A uniform circular array one-dimensional spatial spectrum direction finding signal processing device
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHONGQING HUILING ELECTRONIC NEW TECH CO LTD
- Filing Date
- 2023-01-03
- Publication Date
- 2026-06-26
AI Technical Summary
In existing technologies for radio monitoring, the MUSIC algorithm for uniform circular arrays makes signal quantity estimation errors, leading to spatial signal measurement errors. The Davis transform cannot effectively complete coherent signal direction finding over a wide frequency band, and the DML algorithm has a large computational load, making it difficult to complete DOA estimation in a short time.
By combining the MUSIC and Capon algorithms, and through spatial spectrum search and iterative calculation, the DML algorithm is used to process coherent signals, achieving accurate DOA calculation for incoherent signals. Furthermore, coherent signal direction finding is completed based on incoherent signal direction finding.
It enables accurate calculation of DOA for incoherent signals in radio monitoring and can complete direction finding of coherent signals in a short time, reducing the amount of calculation and error.
Smart Images
Figure CN116125372B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of spatial spectrum direction finding in radio monitoring, and in particular to a uniform circular array one-dimensional spatial spectrum direction finding signal processing device. Background Technology
[0002] In the field of radio monitoring, uniform circular arrays are commonly used for spatial spectrum direction finding. The general direction finding algorithm is the super-resolution MUSIC algorithm, which can achieve direction finding of incoherent signals. However, the MUSIC algorithm requires pre-calculation of the number of arriving signals. Common signal quantity estimation algorithms, such as those based on information theory criteria, often encounter errors in signal quantity estimation during practical use, leading to errors in spatial signal measurement. Common beamforming transformation methods convert uniform circular arrays into virtual uniform linear arrays to achieve coherent signal direction finding. However, due to the wide monitoring frequency bands and relatively small number of array elements in radio monitoring, the Davis transform cannot effectively complete coherent signal direction finding across the entire frequency band. The DML algorithm requires multi-dimensional nonlinear search; if all signals are searched using the DML algorithm, the computational load becomes extremely large, making it difficult to complete direction finding (DOA) estimation within a short timeframe in practical engineering implementations. Summary of the Invention
[0003] In view of at least one deficiency of the prior art, the purpose of this invention is to provide a uniform circular array one-dimensional spatial spectrum direction-finding signal processing device. This device combines the MUSIC algorithm and the Capon algorithm for calculation, and accurately calculates the DOA of incoherent signals through spatial spectrum search and iterative calculation. Based on the direction finding of incoherent signals, the DML algorithm can be used to complete the direction finding of a pair of coherent signals.
[0004] To achieve the above objectives, the present invention adopts the following technical solution: a uniform circular array one-dimensional spatial spectrum direction finding signal processing device, comprising a data processing and covariance matrix module, a MUSIC algorithm processing module, a Capon algorithm processing module, a processing result fusion module, and a DML algorithm processing module;
[0005] The data processing and covariance matrix module is used to preprocess the received signal and calculate the covariance matrix.
[0006] The MUSIC algorithm processing module implements super-resolution spatial spectrum direction finding, calculates the MUSIC spatial spectrum of the receiving array using the input covariance matrix, and inputs the MUSIC spatial spectrum calculation result into the processing result fusion module.
[0007] The Capon algorithm processing module uses a pre-set threshold method and the Capon algorithm to achieve spatial spectrum direction finding. The Capon spatial spectrum of the receiving array is calculated from the covariance matrix, and the Capon spatial spectrum calculation result is input into the processing result fusion module.
[0008] The processing result fusion module analyzes and judges the calculation results of MUSIC spatial spectrum and Capon spatial spectrum to determine whether iterative search is needed to obtain the direction finding results of incoherent signals and to determine whether coherent signals are included.
[0009] The DML algorithm processing module calculates the DOA of a pair of coherent signals and outputs the direction finding results of the space signal. If the processing and fusion module determines that there is a coherent signal in the incoming wave signal, it obtains the DOA of the incoherent signal as prior information and calculates the DOA of a pair of coherent signals.
