A joint ordering QR decomposition method for multi-antenna detection and related device

Abstract

The application discloses a kind of joint ordering for multi-antenna detection QR The application discloses a decomposition method and related device, belonging to the technical field of multiple-input multiple-output signal detection, comprising: initializing orthogonal matrix and upper triangular matrix;Obtain channel matrix H ; Based on the sum of residual energy and its energy joint weighting, extract candidate column from the residual column in the channel matrix H ; According to the candidate column, update the orthogonal matrix and the upper triangular matrix, until all columns in the channel matrix H ; Get the final orthogonal matrix and the upper triangular matrix;The final orthogonal matrix and the upper triangular matrix are used as joint ordering decomposition result for multi-antenna detection QR , this method and related device can cut out the node of super radius as soon as possible, reduce access node, improve the speed of multi-antenna detection while ensuring detection performance.

Description

Technical Field

[0001] This invention belongs to the field of multiple-input multiple-output signal detection technology, and relates to a joint sorting method for multi-antenna detection. QR Decomposition method and related apparatus. Background Technology

[0002] Over the past decade, the surge in mobile internet video, cloud applications, and IoT terminals has placed continuous pressure on cellular networks in terms of capacity, spectrum efficiency, edge coverage, and energy consumption. The industry generally regards multi-antenna gain as one of the core technologies to meet the growing traffic. Massive MIMO (Multiple-Input Multiple-Output) technology, by deploying a large number of antennas at the transmitter / receiver end and performing spatial multiplexing and beamforming, can serve more users on the same time and frequency resources, increase cell capacity, and enhance anti-interference and coverage robustness. It has become a key enabling technology for 5G NR (5G New Radio) and is widely regarded as one of the cornerstone capabilities for moving towards 6G.

[0003] Along with the benefits come the challenge of detection complexity, especially in uplink multi-user / massive antenna scenarios: as the transmit dimension and modulation order increase, the search space for maximum likelihood detection grows exponentially. Therefore, practical systems and a large amount of engineering research tend to use linear detectors or their simpler variants because of their low complexity and mature implementation.

[0004] Sphere decoding is a quasi-ML (Machine Learning) search algorithm: it searches for the lattice point with the minimum distance within a hypersphere of a given radius. The tree-like search structure can significantly reduce the average complexity and achieve the same performance as ML. Although sphere decoding (SD) has superior performance, its complexity fluctuates greatly with channel conditions and signal-to-noise ratio. However, in small-to-medium dimensionality, high reliability, or specific processing scenarios, SD and its improved versions are still frequently used and studied to improve bit error rate performance.

[0005] Almost all mainstream SD processes perform preprocessing on the channel matrix. QR Decomposition (QRDecomposition, orthogonal triangular decomposition) triangulates the system, ensuring that the computation of cumulative metrics at each level and the tree search are performed from bottom to top according to the sign hierarchy. This preprocessing directly determines the shape of the search tree and the distribution of metrics at each level, thus significantly affecting the speed of radius tightening, the order of node expansion, and pruning efficiency. Therefore, QR The quality of the decomposition has a direct impact on the time complexity of SD.

[0006] standard QRThe decomposition is numerically stable, but it is not optimized for pruning in SD (Solution-Oriented Search). The diagonal and upper triangular structure of the upper triangular matrix may not be beneficial for pruning in the early stages of the search. Therefore, sorting is often used in engineering. QR Decompose the columns, rearrange them according to certain criteria, and then... QR Decomposition. A common criterion is column modulus length sorting, selecting the column with the smallest modulus length as the base each time, pushing layers with strong signal-to-noise ratio (SNR) to the bottom, so that the first detected layer has a higher SNR, thereby reducing candidate expansion in spherical decoding. However, this sorting based on column modulus length as the core criterion has two shortcomings:

[0007] In low signal-to-noise ratio / strongly correlated channels, it is impossible to characterize the attenuation effect of the basis on other columns. (Ordering) QR The decomposition only considers the diagonal terms of the diagonal matrix, i.e., the column magnitude, ignoring the off-diagonal terms, i.e., the correlation between columns. As a result, even if a column has a small magnitude, but is highly correlated with the other columns, selecting it may not significantly reduce the energy of the other columns in the orthogonal complement. The Euclidean distance of the first detection layer is relatively small, the radius shrinks slowly, and pruning is delayed.

