Precoding matrix acquisition method, device, electronic device and storage medium

Abstract

The present application relates to the field of wireless communication technology and discloses a method, device, electronic device, and storage medium for obtaining a precoding matrix. The method includes: determining a method for obtaining a precoding matrix; wherein the obtaining method includes online obtaining and obtaining based on offline learning results; when the obtaining method is determined to be online obtaining, the centroid of the spatial grid to which the channel matrix belongs is obtained based on the clustering result of the channel matrix of the target user equipment UE, and the precoding matrix of the target UE is obtained based on the centroid; when the obtaining method is determined to be based on offline learning results, the spatial grid where the target UE is located is obtained, and based on the optimal precoding matrix of each spatial grid learned offline, the optimal precoding matrix of the spatial grid where the target UE is located is obtained as the precoding matrix of the target UE. This method reduces the occupation of channel resources and computing resources by obtaining the precoding matrix, improves the efficiency of obtaining the precoding matrix, reduces costs, and improves the shaping performance of users.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of wireless communication technology, and in particular to a method, device, electronic device, and storage medium for obtaining a precoding matrix. Background Art

[0002] With the advancement of communications technology, the global 5G standard (5G New Radio, 5GNR), featuring a new air interface design based on Orthogonal Frequency Division Multiplexing (OFDM), is becoming the foundation of next-generation cellular technology. In wireless communication systems, precoding (digital beamforming) technology can reduce interference between data streams, improve channel condition factors, and increase data throughput, making it a key technology for improving system performance.

[0003] The precoding schemes currently used in wireless communication systems generally select the precoding matrix based on the precoding matrix indicator (PMI) feedback from the user equipment (UE) or the precoding matrix based on the measurement of the sounding reference signal (SRS).

[0004] However, in the process of precoding using existing precoding schemes, the precoding scheme based on PMI feedback requires the UE to feedback the selection of the PMI codebook, and the precoding scheme based on SRS requires the UE to send a measurement signal. Both will cause additional time-frequency resource occupation and bring about a large channel resource overhead; in addition, the precoding method based on singular value decomposition (SVD) requires the use of huge computing resources for calculation when the number of antennas is large, which brings about a large computing resource overhead; in addition, the precoding scheme based on PMI requires the UE to measure the downlink channel before feedback, and the precoding scheme based on SRS needs to utilize the uplink and downlink reciprocity of the channel. When the UE is in motion, especially when it is in high-speed motion, the channel when using the calculated precoding for transmission is easily inconsistent with the channel when calculating the precoding, resulting in performance loss. Summary of the Invention

[0005] The main purpose of the embodiments of the present application is to propose a precoding matrix acquisition method, device, electronic device and storage medium, aiming to reduce the overhead of precoding matrix acquisition while improving the accuracy and adaptability of precoding matrix acquisition, so as to ensure high-performance data transmission of the communication system as much as possible.

[0006] To achieve the above-mentioned purpose, an embodiment of the present application provides a precoding matrix acquisition method, including: determining a method for acquiring the precoding matrix; wherein the acquisition method includes online acquisition and acquisition based on offline learning results; when the acquisition method is determined to be online acquisition, according to the clustering result of the channel matrix of the target user equipment UE, the centroid of the spatial grid to which the channel matrix belongs is obtained, and the precoding matrix of the target UE is obtained based on the centroid; when the acquisition method is determined to be based on offline learning results, the spatial grid where the target UE is located is obtained, and based on the optimal precoding matrix of each spatial grid learned offline, the optimal precoding matrix of the spatial grid where the target UE is located is obtained as the precoding matrix of the target UE.

[0007] In order to achieve the above-mentioned purpose, an embodiment of the present application also provides a precoding matrix acquisition device, including: a determination module, used to determine a method for obtaining the precoding matrix; wherein the acquisition method includes online acquisition and acquisition based on offline learning results; a first acquisition module, used to obtain the centroid of the spatial grid to which the channel matrix belongs according to the clustering result of the channel matrix of the target user equipment UE when the acquisition method is determined to be online acquisition, and obtain the precoding matrix of the target UE based on the centroid; a second acquisition module, used to obtain the spatial grid where the target UE is located when the acquisition method is determined to be based on offline learning results, and obtain the optimal precoding matrix of the spatial grid where the target UE is located based on the optimal precoding matrix of each spatial grid learned offline, as the precoding matrix of the target UE.

[0008] To achieve the above-mentioned objectives, an embodiment of the present application further provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the precoding matrix acquisition method as described above.

[0009] To achieve the above-mentioned purpose, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, which implements the above-mentioned precoding matrix acquisition method when executed by a processor.