[0010] The data processing and covariance matrix module performs digital down-conversion, filtering, and correlation processing on the received signal to form a covariance matrix. R X Let X represent the received data covariance matrix, and let X represent the received signal data matrix. H Let E[ be the conjugate transpose of X.] ] represents the mathematical expectation.
[0011] The MUSIC algorithm processing module processes R. X Eigenvalue decomposition yields eigenvalues and eigenvectors. The number of incoming waves is calculated using the eigenvalues, and the eigenvectors are decomposed into signal subspaces according to the magnitude of the eigenvalues and the number of incoming waves. With noise subspace ; for the incident azimuth angle of the signal The search yielded the MUSIC spatial spectrum.
[0012] Follow the function below The search yielded the MUSIC spatial spectrum:
[0013] ;
[0014] It is the azimuth angle of the signal incident. The corresponding array steering vector, for a uniform circular array, The expression is:
[0015] ;
[0016] Where R represents the radius of the circular array; M represents the number of array elements; .
[0017] Capon algorithm processing module for R X Inverse matrix for signal incident azimuth angle The search yielded the Capon space spectrum.
[0018] Follow the function below The search yielded the Capon space spectrum:
[0019] .
[0020] The processing result fusion module compares and analyzes the MUSIC and Capon spectral results, and performs mode determination on the MUSIC and Capon spatial spectral results. If the match is poor, the number of incoming waves is changed, and the MUSIC spatial spectrum is iteratively searched. If the match is good, it determines whether coherent signals are included and outputs the DOA detection results for incoherent signals. + Substituting into the Capon spatial spectrum formula, ,in, For noise power, It is R X A diagonal matrix consisting of D large eigenvalues, where D is the number of incident signals in the array, i.e.: ;Will Substituting the spectral function into the Capon space spectral formula: .
[0021] If the incoming wave contains a coherent signal, the DML algorithm measures the incident angle of the coherent signal; matrix It is the array steering matrix of D incident signals Then the guiding matrix projection matrix Represented as For a steering matrix composed of multiple signals, the measured incoherent signal DOA is used as prior information, and the coherent signals are processed according to... Formula search yields the incident angle of the coherent signal. Represents the trace of a matrix; The angle corresponding to the maximum value is the incident angle of the coherent signal.
[0022] The specific working logic of the result fusion module is as follows:
[0023] Results fusion and iterative search of the music spatial spectrum:
[0024] Comparing the peak search results of MUSIC and Capon spectra, pattern identification and iterative search are performed on the search results of MUSIC and Capon spectra.
[0025] Pattern judgment
[0026] Mode 1: Missed Detection in MUSIC. The estimated signal quantity is too low, leading to spectral search errors and missed detections. Increase the estimated signal quantity value and search the MUSIC spatial spectrum.
[0027] Specifically: if a MUSIC spectral peak is missed, the number of incoming waves is increased and the MUSIC spectral peak search is performed again.
[0028] The estimated number of signals is updated, and after iterative search, the correct incoherent signal detection result is obtained.
[0029] Mode 2: Coherent signal identification and spurious peak removal.
[0030] When a coherent signal exists in space: the peak value of the corresponding MUSIC spectrum is similar to that of the Capon spectrum, and the peak is a spurious peak. By removing the spurious peak of the coherent signal, the detection result of the incoherent signal is obtained.
[0031] It can accurately measure the DOA of incoherent signals and determine whether the incoming signal contains a coherent signal.
[0032] If the peak values of the MUSIC spectrum and the Capon spectrum are approximately equal, then the peak is a pseudo-peak formed by the coherent signal.
[0033] Next, the DML algorithm processing module processes a pair of coherent signals for DOA calculation and outputs the spatial direction finding results of the coherent signals.
[0034] If the processing and fusion module determines that there is a coherent signal in the incoming wave signal, then the false peaks of the coherent signal are removed, and the DML algorithm can be used to implement DOA measurement of a pair of coherent signals.
[0035] Obtain the incoherent signal DOA as prior information, and calculate a pair of coherent signals DOA.