[0008] Its adaptability to signal-to-noise ratio and modulation order is limited. (Ordering) QR The decomposition criteria lack adaptability to noise intensity and modulation order. When noise dominates, its preferred ordering may not push the strong layers that truly help with pruning to the first detection position.

[0009] In addition, if we only consider the residual energy of the remaining columns after selecting a certain basis, that is, selecting the candidate that maximizes the total residual energy of the remaining columns in each step, this can directly serve the goal of early pruning. Therefore, it is usually better than sorting in low signal-to-noise ratio situations. QR Decomposition. However, it also has a drawback: at high signal-to-noise ratios, the energy distribution of the selected columns themselves may be ignored. Simply maximizing the residual energy of the remaining columns may prematurely select some columns with high energy but weak correlation with other columns to the upper layer, resulting in an undesirable diagonal decreasing structure of the upper triangular matrix R, which in turn makes it impossible to prune early, ultimately leading to an increase in the number of access nodes and thus reducing the speed of multi-antenna detection. Summary of the Invention

[0010] The purpose of this invention is to overcome the shortcomings of the prior art and provide a joint ranking method for multi-antenna detection. QR The decomposition method and related apparatus can cut out nodes exceeding the radius as early as possible, reduce the number of accessed nodes, and improve the antenna detection speed while ensuring detection performance.

[0011] To achieve the above objectives, this invention discloses a joint ranking method for multi-antenna detection. QR Decomposition methods include:

[0012] Initialize the orthogonal matrix and the upper triangular matrix; obtain the channel matrix. H ;

[0013] Based on the sum of residual energy and its own energy, weighted together, from the channel matrix... H Extract candidate columns from the remaining columns, and update the orthogonal matrix and upper triangular matrix based on the candidate columns until the channel matrix has been traversed. H By analyzing all columns, we obtain the final orthogonal matrix and upper triangular matrix;

[0014] The final orthogonal matrix and upper triangular matrix are used as a joint sorting for multi-antenna detection. QR Decomposition results.

[0015] Furthermore, in the During the extraction of candidate columns, the channel matrix H The number of remaining columns is , The number of transmitting antennas is indicated by the sum of residual energy and its own weighted average, derived from the channel matrix. H The process of extracting candidate columns from the remaining columns is as follows:

[0016] Calculate the channel matrix H Inside The sum of the residual energies of the remaining columns;

[0017] Using the The sum of the residual energies of each remaining column is used to calculate the [specific energy] based on a joint weighted average of its own energies. From the scores of the remaining columns, obtain the remaining column with the highest score;

[0018] The remaining column with the highest score is selected as the candidate column.

[0019] Furthermore, the calculation of the channel matrix H Inside The process of summing the residual energies of the remaining columns is as follows:

[0020] The The remaining column of the th The sum of the residual energy of the column for:

[0021]

[0022] in, For Gram matrices with MMSE (Minimum Mean Square Error) regularization The Middle The MMSE regularization energy of the column Gram matrices with MMSE regularization The Middle The MMSE regularization energy of the column Gram matrices with MMSE regularization The Middle Line 1 The elements corresponding to the column.

[0023] Furthermore, the Gram matrix with MMSE regularization Represented as:

[0024]

[0025] in, Submatrix The conjugate transpose of the submatrix , Indicates the first Residual energy matrix during the second extraction All lines, the first Listed to number A matrix consisting of all elements corresponding to each column. , For the first The orthogonal matrix obtained after the extraction is updated. For the first Arrange the matrix during the next extraction. Let Variance be the noise variance. express unit array, express A complex matrix of dimension 1; express Complex matrices of dimension express The conjugate transpose of . This represents the number of receiving antennas.