[0010] The precoding matrix acquisition method provided in the embodiment of the present application determines the acquisition method of the precoding matrix based on factors such as the current state of the communication system. When the precoding matrix is ​​acquired in an online manner, the spatial grid to which the channel matrix belongs is obtained by clustering based on the channel matrix of the target user equipment UE, and the precoding matrix of the target UE is obtained based on the centroid of the spatial grid. When the precoding matrix is ​​acquired in a manner based on offline learning results, the spatial grid to which the target UE belongs is determined, and the precoding matrix of the target UE is determined from the optimal precoding matrices of each spatial grid learned offline. During online acquisition, the target UE is divided into spatial grids according to the clustering results of the channel matrix obtained in real time, and the precoding matrix of the target UE is obtained based on the centroid of the spatial grid to which it belongs. Without the need for UE feedback, the precoding matrix of the target UE is accurately and efficiently obtained, thereby improving the accuracy of precoding matrix acquisition and the adaptability to the terminal motion state. When acquiring based on offline learning results, the precoding matrix of the target UE is determined according to the spatial grid to which the target UE belongs in the optimal coding matrix of each spatial grid determined based on offline data, thereby greatly reducing the occupation of channel resources and real-time computing resources, improving the timeliness of precoding matrix feedback, reducing the cost of precoding matrix acquisition, and improving the user's shaping performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] One or more embodiments are exemplarily described by the figures in the corresponding drawings, and these exemplified descriptions do not constitute limitations on the embodiments.

[0012] Figure 1 is a flow chart of a method for obtaining a precoding matrix in an embodiment of the present application;

[0013] Figure 2 is a structural diagram of a precoding matrix acquisition device in another embodiment of the present application;

[0014] Figure 3 It is a structural diagram of an electronic device in another embodiment of the present application. DETAILED DESCRIPTION

[0015] As can be seen from the background art, when using current precoding schemes for precoding, obtaining the precoding matrix consumes a lot of resources and has high costs. Furthermore, the precoding matrix has poor adaptability to user devices in motion, which can easily lead to a decrease in overall communication system performance. Therefore, how to simply, accurately, and efficiently obtain the precoding matrix and improve the effectiveness and adaptability of precoding schemes is an urgent technical issue.

[0016] In order to solve the above problems, an embodiment of the present application provides a precoding matrix acquisition method, including: determining a method for acquiring a precoding matrix; wherein the acquisition method includes online acquisition and acquisition based on offline learning results; when the acquisition method is determined to be online acquisition, according to the clustering result of the channel matrix of the target user equipment UE, the centroid of the spatial grid to which the channel matrix belongs is obtained, and the precoding matrix of the target UE is obtained based on the centroid; when the acquisition method is determined to be based on offline learning results, the spatial grid where the target UE is located is obtained, and based on the optimal precoding matrix of each spatial grid learned offline, the optimal precoding matrix of the spatial grid where the target UE is located is obtained as the precoding matrix of the target UE.

[0017] The precoding matrix acquisition method provided in the embodiment of the present application determines the acquisition method of the precoding matrix based on factors such as the current state of the communication system. When the precoding matrix is ​​acquired in an online manner, the spatial grid to which the channel matrix belongs is obtained by clustering based on the channel matrix of the target user equipment UE, and the precoding matrix of the target UE is obtained based on the centroid of the spatial grid. When the precoding matrix is ​​acquired in a manner based on offline learning results, the spatial grid to which the target UE belongs is determined, and the precoding matrix of the target UE is determined from the optimal precoding matrices of each spatial grid learned offline. During online acquisition, the target UE is divided into spatial grids according to the clustering results of the channel matrix obtained in real time, and the precoding matrix of the target UE is obtained based on the centroid of the spatial grid to which it belongs. Without the need for UE feedback, the precoding matrix of the target UE is accurately and efficiently obtained, thereby improving the accuracy of precoding matrix acquisition and the adaptability to the terminal motion state. When acquiring based on offline learning results, the precoding matrix of the target UE is determined according to the spatial grid to which the target UE belongs in the optimal coding matrix of each spatial grid determined based on offline data, thereby greatly reducing the occupation of channel resources and real-time computing resources, improving the timeliness of precoding matrix feedback, reducing the cost of precoding matrix acquisition, and improving the user's shaping performance.

[0018] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, each embodiment of the present application will be described in detail below with reference to the accompanying drawings. However, it will be understood by those skilled in the art that in each embodiment of the present application, many technical details are proposed to enable the reader to better understand the present application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical solutions claimed in the present application can be implemented. The division of the following embodiments is for convenience of description and should not constitute any limitation on the specific implementation of the present application. The various embodiments can be combined and referenced with each other under the premise of no contradiction.

[0019] The following will describe in detail the implementation details of the precoding matrix acquisition method described in this application in combination with specific embodiments. The following content is only the implementation details provided for ease of understanding and is not necessary for implementing this solution.