[0036] Significant Effects: This invention provides a one-dimensional spatial spectrum direction-finding signal processing device for a uniform circular array. This device combines the MUSIC algorithm and the Capon algorithm for calculation, accurately calculating the DOA of incoherent signals through spatial spectrum search and iterative calculation. Based on the results of the MUSIC and Capon algorithms, it achieves DOA estimation for incoherent signals. Furthermore, based on the direction-finding of incoherent signals, it can complete the direction-finding of a pair of coherent signals. Attached Figure Description
[0037] Figure 1 Here is a flowchart of the spatial spectrum orientation finding process for a one-dimensional uniform circular array;
[0038] Figure 2 The MUSIC spatial spectrum when a signal is missed;
[0039] Figure 3 For Capon space spectrum;
[0040] Figure 4 The MUSIC spatial spectrum after iterative search;
[0041] Figure 5The MUSIC spatial spectrum contains a pair of coherent signals;
[0042] Figure 6 It is the Capon spatial spectrum containing a pair of coherent signals. Detailed Implementation
[0043] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.
[0044] like Figures 1-6 As shown, a uniform circular array one-dimensional spatial spectrum direction finding signal processing device is used. The direction finding algorithm is mainly applied in the field of spatial spectrum direction finding for radio monitoring. It adopts a 9-element uniform circular array and uses a signal processing scheme that combines multiple spatial spectrum direction finding algorithms such as Multiple Signal Classification (MUSIC), Capon and Deterministic Maximum Likelihood (DML) to realize the direction finding of multiple incoherent signals and can also realize the direction finding of a pair of coherent signals.
[0045] The technical solution includes five processing modules: data processing and covariance matrix module, MUSIC algorithm processing module, Capon algorithm processing module, processing result fusion module, and DML algorithm processing module.
[0046] The data processing and covariance matrix module preprocesses the received signal and calculates the covariance matrix. This mainly includes digital down-conversion, filtering, and covariance matrix calculation. The generated covariance data can be input into the MUSIC algorithm processing module, the Capon algorithm processing module, and the DML algorithm processing module.
[0047] The MUSIC algorithm processing module implements super-resolution spatial spectrum direction finding. It calculates the spatial spectrum of the receiving array using the input covariance matrix and inputs the spatial spectrum calculation results into the processing result fusion module. This mainly includes covariance eigenvalue decomposition, improved MDL criterion for calculating the signal quantity, and MUSIC spatial spectrum search.
[0048] The Capon algorithm processing module employs a pre-set threshold method and the Capon algorithm to achieve spatial spectrum direction finding. Similar to the MUSIC algorithm, the spatial spectrum of the receiving array is calculated from the covariance matrix, and the spatial spectrum calculation result is input into the processing result fusion module. This mainly includes covariance matrix inversion, Capon spatial spectrum search, and threshold filtering.
[0049] The result fusion module analyzes and judges the calculation results of the MUSIC spatial spectrum and Capon spatial spectrum to determine whether iterative search is needed to obtain the direction finding results of incoherent signals, and to determine whether coherent signals are included. This mainly includes iterative search using the MUSIC algorithm, removal of spurious peaks in the spatial spectrum, and simultaneous determination of whether the received signal contains coherent signals. If no coherent signals are included, this module directly outputs the direction finding results.
[0050] The DML algorithm processing module can calculate the DOA of a pair of coherent signals and output the direction-finding results of the space signal. If the processing and fusion module determines that there is a coherent signal in the incoming wave signal, the DOA of the incoherent signal is input into this module as prior information, and the module calculates the DOA of a pair of coherent signals. This mainly includes the initialization of the incoherent signal DOA and the two-dimensional DML search.
[0051] Working principle
[0052] The uniform circular array spatial spectrum direction finding mainly consists of a data processing and covariance matrix module, a MUSIC algorithm processing module, a Capon algorithm processing module, a processing result fusion module, and a DML algorithm processing module.
[0053] The data processing and covariance matrix module works by digitally down-converting and filtering the received signal, followed by correlation processing to form a covariance matrix. R X Let X represent the received data covariance matrix, and let X represent the received signal data matrix. H Let E[ be the conjugate transpose of X.] ] represents the mathematical expectation.
