. Extract the mirror space spectrum value corresponding to each mirror angle position from the two-dimensional MUSIC space spectrum function, and determine each mirror space spectrum value as a mirror peak; Calculate the average spatial spectrum value of each second candidate peak and the average mirror spatial spectrum value of the mirror peak corresponding to each second candidate peak within a preset time range after the current time. The second candidate peak whose average spatial spectral value is greater than the average mirror spatial spectral value is determined as the third candidate peak.

6. The multi-peak search method based on two-dimensional adaptive threshold according to claim 4, characterized in that, The step of performing a third search on each of the third candidate peaks within a preset angle range, and determining the third candidate peak that satisfies the second preset condition as the target peak, includes: Centered on the azimuth and elevation angles corresponding to each of the third candidate peaks, each of the fine search angle combinations is traversed within the preset azimuth and elevation angle superposition ranges. Extract the spatial spectrum values ​​corresponding to each of the fine search angle combinations from the two-dimensional MUSIC spatial spectrum function, and superimpose the spatial spectrum values ​​to obtain the total power value corresponding to each of the third candidate peaks; Based on the total power values, a maximum peak value search is performed on each of the third candidate peaks, and the third candidate peak that meets the second preset condition is determined as the fourth candidate peak. The second preset condition is that the spatial spectral value corresponding to the third candidate peak is the maximum value. The trajectory smoothing process is performed on each of the fourth candidate peaks to obtain the target peak.

7. The multi-peak search method based on two-dimensional adaptive threshold according to claim 6, characterized in that, The step of superimposing the spatial spectral values ​​to obtain the total power value corresponding to each of the third candidate peaks includes: The total power value corresponding to each of the third candidate peaks is calculated based on the third formula. The third formula is: ; in, This represents the total power value corresponding to the third candidate peak. The azimuth angle within the i-th angle combination in the preset angular neighborhood of the third candidate peak. The pitch angle within the i-th angle combination in the preset angle neighborhood of the third candidate peak is determined. Let be the spatial spectrum value corresponding to the i-th angle combination within the preset angle neighborhood of the third candidate peak, k and K are the preset angle neighborhood parameters of the third candidate peak in azimuth and elevation angles, respectively, m and M are the neighborhood range parameters of the mirror peak in azimuth and elevation angles, and j is the index of the angle combination within the preset angle neighborhood of the mirror peak.

8. A multi-peak search device based on two-dimensional adaptive threshold, characterized in that, The two-dimensional adaptive threshold-based multi-peak search device includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the two-dimensional adaptive threshold-based multi-peak search device to perform the method as described in any one of claims 1-7.

9. A computer program product containing instructions, characterized in that, When the computer program product is run on a two-dimensional adaptive threshold multi-peak search device, the two-dimensional adaptive threshold multi-peak search device performs the method as described in any one of claims 1-7.

10. A computer-readable storage medium comprising instructions, characterized in that, When the instruction is executed on a two-dimensional adaptive threshold multi-peak search device, the two-dimensional adaptive threshold multi-peak search device performs the method as described in any one of claims 1-7.

Images