[0026] Furthermore, the aforementioned The remaining column of the th Column rating results for:

[0027]

[0028] in, These are quantities related to the signal-to-noise ratio and modulation order. For the The remaining column of the th The sum of the residual energies of the column, Indicates the first The MMSE regularization energy of the column.

[0029] Furthermore, the process of updating the orthogonal matrix based on the candidate columns is as follows:

[0030] The first The candidate columns obtained during the extraction are used as the orthogonal matrix. Initial values ​​for the column;

[0031] For the orthogonal matrix, the first... The initial values ​​of the columns are normalized to obtain the first column in the orthogonal matrix. The result of normalizing the initial values ​​of the column ;

[0032] Will As the th in the orthogonal matrix The final value of the column.

[0033] Furthermore, the process of updating the upper triangular matrix based on the candidate columns is as follows:

[0034] upper triangular matrix The first to Okay, number A matrix consisting of all elements corresponding to each column Updated to:

[0035]

[0036] upper triangular matrix The Middle Okay, number Listed to number A matrix consisting of all elements corresponding to each column Updated to:

[0037]

[0038] in, Orthogonal matrix All rows, from the 1st to the 2nd A matrix consisting of all elements corresponding to each column; For the first Residual energy matrix during the second extraction All lines, the first Listed to number A matrix consisting of all elements corresponding to each column. For the first During the next extraction, the permutation matrix The List.

[0039] This invention discloses a joint ranking method for multi-antenna detection. The decomposition system includes:

[0040] The acquisition module is used to initialize the orthogonal matrix and the upper triangular matrix; and to acquire the channel matrix. H ;

[0041] The update module is used to weight the channel matrix based on the sum of residual energy and its own energy. H Extract candidate columns from the remaining columns, and update the orthogonal matrix and upper triangular matrix based on the candidate columns until the channel matrix has been traversed. H By analyzing all columns, we obtain the final orthogonal matrix and upper triangular matrix;

[0042] The configuration module is used to use the final orthogonal matrix and upper triangular matrix as a joint sorting algorithm for multi-antenna detection. Decomposition results.

[0043] This invention discloses a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the joint sorting for multi-antenna detection. The steps of the decomposition method.

[0044] This invention discloses a computer-readable storage medium storing a computer program that, when executed by a processor, implements the joint sorting for multi-antenna detection. The steps of the decomposition method.

[0045] The present invention has the following beneficial effects:

[0046] The joint sorting method for multi-antenna detection described in this invention QR In practical operation, the decomposition method and related devices, based on the combined weighting of the sum of residual energy and its own energy, decompose the energy from the channel matrix. H Candidate columns are extracted from the remaining columns within the matrix, and the orthogonal matrix and upper triangular matrix are updated based on the candidate columns. This pushes the high-energy layers to the bottom, significantly amplifies the increment of erroneous branches, and thus prunes nodes that exceed the radius earlier to reduce the number of accessed nodes. This improves the antenna detection speed while ensuring detection performance, making it highly practical. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a flowchart of the method of the present invention;

[0049] Figure 2 The gain curves of three algorithms for 6-transmit 6-receive 16QAM (Quadrature Amplitude Modulation) modulation are shown in the simulation experiment.

[0050] Figure 3 The gain curves of three 64QAM modulation algorithms with 4 transmit and 4 receive signals are shown in the simulation experiment.

[0051] Figure 4 This is a system structure diagram of the present invention. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Example 1

[0054] refer to Figure 1 The joint sorting method for multi-antenna detection described in this invention QR The decomposition method includes the following steps:

[0055] 1) Construct the variable matrix;

[0056] Let the channel matrix be set. ,in, express Complex matrices of dimension Indicates the number of receiving antennas. Indicates the number of transmitting antennas. , Channel matrix H The Middle Column vectors, express A complex matrix of dimension , let the variance of the noise be . The modulation order is The joint sorting method for multi-antenna detection described in this invention QR The purpose of the decomposition method is to find a permutation vector. Orthogonal matrix and upper triangular matrix , express A complex matrix of dimension , satisfying ,in, Represents the channel matrix The columns in the permutation vector The order of the elements in the text is rearranged.