[0054] The MUSIC algorithm processing module processes R... X Eigenvalue decomposition yields eigenvalues and eigenvectors. The number of incoming waves is calculated using the eigenvalues, and the eigenvectors are decomposed into signal subspaces according to the magnitude of the eigenvalues and the number of incoming waves. With noise subspace Follow the function below to... The search yielded the MUSIC spatial spectrum:
[0055] ;
[0056] It is the azimuth angle of the signal incident. The corresponding array steering vector, for a uniform circular array, The expression is:
[0057] ;
[0058] Where R represents the radius of the circular array; M represents the number of array elements; .
[0059] Capon algorithm processing module for R X To find the inverse of a matrix, follow the function below. The search yielded the Capon space spectrum:
[0060] ;
[0061] The processing result fusion module compares and analyzes the MUSIC and Capon spectral results, and performs mode determination on the MUSIC spectrum and Capon results. If the match is poor, the number of incoming waves is changed, and the MUSIC spatial spectrum is iteratively searched. If the match is good, it determines whether coherent signals are included and outputs the DOA detection results for incoherent signals. + Substituting into the Capon spatial spectrum formula, ,in, For noise power, It is R X A diagonal matrix consisting of D large eigenvalues, where D is the number of incident signals in the array, i.e.: ;Will Substituting the spectral function into the Capon space spectral formula: ;
[0062] If the incoming wave contains a coherent signal, the DML algorithm measures the angle of incidence of the coherent signal. Matrix It is the array steering matrix of D incident signals Then the guiding matrix projection matrix Represented as For a steering matrix composed of multiple signals, the measured incoherent signal DOA is used as prior information, and the coherent signals are processed according to... Formula search yields the incident angle of the coherent signal. Represents the trace of a matrix.
[0063] Technical effects of this patent
[0064] By performing mode identification and iterative search using MUSIC and Capon spectrum search results, we can accurately measure the DOA of incoherent signals, determine whether the incoming signal contains coherent signals, remove spurious peaks of coherent signals, and use the DML algorithm to measure the DOA of a pair of coherent signals.
[0065] (1) Iterative search for detecting incoherent signals
[0066] Mode 2: Missed Detections in MUSIC. An underestimated signal quantity leads to spectral search errors and missed detections. Increase the estimated signal quantity value and search the MUSIC spatial spectrum.
[0067] Simulation example: Uniform circular array, 9 elements, 5 incident signals, frequency: 300MHz, angles: (25°, 50°, 75°, 100°, 125°), signal-to-noise ratio 10dB, spectral peak search step is 1°.
[0068] The results are as follows Figure 2 and Figure 3 As shown.
[0069] The estimated signal quantity is updated, and after iterative search, the correct incoherent signal detection result is obtained. The result is as follows: Figure 4 As shown. A total of 5 incoherent incident signals were measured, with angles of (25°, 50°, 76°, 101°, and 125°).
[0070] (2) Mode 2: Coherent signal judgment and spurious peak removal. When a coherent signal exists in space, the corresponding MUSIC spectral peak and Capon spectral peak are similar in value, and the spectral peak is a spurious peak. The spurious peak of the coherent signal is removed to obtain the non-coherent signal detection result.
[0071] Simulation example: Uniform circular array, 9 elements, incident signal frequency: 300MHz, angle (25° 50° 75° 100° 125°), signal-to-noise ratio 10dB, where signal 1 and signal 2 are coherent.
[0072] By analyzing the MUSIC and Capon spectral peaks, the incident signal contained coherent signals. The locations of these coherent signals were determined, yielding incoherent detection results: three incoherent signals with incident angles of (75°, 100°, and 125°), and coherent signals were also observed. The results are as follows... Figure 5 and Figure 6 As shown.
[0073] (3) Coherent signal detection
[0074] The DML algorithm is a multidimensional nonlinear search with a very high computational cost. Currently, this algorithm can be used to detect a pair of coherent signals. According to ( Initialize the array steering matrix (75°, 100°, 125°), where , The incident angle of the coherent signal to be calculated. After a two-dimensional search, the detection results can be obtained: 5 signals, incident angles (25°, 50°, 75°, 100°, 125°), and the first and second signals are coherent signals.