[0020] The first aspect of the embodiment of the present application relates to a method for obtaining a precoding matrix. The specific process of the method for obtaining a precoding matrix is ​​shown in FIG. Figure 1 The precoding matrix acquisition method can be applied to a base station of a communication system, or a terminal device connected to a base station of the communication system. This embodiment is described using the base station as an example. The precoding matrix acquisition method includes at least but is not limited to the following steps:

[0021] Step 101: Determine a method for obtaining a precoding matrix; wherein the obtaining method includes online obtaining and obtaining based on offline learning results.

[0022] Specifically, after being put into operation, a base station begins setting precoding matrices for data transmissions from different user devices. Upon acquiring the channel matrix for any user device within its jurisdiction, the base station checks its database status and, based on the database status detection results, determines how to acquire the user device's precoding matrix. These acquisition methods include online acquisition and acquisition based on offline learning results. The precoding matrix acquisition method is determined based on the database status to ensure the accuracy of the acquired precoding matrix.

[0023] In one example, a base station determines a method for acquiring a precoding matrix, including: determining the method for acquiring the precoding matrix based on the amount of channel matrix data of a UE stored in a database; wherein, if the amount of data is greater than a preset threshold, determining the acquisition method to be based on offline learning results; and if the amount of data is less than or equal to the preset threshold, determining the acquisition method to be online acquisition. Specifically, when determining the method for acquiring the precoding matrix, the base station may detect the amount of channel matrix data of the user equipment stored in its database. If the amount of channel matrix data stored in the database is greater than a preset threshold, the base station determines the acquisition method to be based on offline learning results; and if the amount of channel matrix data stored in the database is less than or equal to the preset threshold, the base station determines the acquisition method to be online acquisition. By determining the acquisition method based on the amount of channel matrix data stored in the database, the precoding matrix can be acquired through online learning when the amount of stored data is insufficient, thereby ensuring the accuracy and timeliness of precoding matrix acquisition. When the amount of stored data is sufficient, the precoding matrix can be acquired through offline learning results, thereby ensuring the accuracy of precoding matrix acquisition while avoiding the occupation of channel resources and real-time computing resources.

[0024] It is worth mentioning that, upon detecting a change in the channel environment, the base station can also test the validity of the channel matrix stored in its own database based on the current channel environment, delete historical data that is no longer suitable for the current channel environment, or update historical data based on the current signal environment. By testing and dynamically maintaining the validity of the stored channel matrix data based on the current channel environment, the accuracy of the precoding matrix obtained based on offline learning results can be further improved.

[0025] Step 102: When the acquisition mode is determined to be online acquisition, the centroid of the spatial grid to which the channel matrix belongs is obtained according to the clustering result of the channel matrix of the target user equipment UE, and the precoding matrix of the target UE is obtained based on the centroid.

[0026] Specifically, when the base station determines that the acquisition method of the precoding matrix is ​​online acquisition based on the database status detection result, it obtains the channel matrix of the target user equipment for which the precoding matrix needs to be determined, and clusters the channel matrix of the target user equipment. Based on the clustering result of the channel matrix of the target UE, it determines the spatial grid to which the channel matrix of the target UE belongs, then obtains the centroid of the spatial grid to which the channel matrix belongs, and obtains the precoding matrix of the target UE based on the obtained centroid. By obtaining the centroid of the spatial grid to which the channel matrix of the target UE belongs based on the clustering result of the channel matrix of the target UE, the precoding matrix of the target UE is obtained based on the obtained centroid, and the precoding matrix of the target UE is accurately obtained using the channel matrix obtained in real time, and the accurate acquisition of the precoding matrix is ​​completed without the help of feedback from the target UE.

[0027] In an example, the base station obtains the centroid of the spatial grid to which the channel matrix belongs based on the clustering result of the channel matrix of the target user equipment UE, including: determining whether there is a spatial grid to which the channel matrix belongs based on the clustering result of the channel matrix of the target UE; if there is no spatial grid to which the channel matrix belongs, creating a spatial grid as the spatial grid to which the target UE belongs, and determining the centroid of the spatial grid to which the target UE belongs based on the channel matrix of the target UE; if there is a spatial grid to which the channel matrix belongs, updating the centroid of the spatial grid to which the channel matrix belongs based on the channel matrix of the target UE.

[0028] Specifically, the base station first clusters the target UE's channel matrix and detects whether a spatial grid to which the target UE's channel matrix belongs exists. If a spatial grid to which the target UE's channel matrix belongs is detected, the centroid of the spatial grid to which the target UE's channel matrix belongs is updated based on the target UE's channel matrix. The updated centroid is then used as the centroid of the spatial grid to which the target UE's channel matrix belongs, and subsequent precoding matrix acquisition is continued. If the spatial grid to which the target UE's channel matrix belongs is not detected, a new spatial grid is created based on the target UE's channel matrix as the spatial grid to which the target UE belongs. The centroid of the created new spatial grid is calculated based on the target UE's channel matrix, and subsequent precoding matrix acquisition is continued based on the calculated centroid of the new spatial grid. By promptly updating the centroid of the spatial grid when the target UE's channel matrix exists, and creating the spatial grid to which the target UE's channel matrix belongs and calculating the centroid of the spatial grid when the target UE's channel matrix does not exist, the centroid of the spatial grid to which the target UE's channel matrix belongs is ensured to be accurately obtained, thereby ensuring the accuracy of the precoding matrix determined by online learning.