[0057] set up The residual energy matrix, The residual energy matrix at the time of the first extraction is initialized to... That is, the channel matrix The value is assigned to the residual energy matrix. Permutation matrix Used to store the channel matrix Sort and arrange the columns in the table;

[0058] 2) From Begin by traversing the channel matrix. H All columns are jointly sorted and decomposed based on the sum of residual energies;

[0059] In step 2), at the During extraction, orthogonal matrix Middle front Column It has been determined, then the first Residual energy matrix during the second extraction All lines, the first Listed to number All elements corresponding to the column for:

[0060]

[0061] in, For the first The orthogonal matrix during extraction; when hour, It is a matrix of all zeros. An identity matrix, an upper triangular matrix The matrix is ​​all zeros, and the permutation vector is... , Indicates the first During the next extraction, the permutation matrix All lines, the first Listed to number A matrix consisting of all elements corresponding to each column. express The conjugate transpose of .

[0062] The specific process of step 2) is as follows:

[0063] 21) Calculate the sum of the residual energies of all remaining columns. ;

[0064] In the channel matrix H Remaining To select the most suitable column from the remaining columns, it is necessary to iterate through the rest. Each column is used to calculate the residual energy and column modulus length, and a submatrix is ​​defined. , express Complex matrices of dimension Indicates the first During the second extraction, the residual energy matrix All lines, the first Listed to number A matrix consisting of all elements corresponding to a column, a submatrix The List as Considering the influence of noise, calculate the Gram matrix with MMSE regularization. for:

[0065]

[0066] Among them, Gram matrices with MMSE regularization The element corresponding to the j-th row and k-th column , This means taking all elements on the diagonal of the matrix and forming a vector. Indicates the first The MMSE regularization energy of the column express unit array, express A complex matrix of dimension 1.

[0067] When the remaining The first in the column Using column j as a candidate, normalize the j-th column to obtain the normalized result of column j. The first Listed in Energy in direction for:

[0068]

[0069] Then the first When column 1 is used as a candidate column, the first column 2 Residual energy of the column , For the first The MMSE regularization energy of the column, therefore, we obtain, when the first When a column is selected as a candidate column, the sum of the residual energies of all remaining columns. for:

[0070]

[0071] 22) Adaptive weighted joint scoring:

[0072] After traversing all remaining... After each column, obtain The sum of the residual energies of all remaining columns, considering both residual energy and column modulus, when choosing the th column... When column 1 is used as a candidate column, the first column 2 Column rating results for:

[0073]

[0074] in, This is a quantity related to the signal-to-noise ratio and modulation order, and the specific calculation process is as follows:

[0075] Calculate the equivalent signal-to-noise ratio:

[0076]

[0077] Based on equivalent signal-to-noise ratio Determine weights The specific values ​​are:

[0078]

[0079] Finally, the candidate column that results in the highest score is selected. As an orthogonal matrix The Middle The initial value of the column, i.e.:

[0080]

[0081] 23) Column permutation and renew;

[0082] orthogonal matrix The Middle The initial value of the column is in the first position. Permutation matrix during the next extraction The column index in the text is pick, when ,get:

[0083]

[0084] in, Indicates the first The permutation matrix is ​​swapped during the next extraction. The Column and number Column, will All lines, the first Assigning the element of column to the first Column, will All lines, the first Assigning the element of column to the first List; Indicates the first Exchange residual energy matrix during the second extraction The Column and number List; Represents the commutation vector of the th permutation vector The element and the first Each element.

[0085] For orthogonal matrices The Middle The initial values ​​of the columns are normalized; specifically, the Gram-Schmidt orthogonalization method is used to obtain orthogonal matrices. The Middle The result of normalizing the initial values ​​of the column for:

[0086]

[0087] in, For the first Residual energy matrix during the second extraction The List.