[0075] The procedure for spatial spectrum orientation finding of a one-dimensional uniform circular array is as follows: Figure 1 As shown.
[0076] (1) Data processing, covariance matrix
[0077] Correlation processing is performed on the multi-shot received signal X(t), and the covariance matrix is: ,in It is the received signal data vector. It is the conjugate transpose matrix, E[ [] represents the mathematical expectation.
[0078] Eigenvalue decomposition yields , yes eigenvectors, yes The conjugate transpose of the matrix. Represents the eigenvalue. Based on the eigenvalue... Calculate the number of incoming signals, then It can be represented as: , This represents the eigenvector corresponding to the incident signal. These are the characteristic values of the incident signal; It is the feature vector corresponding to the noise. It is a noise characteristic value.
[0079] The inverse can be obtained + .
[0080] (2) MUSIC algorithm processing
[0081] Azimuth Up, 0.1° search step, search according to the function below. The search yielded the MUSIC spatial spectrum:
[0082] ;
[0083] It is the azimuth angle of the signal incident. The corresponding array steering vector, for a uniform circular array, The expression is:
[0084] ;
[0085] Where R represents the radius of the circular array; M represents the number of array elements; .
[0086] 3) Capon algorithm processing
[0087] The Capon algorithm processing module processes the covariance R. X Matrix inversion, with a search step of 0.1°, follows the function below. The search yielded the Capon space spectrum:
[0088] 4) Processing result fusion and iterative search of the music spatial spectrum
[0089] Compare the MUSIC and Capon spectral peak search results. If a MUSIC spectral peak is missed, increase the number of incoming waves and re-search for the MUSIC spectral peak. If the peak values of the MUSIC and Capon spectral peaks are approximately equal, then the peak is a spurious peak formed by a coherent signal.
[0090] If the incoming signal does not contain a coherent signal, the direction finding result of the MUSIC search will be output; if the incoming signal contains a coherent signal, the incoherent signal DOA will be used as a priori signal, and the DML algorithm will be used to measure the incoming signal DOA.
[0091] (5) DML measurement of coherent signal DOA
[0092] Array guiding matrix projection matrix Represented as . It is a steering matrix formed by multiple signals in space. The incoherent signal DOA obtained in step (4) is used as prior information and substituted into the system. In the middle; the angle of paired coherent signals ( , ), and In two-dimensional space, for angles ( , )according to Search for formulas. Represents the trace of a matrix. The angle corresponding to the maximum value is the incident angle of the coherent signal.
[0093] Finally, it should be noted that the above are only specific embodiments of the present invention. Of course, those skilled in the art can make modifications and variations to the present invention. If these modifications and variations fall within the scope of the claims of the present invention and their equivalents, they should be considered as being within the protection scope of the present invention.
Claims
1. A uniform circular array one-dimensional spatial spectrum direction finding signal processing device, comprising a data processing and covariance matrix module, a MUSIC algorithm processing module, a Capon algorithm processing module, a MUSIC and Capon processing result fusion module, and a DML algorithm processing module; The data processing and covariance matrix module is used to preprocess the received signal and calculate the covariance matrix. The MUSIC algorithm processing module implements super-resolution spatial spectrum direction finding, calculates the MUSIC spatial spectrum of the receiving array using the input covariance matrix, and inputs the MUSIC spatial spectrum calculation result into the processing result fusion module. The Capon algorithm processing module uses a pre-set threshold method and the Capon algorithm to achieve spatial spectrum direction finding. The Capon spatial spectrum of the receiving array is calculated from the covariance matrix, and the Capon spatial spectrum calculation result is input into the processing result fusion module. The processing result fusion module will + Substituting into the Capon spatial spectrum formula, ,in, For noise power, It is R X A diagonal matrix consisting of D large eigenvalues, where D is the number of incident signals in the array, i.e.: ;Will Substituting the spectral function into the Capon space spectral formula: ; U s For the signal subspace, U n For the noise subspace, α(θ) i θ is the incident azimuth angle of the signal. i The corresponding array steering vector, , ,..., They are all eigenvalues; The processing result fusion module analyzes and judges the calculation results of the MUSIC spatial spectrum and the Capon spatial spectrum to determine whether iterative search is needed to obtain the direction finding results of