[0029] It is worth mentioning that when detecting the spatial grid to which the target UE channel matrix belongs, not only can clustering be performed based on the autocorrelation matrix, but also detection and judgment of the spatial grid can be performed based on the 1-norm, 2-norm, ∞-norm, Cosine distance, etc. This embodiment does not limit the specific detection method adopted.

[0030] Furthermore, the base station determines whether there is a spatial grid to which the channel matrix belongs based on the clustering result of the channel matrix of the target UE, including: obtaining the autocorrelation matrix of the channel matrix of the target UE; detecting the correlation between the autocorrelation matrix and the centroids of each existing spatial grid, and obtaining the correlation result S of each centroid. k ; Where k is the number of the spatial grid, and the maximum value of k is the number of existing spatial grids; k When the values ​​of S and S are all less than the preset correlation threshold, it is determined that the spatial grid to which the channel matrix belongs does not exist; k If there is a correlation threshold greater than or equal to the preset correlation threshold, S k The spatial grid corresponding to the maximum value in is determined as the spatial grid to which the channel matrix belongs.

[0031] Specifically, when the base station performs spatial grid detection on the channel matrix of the target UE, it first obtains the channel matrix H of the target UE in a preset manner. For example, the channel matrix H of the target UE is calculated based on the SRS measurement result, and then the autocorrelation matrix H of the channel matrix H is calculated based on the channel matrix H of the target UE. HH, where the superscript H represents the conjugate transpose of the channel matrix H. Then, based on the calculated autocorrelation matrix, the correlation between the autocorrelation matrix and the centroid of each existing spatial grid is calculated. For example, the correlation S between the autocorrelation matrix and the centroid of each existing spatial grid is calculated using the following formula: k :

[0032] S k =cov(H H H,C k ), k∈R

[0033] Among them, k is the number of the current spatial grid in the grid set, R is the number set of the grid set, C k is the centroid of the k-th spatial grid, cov(H H H,C k ) is a preset correlation calculation function. For example, the following correlation calculation function can be used for calculation:

[0034] cov(A,B)=real(trace(A H B) / sqrt(trace(A H A)) / sqrt(trace(B H B)))

[0035] Wherein, real() indicates taking the real part, trace() indicates taking the trace of the matrix, and sqrt() indicates taking the square root. This embodiment does not limit the specific correlation calculation function used.

[0036] After obtaining the correlation between the autocorrelation matrix and the centroids of the existing spatial grids, for each correlation S k The size relationship between the S k When all S k When all of them are less than 0.2, it is determined that the channel matrix does not belong to any existing spatial grid, that is, there is no spatial grid to which the channel matrix belongs; when one or more S k In the case of S k The spatial grid corresponding to the maximum value is used as the spatial grid to which the target UE channel matrix belongs. By detecting the correlation between the autocorrelation matrix corresponding to the target UE channel matrix and the centroid of each existing spatial grid, it is accurately detected whether the spatial grid to which the target UE channel matrix belongs exists in the existing spatial grid, facilitating the subsequent use of the corresponding method to accurately obtain the precoding matrix.

[0037] It is worth mentioning that the preset correlation threshold can be set based on experience or determined based on the networking characteristics of the current communication system. In addition, when there is a tendency to add more different spatial grids, the preset correlation threshold can be set larger. When there is a tendency to update the existing spatial grid, the preset correlation threshold can be set smaller. This embodiment does not limit the specific determination method and setting of the preset correlation threshold.

[0038] Furthermore, before detecting the correlation between the autocorrelation matrix and the centroids of each existing spatial grid, the base station further includes: detecting whether there is currently a spatial grid; if there is currently a spatial grid and the number of existing spatial grids is less than a preset number, then detecting the correlation between the autocorrelation matrix and the centroids of each existing spatial grid; if there is currently no spatial grid or the number of existing spatial grids is greater than or equal to the preset number, determining that there is no spatial grid to which the channel matrix belongs. Specifically, before calculating the correlation between the autocorrelation matrix and the centroids of each existing spatial grid, the base station detects whether there is currently an existing spatial grid and the number of existing spatial grids; if there is currently a spatial grid and the number of existing spatial grids is less than the preset number, determining that there may be a spatial grid to which the channel matrix belongs in the current spatial grid; then detecting the correlation between the autocorrelation matrix and the centroids of each existing spatial grid; if there is currently no spatial grid or the number of existing spatial grids is greater than or equal to the preset number, determining that there is no spatial grid to which the channel matrix belongs, and directly creating a new spatial grid. By setting the upper limit of the category of the spatial grid, the acquisition process of the spatial grid to which the signal matrix belongs is optimized, thereby improving the acquisition efficiency and accuracy.