[0088] Then As an orthogonal matrix The Middle The final value of the column;

[0089] For upper triangular matrices The first to Okay, number Update the columns to make the permutation matrix W The i-th column ,when Then the upper triangular matrix The first to Okay, number The column is updated to:

[0090]

[0091] in, Represents an upper triangular matrix The 1st to the 1st A matrix consisting of rows and all corresponding elements of each column; Represents an orthogonal matrix All rows, from the 1st to the 2nd The matrix formed by the corresponding elements of each column; the superscript H represents the conjugate transpose of the matrix.

[0092] upper triangular matrix The Middle Okay, number Listed to number A matrix consisting of all elements corresponding to each column Updated to:

[0093]

[0094] in, Orthogonal matrix All rows, from the 1st to the 2nd A matrix consisting of all elements corresponding to each column; For the first Residual energy matrix during the second extraction All lines, the first Listed to number A matrix consisting of all elements corresponding to each column. For the first During the next extraction, the permutation matrix The List, for The conjugate transpose of .

[0095] Simultaneously, the residual energy matrix Updated to:

[0096]

[0097] in, Indicates the first Residual energy matrix during extraction All lines, the first Listed to number A matrix consisting of all elements corresponding to each column. For the first Residual energy matrix during extraction All lines, the first Listed to number A matrix consisting of all elements corresponding to each column.

[0098] Finally, the first The upper triangular matrix extracted in this step Residual energy matrix Orthogonal matrix Then proceed to the next step.

[0099] 24) Judgment Is it equal to ,when When, then the current orthogonal matrix is... Upper triangular matrix The permutation vector p is used as the joint sorting of the channel matrix. Decomposition results; when At that time, then Then proceed to step 21).

[0100] Simulation Experiment

[0101] In MATLAB, this experiment sets up different modulation methods and the number of antennas to simulate different algorithms compared to the original. QR Decomposition, in sorting QR During the decomposition process, the number of search nodes decreased by a certain percentage, defined as follows:

[0102]

[0103] Set the number of Monte Carlo operations for each signal-to-noise ratio. And the number of time slots, that is, the number of transmit and receive processes simulated for each slot at each signal-to-noise ratio, for a total simulation. Next. Calculate the reduction percentage for each sending and receiving process, then... By averaging the number of time slots, we can obtain the reduction ratio under the current signal-to-noise ratio.

[0104] This experiment simulated two different modulation schemes and antenna numbers, using the same random seed each time to ensure identical channel matrix and noise levels, guaranteeing fairness. The first scheme was 6-transmit, 6-receive 16QAM modulation, with the number of time slots set to 500 for each signal-to-noise ratio, and the number of Monte Carlo iterations... The simulation results are as follows Figure 2 As shown. The second method is 4 transmit, 4 receive 64QAM modulation, with the number of time slots set to 500 for each signal-to-noise ratio, and the number of Monte Carlo iterations... The simulation results are as follows Figure 3 As shown in the figure. The solid lines marked with diamonds represent JO-QRD decomposition (joint-order qr decomposition). QR The result of the decomposition is the result corresponding to the algorithm of this invention; the dotted line of the square mark is SQRD (Sorted QR Decomposition). QR The result of the decomposition is that only the column with the lowest energy is selected as the candidate column; the dotted line marked with a circle represents RE- QR The result of D-decomposition (residual-energy QR decomposition) is that, without the scoring step, the column with the largest sum of residual energies is directly selected as the candidate. Figure 2 and Figure 3 It can be seen that, compared to those that are not sorted... QRBy decomposition and comparison, this invention, while ensuring the equivalent BER, combines... QR The average number of exploration nodes in the decomposition is significantly reduced, specifically by 25% to 40% in the 5dB to 20dB range, and even at lower and higher signal-to-noise ratios or when column correlation is significant, the effect on ranking is also improved. QR Decomposition also has the advantage of stability.

