A modular XL-MIMO near-field beam training method

Abstract

This invention discloses a modular XL-MIMO near-field beam training method, belonging to the field of XL-MIMO beam training technology. The method comprises: activating the antennas of the first and last modules in the array to perform far-field beam scanning, determining the codeword index with the maximum beam gain of the first and last modules, estimating the codeword index with the maximum beam gain of the entire array, and using this as the center, selecting several candidate angles, using a polar codebook for non-uniform distance sampling to generate candidate sampling points, assigning each candidate sampling point to a collaborative module for training, determining the energy of each candidate sampling point, and then determining the optimal polar near-field codeword index to complete the modular XL-MIMO near-field beam training. This method not only effectively reduces training overhead, but also maximizes the accuracy of beam gain and angle estimation under resource-limited conditions.

Description

Technical Field

[0001] The present invention relates to a modular XL-MIMO near-field beam training method, belonging to the technical field of XL-MIMO beam training. Background Art

[0002] XL-MIMO (Extremely Large-Scale Antenna Array) technology significantly improves the spatial resolution and spectral efficiency of wireless communications by adopting extremely large-scale antenna arrays (ELAA), becoming a key enabling technology for sixth-generation (6G) wireless networks.

[0003] Beam training sends a predefined training sequence through interaction between the base station and the user, and uses an algorithm to select candidate beam directions from the precoding codebook. The user end measures the channel quality corresponding to each beam and feeds back the optimal beam index or complete channel state information to the base station. Subsequently, the base station designs the beamforming vector based on the feedback information to concentrate the energy on the user's main propagation path. As a flexible technology, beam training significantly enhances the overall performance of wireless communication systems by dynamically optimizing the beam direction to reduce interference, improve signal quality and system capacity. In XL-MIMO systems, the increase in the number of antennas leads to a significant increase in the number of symbols required for beam training, which greatly increases the training overhead. At the same time, traditional far-field beam training methods often lead to significant performance degradation when applied to near-field communication systems.

[0004] To address these issues, a polarization-domain codebook specifically designed for the near field was proposed. The angular domain is uniformly sampled, while the range domain is non-uniformly sampled. This reduces the codebook size and mitigates the high training cost associated with complex near-field codebooks. To reduce the training overhead of beam training, an efficient two-stage beam training method was proposed. Specifically, the first stage determines candidate user positions through far-field beam scanning. The second stage uses a customized polarization-domain codebook to find the optimal effective range for the user given the candidate angles. To further reduce training overhead, a hierarchical beam training method was proposed. In the first stage, a far-field hierarchical codebook search is performed using the central subarray of the ELAA to obtain a coarse user direction estimate. In the second stage, based on the coarse user direction, a designed polar-domain hierarchical codebook is used to further refine the user direction and range information. Furthermore, a near-field beam training method based on deep learning was proposed. Specifically, a near-field codebook containing angle and range information is used to train deep neural networks (DNNs) to reduce training overhead. A near-field beam training method with supplementary codewords is proposed to improve beamforming performance. However, existing XL-MIMO beam training schemes still face challenges in terms of training overhead and computational complexity.

[0005] Although non-modular beam training schemes have significant advantages in signal processing accuracy and system coordination, their limitations in scalability and computational complexity have prompted researchers to turn to the exploration of modular beam training. Modular beam training can effectively share the computational load and improve processing speed and efficiency by distributing computational tasks to multiple nodes. Specifically, a modular two-stage beam training scheme is proposed. In the first stage, each subarray independently uses multiple far-field channel steering vectors for analog combining. In the second stage, for each codeword in the predefined mixed field codebook, a dedicated digital combiner is designed to combine the outputs of the analog combiner in the first stage. The codeword corresponding to the dedicated digital combiner is selected from the mixed field codebook to achieve maximum combining power. However, in the case of downlink beam training, this scheme still faces large beam training overhead and high computational complexity. Summary of the Invention

[0006] The purpose of the present invention is to provide a modular XL-MIMO near-field beam training method, which can not only effectively reduce the training overhead, but also maximize the accuracy of beam gain and angle estimation under resource-limited conditions.