[0039] In another example, the base station determines the centroid of the spatial grid to which it belongs based on the channel matrix of the target UE, including: determining the autocorrelation matrix of the channel matrix of the target UE as the centroid of the spatial grid to which it belongs; updating the centroid of the spatial grid to which it belongs based on the channel matrix of the target UE, including: obtaining the centroid of the spatial grid to which it belongs; and performing weighted averaging on the obtained centroid and the autocorrelation matrix of the channel matrix of the target UE to obtain the updated centroid of the spatial grid to which it belongs. Specifically, when the base station determines the centroid of the spatial grid to which the channel matrix belongs based on the channel matrix of the target UE, if the spatial grid to which the channel matrix of the target UE belongs is not detected, the base station enters the spatial grid creation process, creates a new spatial grid as the spatial grid to which the channel matrix of the target UE belongs, and then uses the autocorrelation matrix of the channel matrix of the target UE as the centroid of the newly created spatial grid; if the spatial grid to which the channel matrix of the target UE belongs is detected, the base station enters the spatial grid update process, obtains the current centroid of the spatial grid to which the channel matrix of the target UE belongs, and then performs weighted averaging on the obtained current centroid of the spatial grid and the autocorrelation matrix of the target UE channel matrix, and uses the result of the weighted averaging as the latest centroid of the spatial grid to which the channel matrix of the target UE belongs, and uses the latest centroid as the centroid of the spatial grid to which the channel matrix of the target UE belongs for subsequent acquisition of the precoding matrix. The centroid of the spatial grid to which the target UE belongs is acquired in a corresponding manner according to the detection result to ensure the accuracy of the precoding matrix subsequently acquired.

[0040] For example, if the spatial grid to which the channel matrix of the target UE belongs is not detected, a spatial grid numbered K is created, and the value of K can be the current spatial grid number plus 1. Then the channel matrix H of the target UE is transformed to obtain the autocorrelation matrix H of the channel matrix H. H H, and H H H is the centroid of the spatial grid K to which the target UE belongs. When the spatial grid to which the channel matrix of the target UE belongs is detected, the centroid C of the spatial grid k to which it belongs is obtained. k , then according to the autocorrelation matrix H of the target UE's channel matrix H H H, obtain the updated centroid C of the spatial grid k according to the following formula k′ :

[0041]

[0042] Where N is the total number of existing channel matrices in the spatial grid k.

[0043] Before updating the centroid of the spatial grid to which the target UE channel matrix belongs, the spatial grid k to which the target UE channel matrix belongs can also be obtained according to the following formula:

[0044] k=arg max(cov(H H H,C k ))

[0045] The spatial grid k whose centroid has the largest correlation with the autocorrelation matrix of the target UE channel matrix is ​​selected as the spatial grid to which the target UE channel matrix belongs.

[0046] In another example, the base station obtains the precoding matrix of the target UE based on the centroid, including: performing singular value decomposition on the centroid to obtain a right singular matrix V; obtaining the first RI columns of V as the precoding matrix of the target UE; wherein RI is a pre-acquired rank indicator. When the base station calculates the precoding matrix of the target UE based on the centroid, it can perform the calculation based on singular value decomposition and extract the target precoding matrix based on the current rank indicator pre-provided by the adaptive modulation and coding module (Adaptive Modulation and Coding, AMC). This ensures the accuracy of the precoding matrix acquisition while simplifying the calculation process and improving acquisition efficiency.

[0047] Step 103, when it is determined that the acquisition method is based on offline learning results, the spatial grid where the target UE is located is obtained, and based on the optimal precoding matrix of each spatial grid learned offline, the optimal precoding matrix of the spatial grid where the target UE is located is obtained as the precoding matrix of the target UE.

[0048] Specifically, when the base station determines that the precoding matrix is ​​obtained based on the offline learning result according to the database status detection result, it directly obtains the spatial grid where the target UE is located, and then uses the optimal precoding matrix corresponding to the spatial grid where the target UE is located as the precoding matrix of the target UE among the optimal precoding matrices of each spatial grid determined based on the offline learning result.