[0105] Example 2

[0106] refer to Figure 4 The joint sorting method for multi-antenna detection described in this invention QR The decomposition system includes:

[0107] The acquisition module is used to initialize the orthogonal matrix and the upper triangular matrix; and to acquire the channel matrix. H ;

[0108] The update module is used to weight the channel matrix based on the sum of residual energy and its own energy. H Extract candidate columns from the remaining columns, and update the orthogonal matrix and upper triangular matrix based on the candidate columns until the channel matrix has been traversed. H By analyzing all columns, we obtain the final orthogonal matrix and upper triangular matrix;

[0109] The configuration module is used to use the final orthogonal matrix and upper triangular matrix as a joint sorting algorithm for multi-antenna detection. QR Decomposition results.

[0110] The module division in this embodiment is illustrative and represents only one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional modules in each embodiment of this application can be integrated into a single processor, exist as separate physical entities, or be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0111] Example 3

[0112] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the joint sorting for multi-antenna detection. QR The steps of the decomposition method, for example, include: initializing the orthogonal matrix and the upper triangular matrix; obtaining the channel matrix. H Based on the sum of residual energy and its own energy, weighted together, from the channel matrix... H Extract candidate columns from the remaining columns, and update the orthogonal matrix and upper triangular matrix based on the candidate columns until the channel matrix has been traversed. HExtracting all columns from the given matrix yields the final orthogonal matrix and upper triangular matrix; this final orthogonal matrix and upper triangular matrix are then used as the joint sorting algorithm for multi-antenna detection. QR The decomposition results are as follows: The memory may include main memory, such as high-speed random access memory (RAM), and may also include non-volatile memory, such as at least one disk storage device. The processor, network interface, and memory are interconnected via an internal bus, which can be an industry-standard architecture bus, a peripheral component interconnection standard bus, an extended industry-standard architecture bus, etc. The bus can be divided into address bus, data bus, control bus, etc. The memory is used to store programs; specifically, the program may include program code, which includes computer operation instructions. The memory may include main memory and non-volatile memory, and provides instructions and data to the processor.

[0113] Example 4

[0114] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the joint sorting for multi-antenna detection. QR The steps of the decomposition method, for example, include: initializing the orthogonal matrix and the upper triangular matrix; obtaining the channel matrix. H Based on the sum of residual energy and its own energy, weighted together, from the channel matrix... H Extract candidate columns from the remaining columns, and update the orthogonal matrix and upper triangular matrix based on the candidate columns until the channel matrix has been traversed. H Extracting all columns from the given matrix yields the final orthogonal matrix and upper triangular matrix; this final orthogonal matrix and upper triangular matrix are then used as the joint sorting algorithm for multi-antenna detection. QR The decomposition results are as follows. Specifically, the computer-readable storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. The volatile memory may include random access memory (RAM) and / or cache memory, etc. The non-volatile memory may include read-only memory (ROM), hard disk, flash memory, optical disk, magnetic disk, etc.

[0115] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and disclosure of the invention. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein.

[0116] It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

[0117] The above description is merely a preferred embodiment of the present invention and does not constitute any limitation on the present invention. Any simple modifications, alterations, or equivalent structural changes made to the above embodiments based on the technical essence of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A joint ranking method for multi-antenna detection QR The decomposition method is characterized by, include: Initialize the orthogonal matrix and the upper triangular matrix; obtain the channel matrix. H ; Based on the sum of residual energy and its own energy, weighted together, from the channel matrix... H Extract candidate columns from the remaining columns, and update the orthogonal matrix and upper triangular matrix based on the candidate columns until the channel matrix has been traversed. H By analyzing all columns, we obtain the final orthogonal matrix and upper triangular matrix; The final orthogonal matrix and upper triangular matrix are used as a joint sorting for multi-antenna detection. QR Decomposition results; In the During the extraction of candidate columns, the channel matrix H The number of remaining columns is , The number of transmitting antennas is indicated by the sum of residual energy and its own weighted average, derived from the channel matrix. H The process of extracting candidate columns from the remaining columns is as follows: Calculate the channel matrix H Inside The sum of the residual energies of the remaining columns; Using the The sum of the residual energies of each remaining column is used to calculate the [specific energy] based on a joint weighted average of its own energies. From the scores of the remaining columns, obtain the remaining column with the highest score; The remaining column with the highest score is selected as the candidate column.