[0007] In order to achieve the above object, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides a modular XL-MIMO near-field beam training method, which is applied to a modular XL-MIMO system in which a base station deploys a uniform linear array. The array is divided into a plurality of modules, each module having a plurality of antennas, and each module serving a user through a radio frequency chain. The method comprises:

[0009] Activate the antennas of the first and last modules in the array to perform far-field beam scanning and determine the codeword index with the largest beam gain for the first and last modules;

[0010] Estimate the codeword index with the largest beam gain of the entire array based on the codeword index with the largest beam gain of the first and last modules;

[0011] Taking the codeword index with the largest beam gain of the entire array as the center, several candidate angles are selected to form a candidate angle index set;

[0012] For each candidate angle in the candidate angle index set, a polar codebook is used to perform non-uniform distance sampling to generate candidate sampling points, and a candidate sampling point index set is formed from all candidate sampling points;

[0013] Assign each candidate sampling point in the candidate sampling point index set to the collaborative module for training, and determine the energy of each candidate sampling point;

[0014] Based on the energy of each candidate sampling point, the optimal polar near-field codeword index is determined to complete modular XL-MIMO near-field beam training.

[0015] In combination with the first aspect, further, the codeword index with the largest beam gain of the entire array is:

[0016] ;

[0017] in, Indicates the codeword index with the largest beam gain of the entire array, Indicates rounding. Indicates the total number of modules in the array, Indicates the number of antennas per module, The sine of the actual angle where the beam gain of the entire array is maximum, ,in, 、 Represents the codeword index with the largest beam gain of the first and last modules respectively.

[0018] In combination with the first aspect, further, the candidate angle index set is:

[0019] ;

[0020] in, represents the candidate angle index set, Indicates the codeword index with the largest beam gain of the entire array, Indicates rounding down. Indicates the total number of candidate angles.

[0021] In combination with the first aspect, further, the sampling distance corresponding to each candidate sampling point is:

[0022] ;

[0023] in, Indicates the Candidate angles No. The sampling distance corresponding to the candidate sampling points, Indicates the total number of candidate sampling points corresponding to each candidate angle, represents the threshold distance that limits the column coherence between near-field steering vectors, ,in, represents the correlation coefficient between codewords, represents the wavelength of the spherical wave, Indicates the dimensions of the array.

[0024] In combination with the first aspect, further, the cooperation module is composed of the first modules and modules, ,in, Indicates the total number of modules in the array, Indicates rounding down;

[0025] When each candidate sampling point in the candidate sampling point index set is assigned to the collaborative module for training, it is also assigned to two modules in the collaborative module.

[0026] In combination with the first aspect, further determining the energy of each candidate sampling point includes:

[0027] Calculate the far-field codeword indexes of the two modules in the collaborative module corresponding to each candidate sampling point;

[0028] Based on the far-field codeword indexes of the two modules in the collaborative module corresponding to each candidate sampling point, the two modules in the collaborative module corresponding to each candidate sampling point are controlled to send pilot signals respectively. The received signals of the user based on the collaborative module corresponding to each candidate sampling point are superimposed to obtain the energy of each candidate sampling point.

[0029] In combination with the first aspect, further, calculating the far-field codeword indexes of two modules in the collaborative module corresponding to each candidate sampling point includes:

[0030] Based on the trigonometric function relationship, the sine values ​​of the two modules in the collaborative module corresponding to each candidate sampling point are calculated;

[0031] Calculate the far-field codeword indexes of the two modules in the collaborative module corresponding to each candidate sampling point based on the sine values ​​of the two modules in the collaborative module corresponding to each candidate sampling point;

[0032] The calculation formula for the sine values ​​of the two modules in the collaborative module corresponding to each candidate sampling point is:

[0033] ;

[0034] in, 、 Indicates the Candidate angles No. The sine values ​​of the two modules in the collaborative module corresponding to the candidate sampling points, Indicates the Candidate angles No. The angle of the candidate sampling point relative to the center antenna of the entire array, Indicates the Candidate angles No. The sampling distance corresponding to the candidate sampling points, 、 Indicates the The position coordinates of the central antennas of the two modules in the collaborative module, Indicates the number of antennas per module, Indicates the spacing distance between adjacent antennas;

[0035] The calculation formula for the far-field codeword index of the two modules in the collaborative module corresponding to each candidate sampling point is:

[0036] ;

[0037] in, 、 Indicates the Candidate angles No. The far-field codeword indexes of the two modules in the collaborative module corresponding to the candidate sampling points, 、 Indicates that for For two modules in a collaborative module, to Divide evenly into After the The sine of the angle;

[0038] The received signal of the user based on the collaborative module corresponding to each candidate sampling point is:

[0039] ;

[0040] in, Indicates when hour The value of Indicates when hour The value of 、 Indicates that the user is based on Candidate angles No. The received signals of the two modules in the collaborative module corresponding to the candidate sampling points, 、 Indicates the The channel between two modules and users in a collaborative module, represents the transposed conjugate, 、 Respectively express about 、 The steering vector, Indicates the pilot signals sent by the two modules in each collaborative module, 、 Indicates the Additive Gaussian white noise of two modules in a collaborative module;

[0041] The energy of each candidate sampling point is:

[0042] ;

[0043] in, Indicates the Candidate angles No. The energy of the candidate sampling points.