[0049] In one example, the base station obtains the spatial grid where the target UE is located, including: obtaining the measurement quantity reported by the target UE; wherein the measurement quantity is used to obtain the precoding matrix recommended by the UE side, the uplink channel matrix between the target UE and the base station, or the location information of the target UE; and obtaining the spatial grid where the target UE is located based on the measurement quantity. Specifically, when determining the precoding matrix of the target UE, the base station obtains the measurement quantity that can reflect the spatial grid where the target UE is located, such as the Type IPMI, Type II PMI reported by the target UE, SRS information obtained by the base station, or the location information of the UE, and then obtains the precoding matrix recommended by the UE side, the uplink channel matrix between the target UE and the base station, or the location information of the target UE based on the obtained measurement quantity. For example, the precoding matrix recommended by the UE side is obtained based on the Type II PMI reported by the target UE. Then, based on the information calculated based on the measurement quantity, the spatial grid where the target UE is located is obtained. For example, after obtaining the precoding matrix recommended by the UE, power spectrum analysis is used to obtain the spatial angle capability distribution and calculate the beam direction during data transmission. Based on the calculated horizontal and vertical angles of the beam direction, the spatial grid where the target UE is located is determined and the spatial grid where the target UE is located is obtained. By accurately determining the spatial grid where the target UE is located based on the obtained measurement data, it is convenient to subsequently accurately obtain the precoding matrix.

[0050] In another example, the base station learns the optimal precoding matrix for each spatial grid offline in the following manner: based on the location information or beam directions of multiple UEs stored in a database and the spatial grid divided based on preset angular intervals, the spatial grid to which each UE belongs is determined, and a set of channel matrices for the UEs in each spatial grid is obtained; for each spatial grid, a constrained objective function is solved based on the set of channel matrices for the UEs in the spatial grid to obtain the optimal precoding matrix for the spatial grid, where the constraints are constraints on the precoding matrix.

[0051] Specifically, when the base station performs offline learning, it reads the stored data from the database regularly (the period can be changed) for offline learning, for example, once a day. When performing offline learning, the location information of multiple UEs stored in the database is read, and the horizontal and vertical angles of each location to the base station are calculated based on the location information. The spatial grid is then divided according to the preset angle intervals to determine the spatial grid to which each UE belongs. For example, the spatial grid is divided at intervals of 5 degrees for horizontal angle and 5 degrees for vertical angle. After completing the grid division for each spatial grid and determining the spatial grid to which each UE belongs, the channel matrix set of the UEs contained in each spatial grid is obtained based on the UEs contained. For each spatial grid, the preset constrained objective function is solved according to the channel matrix set corresponding to the spatial grid to obtain the optimal precoding matrix of the spatial grid. For example, the optimal precoding matrix in the spatial grid k is obtained by solving the following constrained function:

[0052]

[0053] ST‖W‖ 2 =1

[0054] Where C is the channel capacity;

[0055] C(H, W) = ∑log2(1+SINR), The superscript H represents the conjugate transpose, the superscript -1 represents the matrix inversion, the expression diag is to take the diagonal elements of the matrix in the brackets, Rnn is the noise covariance matrix, and if the noise covariance matrix cannot be obtained, it can be set to the unit matrix, RI is the current rank indicator obtained in advance, I RI is the identity matrix with dimension equal to the current RI.

[0056] Based on a set of candidate precoding matrices W whose squared modulus is 1, the channel capacity is calculated for different W values, and the W that maximizes the objective function value is selected as the optimal precoding matrix within the spatial grid. By solving the objective function with precoding matrix constraints using the set of channel matrices within the spatial grid during offline learning, the optimal precoding matrix for each spatial grid is accurately obtained without excessively consuming real-time computing resources, improving the communication system's adaptability to scenarios such as low signal-to-noise ratios and UE mobility.

[0057] It is worth mentioning that when performing spatial grid division, the spatial grid can also be divided according to the UE's geographical location on the map or other information that can be used to calculate or characterize the UE's location. This embodiment does not limit the specific information used when dividing the spatial grid.

[0058] Furthermore, objective functions with constraints include: maximizing channel capacity, maximizing minimum channel capacity, maximizing average signal-to-noise ratio, or maximizing minimum signal-to-noise ratio, etc. In specific applications, an appropriate objective function can be selected as needed. This embodiment does not limit the specific objective function used.

[0059] In another example, before solving a constrained objective function based on a set of channel matrices for UEs within a spatial grid, the base station further includes: performing singular value decomposition on each channel matrix within the set of channel matrices to obtain a right singular matrix V for each channel matrix; and obtaining the first RI columns of each V as a precoding matrix to be substituted into the objective function, where RI is a pre-obtained rank indicator. Specifically, when obtaining the optimal precoding matrix corresponding to the spatial grid, to reduce computational complexity, the base station may pre-singular value decomposition on each channel matrix to obtain a right singular matrix corresponding to each channel matrix. Then, based on the pre-obtained current rank indicator, multiple candidate precoding matrices consisting of the first RIs of the right singular matrix V corresponding to each channel matrix are obtained. When obtaining the optimal precoding matrix, the base station directly solves the constrained objective function based on the set of multiple candidate precoding matrices, and selects the candidate precoding matrix that maximizes the objective function as the optimal precoding matrix for the spatial grid. By narrowing the set of candidate precoding matrices, the computational complexity is greatly reduced, and the efficiency of precoding matrix generation is improved.