incoherent signals and to determine whether coherent signals are included. The DML algorithm processing module calculates the DOA of a pair of coherent signals and outputs the spatial direction finding results of the coherent signals. If the processing and fusion module determines that there is a coherent signal in the incoming wave signal, it obtains the DOA of the incoherent signal as prior information and calculates the DOA of a pair of coherent signals. The data processing and covariance matrix module performs digital down-conversion, filtering, and correlation processing on the received signal to form a covariance matrix. R X Let X represent the received data covariance matrix, and let X represent the received signal data matrix. H Let E[ be the conjugate transpose of X.] [] represents the expected value; Results fusion and iterative search of the music spatial spectrum: The specific working logic of the result fusion module is as follows: Comparing the peak search results of MUSIC and Capon spectra, pattern identification and iterative search are performed on the search results of MUSIC and Capon spectra. Pattern judgment Mode 1: MUSIC Missed Detection: The estimated signal quantity is too small, leading to spectral search errors and missed detections. Increase the estimated signal quantity value and search the MUSIC spatial spectrum. Specifically: if the MUSIC spectral peaks are missed, the number of incoming waves is increased, the MUSIC spectral peak search is performed again, the signal quantity estimate is updated, and after iterative search, the correct incoherent signal detection result is obtained; Mode 2: Coherent signal judgment and spurious peak removal: When a coherent signal exists in space, the corresponding MUSIC spectral peak and Capon spectral peak have similar peak values, and the spectral peak is a spurious peak; by removing the coherent signal spurious peak, the non-coherent signal detection result is obtained, which can accurately realize the DOA measurement of non-coherent signals; Determine whether the incoming signal contains a coherent signal: If the peak values of the MUSIC spectrum and the Capon spectrum are approximately equal, then the peak is a pseudo-peak formed by the coherent signal. The DML algorithm processing module processes a pair of coherent signals for DOA calculation and outputs the spatial direction finding results of the coherent signals: If the processing and fusion module determines that there is a coherent signal in the incoming wave signal, then the false peaks of the coherent signal are removed, and the DML algorithm can be used to implement DOA measurement of a pair of coherent signals. Obtain the incoherent signal DOA as prior information, and calculate a pair of coherent signals DOA.
2. The uniform circular array one-dimensional spatial spectrum direction-finding signal processing device according to claim 1, characterized in that: The MUSIC algorithm processing module processes R. X Eigenvalue decomposition yields eigenvalues and eigenvectors. The number of incoming waves is calculated using the eigenvalues, and the eigenvectors are decomposed into signal subspaces according to the magnitude of the eigenvalues and the number of incoming waves. With noise subspace ; for the incident azimuth angle of the signal The search yielded the MUSIC spatial spectrum.
3. The uniform circular array one-dimensional spatial spectrum direction-finding signal processing device according to claim 2, characterized in that: Follow the function below The search yielded the MUSIC spatial spectrum: ； It is the azimuth angle of the signal incident. The corresponding array steering vector, for a uniform circular array, The expression is: ； Where R represents the radius of the circular array; M represents the number of array elements; .
4. The uniform circular array one-dimensional spatial spectrum direction-finding signal processing device according to claim 3, characterized in that: Capon algorithm processing module for R X Matrix inversion for the incident azimuth angle of the signal The search yielded the Capon space spectrum.
5. The uniform circular array one-dimensional spatial spectrum direction-finding signal processing device according to claim 4, characterized in that: Follow the function below The search yielded the Capon space spectrum: 。 6. The uniform circular array one-dimensional spatial spectrum direction-finding signal processing device according to claim 5, characterized in that: If the incoming wave contains a coherent signal, the DML algorithm measures the incident angle of the coherent signal; matrix It is the array steering matrix of D incident signals Then the guiding matrix projection matrix Represented as For a steering matrix composed of multiple signals, the measured incoherent signal DOA is used as prior information, and the coherent signals are processed according to... Formula search yields the incident angle of the coherent signal. Represents the trace of a matrix; The angle corresponding to the maximum value is the incident angle of the coherent signal.