[0044] In combination with the first aspect, further, the optimal extreme near-field codeword index is:

[0045] ;

[0046] in, The best angle among the corresponding candidate angles, The best sampling point among the corresponding candidate sampling points, Indicates the Candidate angles No. The energy of the candidate sampling points.

[0047] In a second aspect, the present invention provides a computer device, comprising:

[0048] Storage medium for storing computer programs;

[0049] A processor is configured to execute the computer program to implement the modular XL-MIMO near-field beam training method according to the first aspect.

[0050] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the modular XL-MIMO near-field beam training method described in the first aspect.

[0051] In a fourth aspect, the present invention provides a computer program product, comprising a computer program, which, when executed by a processor, implements the modular XL-MIMO near-field beam training method according to the first aspect.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] The modular XL-MIMO near-field beam training method provided by the present invention adopts a partial antenna activation strategy, activating only the antennas of the first and last modules of the array for far-field beam scanning. This can determine the candidate angles of the first and last modules while greatly reducing the number of symbols required for beam training. Since the antenna array is divided into multiple modules, the number of antennas in each module is reduced compared to the entire array, which results in a corresponding reduction in the number of codewords and resolution. In order to improve the estimation accuracy of the actual angle with the maximum energy, the two modules that are farthest apart are selected for calculation. This strategy can not only effectively reduce training overhead, but also maximize the accuracy of beam gain and angle estimation under resource-limited conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 1 is a schematic diagram showing a performance comparison between the modular XL-MIMO near-field beam training method provided by an embodiment of the present invention and the existing mixed-field beam training method and fast beam training method under different signal-to-noise ratios;

[0055] Figure 2 1 is a schematic diagram showing a performance comparison of the modular XL-MIMO near-field beam training method provided by an embodiment of the present invention compared with the existing mixed-field beam training method and fast beam training method under different overhead constraints;

[0056] Figure 3 This is a schematic diagram comparing the performance of the modular XL-MIMO near-field beam training method provided by an embodiment of the present invention compared with the existing mixed-field beam training method and the fast beam training method when the number of antennas in each module is fixed and the number of modules is changed. DETAILED DESCRIPTION

[0057] The technical solution of the present invention will be further described in detail below in conjunction with specific implementation methods.

[0058] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention and are not to be construed as limiting the present invention. The embodiments of the present invention and the technical features in the embodiments may be combined with each other unless there is a conflict.

[0059] An embodiment of the present invention provides a modular XL-MIMO near-field beam training method, which is applied to a modular XL-MIMO system in which a base station deploys a uniform linear array. The array is divided into several modules, each module has several antennas, and each module serves users through a radio frequency chain.

[0060] In this embodiment, the modular XL-MIMO near-field beam training method includes:

[0061] Activate the antennas of the first and last modules in the array to perform far-field beam scanning and determine the codeword index with the largest beam gain for the first and last modules;

[0062] Estimate the codeword index with the largest beam gain of the entire array based on the codeword index with the largest beam gain of the first and last modules;

[0063] Taking the codeword index with the largest beam gain of the entire array as the center, several candidate angles are selected to form a candidate angle index set;

[0064] For each candidate angle in the candidate angle index set, a polar codebook is used to perform non-uniform distance sampling to generate candidate sampling points, and a candidate sampling point index set is formed from all candidate sampling points;

[0065] Assign each candidate sampling point in the candidate sampling point index set to the collaborative module for training, and determine the energy of each candidate sampling point;

[0066] Based on the energy of each candidate sampling point, the optimal polar near-field codeword index is determined to complete modular XL-MIMO near-field beam training.

[0067] The embodiment of the present invention provides a modular XL-MIMO near-field beam training method, which aims to solve the problems of high training overhead and high computational complexity in traditional near-field beam training, while meeting the system's requirements for average achievable rate. By utilizing the geometric relationship between modules and a partially connected hybrid architecture, efficient and flexible beamforming is achieved by distributing computing tasks to multiple collaborative modules. In the first stage, candidate angles are determined by partial far-field beam scanning, significantly narrowing the search range and thus reducing training overhead; in the second stage, the polar codebook and modular beam training method are combined to collaboratively search for near-field codewords under candidate angles to accurately determine the optimal distance of the user and improve the accuracy and efficiency of distance estimation.