[0060] In addition, it should be understood that the step division of the various methods above is only for the purpose of clear description. During implementation, they can be combined into one step or some steps can be split and decomposed into multiple steps. As long as they include the same logical relationship, they are all within the scope of protection of this patent; adding insignificant modifications to the algorithm or process or introducing insignificant designs without changing the core design of the algorithm and process are all within the scope of protection of this patent.

[0061] Another aspect of the present invention provides a precoding matrix acquisition device. Figure 2 ,include:

[0062] The determination module 201 is used to determine a method for obtaining a precoding matrix; wherein the obtaining method includes online obtaining and obtaining based on offline learning results.

[0063] The first acquisition module 202 is configured to, when determining that the acquisition mode is online acquisition, acquire the centroid of the spatial grid to which the channel matrix belongs according to the clustering result of the channel matrix of the target user equipment UE, and acquire the precoding matrix of the target UE based on the centroid.

[0064] The second acquisition module 203 is used to obtain the spatial grid where the target UE is located when it is determined that the acquisition method is based on offline learning results, and based on the optimal precoding matrix of each spatial grid learned offline, obtain the optimal precoding matrix of the spatial grid where the target UE is located as the precoding matrix of the target UE.

[0065] It is not difficult to find that this embodiment is an apparatus embodiment corresponding to the method embodiment applied to the functional network element, and this embodiment can be implemented in conjunction with the method embodiment. The relevant technical details mentioned in the method embodiment are still valid in this embodiment, and to reduce repetition, they are not repeated here. Accordingly, the relevant technical details mentioned in this embodiment can also be applied to the method embodiment.

[0066] It is worth noting that all modules involved in this embodiment are logical modules. In actual applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. In addition, to highlight the innovations of the present invention, this embodiment does not include units that are not closely related to solving the technical problems proposed by the present invention. However, this does not mean that other units do not exist in this embodiment.

[0067] Another aspect of the present application embodiment further provides an electronic device, Figure 3 , including: at least one processor 301; and a memory 302 communicatively connected to the at least one processor 301; wherein the memory 302 stores instructions that can be executed by the at least one processor 301, and the instructions are executed by the at least one processor 301 to enable the at least one processor 301 to execute the precoding matrix acquisition method described in any of the above method embodiments.

[0068] The memory 302 and processor 301 are connected using a bus. The bus can include any number of interconnected buses and bridges, connecting various circuits of one or more processors 301 and memory 302. The bus can also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits. These are all well known in the art and are therefore not described further herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be a single component or multiple components, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor 301 is transmitted over a wireless medium via an antenna. Furthermore, the antenna receives data and transmits it to the processor 301.

[0069] The processor 301 is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. The memory 302 can be used to store data used by the processor 301 when performing operations.

[0070] The embodiment of the present invention further provides a computer-readable storage medium storing a computer program, which implements the above method embodiment when executed by a processor.

[0071] That is, those skilled in the art will understand that all or part of the steps in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a program, which is stored in a storage medium and includes a number of instructions for causing a device (which may be a single-chip microcomputer, chip, etc.) or a processor to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc., various media that can store program code.

[0072] Those skilled in the art will appreciate that the above embodiments are specific embodiments for implementing the present application, and that in actual applications, various changes may be made thereto in form and detail without departing from the spirit and scope of the present application.

Claims

1. A method for obtaining a precoding matrix, characterized in that: include: Determining a method for obtaining a precoding matrix; wherein the method for obtaining the matrix includes online obtaining and obtaining the matrix based on offline learning results; When it is determined that the acquisition mode is online acquisition, obtaining, according to a clustering result of the channel matrix of the target user equipment UE, a centroid of a spatial grid to which the channel matrix belongs, and obtaining a precoding matrix of the target UE based on the centroid; When it is determined that the acquisition method is based on offline learning results, the spatial grid where the target UE is located is obtained, and based on the optimal precoding matrix of each spatial grid learned offline, the optimal precoding matrix of the spatial grid where the target UE is located is obtained as the precoding matrix of the target UE.

2. The method for obtaining a precoding matrix according to claim 1, wherein: The obtaining, based on the clustering result of the channel matrix of the target user equipment UE, the centroid of the spatial grid to which the channel matrix belongs, includes: Determining, according to a clustering result of the channel matrix of the target UE, whether there is a spatial grid to which the channel matrix belongs; In a case where the spatial grid does not exist, creating a spatial grid as the spatial grid to which the target UE belongs, and determining the centroid of the spatial grid according to the channel matrix of the target UE; In a case where the spatial grid exists, the centroid of the spatial grid is updated according to the channel matrix of the target UE.