2. The joint ranking method for multi-antenna detection according to claim 1 QR The decomposition method is characterized by, The calculated channel matrix H Inside The process of summing the residual energies of the remaining columns is as follows: The The remaining column of the th The sum of the residual energy of the column for: in, Gram matrices with MMSE regularization The Middle The MMSE regularization energy of the column Gram matrices with MMSE regularization The Middle The MMSE regularization energy of the column Gram matrices with MMSE regularization The Middle Line 1 The elements corresponding to the column.

3. The joint sorting method for multi-antenna detection according to claim 2 QR The decomposition method is characterized by, The Gram matrix with MMSE regularization Represented as: in, Submatrix The conjugate transpose of the submatrix , Indicates the first Residual energy matrix during the second extraction All lines, the first Listed to number All elements corresponding to the column, , For the first The orthogonal matrix obtained after the extraction is updated. For the first Arrange the matrix during the next extraction. Let Variance be the noise level. express unit array, express A complex matrix of dimension 1; express Complex matrices of dimension express The conjugate transpose of . This represents the number of receiving antennas.

4. The joint ranking method for multi-antenna detection according to claim 1 QR The decomposition method is characterized by, The The remaining column of the th Column rating results for: in, These are quantities related to the signal-to-noise ratio and modulation order. For the The remaining column of the th The sum of the residual energies of the column, Indicates the first The MMSE regularization energy of the column.

5. Joint ranking for multi-antenna detection as described in claim 1 QR The decomposition method is characterized by, The process of updating the orthogonal matrix based on the candidate columns is as follows: The first The candidate columns obtained during the extraction are used as the th in the orthogonal matrix. Initial values ​​for the column; For the orthogonal matrix of the th The initial values ​​of the columns are normalized to obtain the first column in the orthogonal matrix. The result of normalizing the initial values ​​of the column ; Will As the th in the orthogonal matrix The final value of the column.

6. Joint ranking for multi-antenna detection as described in claim 5 QR The decomposition method is characterized by, The process of updating the upper triangular matrix based on the candidate columns is as follows: upper triangular matrix The first to Okay, number A matrix consisting of all elements corresponding to each column Updated to: upper triangular matrix The Middle Okay, number Listed to number A matrix consisting of all elements corresponding to each column Updated to: in, Orthogonal matrix All rows, from the 1st to the 2nd A matrix consisting of all elements corresponding to each column; For the first Residual energy matrix during the second extraction All lines, the first Listed to number A matrix consisting of all elements corresponding to each column. For the first During the next extraction, the permutation matrix The List, for The conjugate transpose of .

7. A joint ranking method for multi-antenna detection QR The decomposition system is characterized by, include: The acquisition module is used to initialize the orthogonal matrix and the upper triangular matrix; and to acquire the channel matrix. H ; The update module is used to weight the channel matrix based on the sum of residual energy and its own energy. H Extract candidate columns from the remaining columns, and update the orthogonal matrix and upper triangular matrix based on the candidate columns until the channel matrix has been traversed. H By analyzing all columns, we obtain the final orthogonal matrix and upper triangular matrix; The configuration module is used to use the final orthogonal matrix and upper triangular matrix as a joint sorting algorithm for multi-antenna detection. QR Decomposition results; In the During the extraction of candidate columns, the channel matrix H The number of remaining columns is , The number of transmitting antennas is indicated by the sum of residual energy and its own weighted average, derived from the channel matrix. H The process of extracting candidate columns from the remaining columns is as follows: Calculate the channel matrix H Inside The sum of the residual energies of the remaining columns; Using the The sum of the residual energies of each remaining column is used to calculate the [specific energy] based on a joint weighted average of its own energies. From the scores of the remaining columns, obtain the remaining column with the highest score; The remaining column with the highest score is selected as the candidate column.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the joint sorting for multi-antenna detection as described in any one of claims 1-6. QR The steps of the decomposition method.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the joint sorting for multi-antenna detection as described in any one of claims 1-6. QR The steps of the decomposition method.

Images