3. The method for obtaining a precoding matrix according to claim 2, wherein: The determining, based on the clustering result of the channel matrix of the target UE, whether there is a spatial grid to which the channel matrix belongs includes: Obtaining an autocorrelation matrix of a channel matrix of the target UE; Detect the correlation between the autocorrelation matrix and the centroids of each existing spatial grid, and obtain the correlation results S of each centroid k ; Where k is the number of the spatial grid, and the maximum value of k is the number of existing spatial grids; In the S k When all of them are smaller than a preset correlation threshold, it is determined that the spatial grid to which the channel matrix belongs does not exist; In the S k If there is a correlation threshold greater than or equal to the preset correlation threshold, the S k The spatial grid corresponding to the maximum value in is determined as the spatial grid to which the channel matrix belongs.

4. The method for obtaining a precoding matrix according to claim 3, wherein: Before detecting the correlation between the autocorrelation matrix and the centroids of each existing spatial grid, the method further includes: Check whether there is a spatial grid; In the case that there are currently existing spatial grids and the number of existing spatial grids is less than a preset number, detecting the correlation between the autocorrelation matrix and the centroids of each existing spatial grid is performed again; In a case where no spatial grid currently exists or the number of existing spatial grids is greater than or equal to a preset number, it is determined that no spatial grid to which the channel matrix belongs exists.

5. The method for obtaining a precoding matrix according to claim 3 or 4, wherein: The determining, according to the channel matrix of the target UE, the centroid of the spatial grid to which the target UE belongs, includes: Determining the autocorrelation matrix of the channel matrix of the target UE as the centroid of the spatial grid; The updating the centroid of the spatial grid according to the channel matrix of the target UE includes: Obtaining the centroid of the spatial grid to which it belongs; A weighted average is performed on the obtained centroid and the autocorrelation matrix of the channel matrix of the target UE to obtain an updated centroid of the spatial grid.

6. The method for obtaining a precoding matrix according to any one of claims 1 to 4, characterized in that: The acquiring, based on the centroid, a precoding matrix of the target UE, includes: Performing singular value decomposition on the centroid to obtain a right singular matrix V; Obtain the first RI column of the V as the precoding matrix of the target UE; wherein the RI is a pre-acquired rank indication.

7. The method for obtaining a precoding matrix according to claim 1, wherein: The method further includes offline learning an optimal precoding matrix for each spatial grid by: Determine the spatial grid to which each UE belongs based on the location information or beam directions of multiple UEs stored in the database and the spatial grid divided based on the preset angular interval, and obtain a channel matrix set for the UE in each spatial grid; For each of the spatial grids, an objective function with constraints is solved according to a set of channel matrices of UEs within the spatial grid to obtain an optimal precoding matrix for the spatial grid, wherein the constraints are constraints of the precoding matrix.

8. The method for obtaining a precoding matrix according to claim 7, wherein: The objectives of the objective function include: maximizing channel capacity, maximizing minimum channel capacity, maximizing average signal-to-noise ratio, or maximizing minimum signal-to-noise ratio.

9. The method for obtaining a precoding matrix according to claim 7, wherein: Before solving the objective function with constraints according to the channel matrix set of UEs in the spatial grid, the method further includes: Performing singular value decomposition on each channel matrix in the channel matrix set to obtain a right singular matrix V of each channel matrix; Obtain the first RI column of each of the V as a precoding matrix substituted into the objective function, wherein RI is a pre-obtained rank indicator.

10. The method for obtaining a precoding matrix according to any one of claims 7 to 9, wherein: The acquiring the spatial grid where the target UE is located includes: Obtaining a measurement value reported by the target UE; wherein the measurement value is used to obtain a precoding matrix recommended by the UE side, an uplink channel matrix between the target UE and the base station, or location information of the target UE; The spatial grid where the target UE is located is acquired according to the measurement amount.

11. The method for obtaining a precoding matrix according to claim 1, wherein: The method of determining the acquisition of the precoding matrix includes: Determine a method for obtaining a precoding matrix based on the amount of data of the UE's channel matrix stored in the database; Wherein, when the data amount is greater than a preset threshold, the acquisition method is determined to be the acquisition based on offline learning results; when the data amount is less than or equal to the preset threshold, the acquisition method is determined to be the online acquisition.

12. A precoding matrix acquisition device, characterized in that: include: A determination module, configured to determine a method for obtaining a precoding matrix; wherein the method for obtaining the precoding matrix includes online acquisition and acquisition based on offline learning results; a first acquisition module, configured to, when determining that the acquisition mode is online acquisition, acquire, according to a clustering result of the channel matrix of the target user equipment UE, a centroid of a spatial grid to which the channel matrix belongs, and acquire a precoding matrix of the target UE based on the centroid; The second acquisition module is used to obtain the spatial grid where the target UE is located when it is determined that the acquisition method is based on offline learning results, and based on the optimal precoding matrix of each spatial grid learned offline, obtain the optimal precoding matrix of the spatial grid where the target UE is located as the precoding matrix of the target UE.

13. An electronic device, characterized in that: include: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the precoding matrix acquisition method according to any one of claims 1 to 11.

14. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the precoding matrix acquisition method according to any one of claims 1 to 11 is implemented.

Images