[0068] The embodiment of the present invention provides a modular XL-MIMO near-field beam training method, which is applied to a modular XL-MIMO system in which a base station deploys a uniform linear array. The array is divided into modules, each with antennas, the base station has a common antennas, each module serves a single antenna user through a radio frequency chain, and the base station A radio frequency chain communicates with a single-antenna user and transmits signals to the single-antenna user.

[0069] In this embodiment, the modular XL-MIMO near-field beam training method specifically includes the following steps:

[0070] Step 1: Build a channel model;

[0071] In this embodiment, the channel model is:

[0072] ;

[0073] in, Indicates that the user receives the signal, , 、…、 Respectively represent the user based on the 1st, ..., The receiving signal of each module, represents the symbol matrix transmitted by the base station to the user, represents the simulated beamforming matrix for the entire array, The value of depends on the channel state information. Used for Perform operations to optimize system performance, represents the channel between the base station and the user, , 、…、 Respectively represent the 1st, ..., The channel between the root antenna and the user, represents the transposed conjugate, represents additive white Gaussian noise.

[0074] Considering the single scattering loop and Rayleigh fading model, the channel between the module and the user can be expressed as:

[0075] ;

[0076] in, Indicates the The channel between the module and the user, It obeys Gaussian distribution, represents the channel covariance matrix, .

[0077] Assume that the scattering ring is relative to the The location of the modules ,in, represents the radius of the scattering ring, represents the angle along the scattering ring, Indicates the center of the scattering ring and the The distance between the center antennas of each module, Indicates the center of the scattering ring and the The angle between the center antennas of each module, Represents transposition, then the specific form of channel covariance can be expressed as:

[0078] ;

[0079] in, express No. Rank Column elements, Indicates the average received energy of the module's central antenna, represents the wavelength of the spherical wave, Indicates the scattering ring and the The distance between the center antennas of each module, ,in, represents the distance between the scattering ring and the center antenna of the entire array, Indicates the The position coordinates of the central antenna of each module, represents the angle of the scattering ring relative to the reference antenna, Express about The probability density function of follows the von Mises distribution, ,in, represents the zero-order Bessel function, represents the concentration of the probability distribution, , Indicates the angle at which the probability density function peaks.

[0080] Defining the steering vector ,in, Indicates the angle of the user relative to the center antenna of the module, Indicates the distance between the user and the module's central antenna, Indicates the user and module For each module, the user is usually in the far field area, that is, , then the steering vector can also be expressed as:

[0081] .

[0082] Assumptions ,in, 、…、 Respectively represent the 1st, ..., The analog beamforming matrix of each module is calculated by the module's far-field discrete Fourier transform codebook. To design, ,in, 、…、 、…、 Respectively 1st, ..., 、…、 Code words, ,in, Indicates about The steering vector, Indicates that to Divide evenly into After the The sine of the angle, .

[0083] Base station selection The codeword in is sent to the user, and the user Module No. Received signal for:

[0084] ;

[0085] in, Indicates the Modules from The selected Code words, Indicates the The pilot signal sent by each module, express No. elements.

[0086] Step 2: Convert the beam training problem into the problem of designing the precoding matrix;

[0087] In this embodiment, the beam training problem is transformed into a precoding matrix design problem, and the spectrum efficiency is maximized by designing the precoding matrix. To maximize the spectrum efficiency. Therefore, the objective function of the beam training problem can be expressed as:

[0088] .

[0089] To solve the objective function, Every codeword in the code needs to be tested, and this method is called far-field beam scanning.

[0090] Definition Combiners , then the user is based on The received signal for:

[0091] .

[0092] Step 3: Activate the antennas of the first and last modules in the array to perform far-field beam scanning and determine the codeword index with the largest beam gain for the first and last modules;

[0093] In this embodiment, the huge overhead required for far-field beam scanning is addressed by selecting candidate angles. A partial antenna activation strategy is adopted, where only the antennas of the first and last modules in the array are activated for far-field beam scanning. This allows the candidate angles of the first and last modules in the array to be determined, while significantly reducing the number of symbols required for beam training. Since the array is divided into multiple modules, the number of antennas in each module is reduced compared to the entire array. This results in a corresponding reduction in the number of codewords and resolution. To improve the accuracy of estimating the actual angle with the greatest beam gain, the two modules farthest apart can be selected for calculation. This strategy not only effectively reduces training overhead but also maximizes the accuracy of beam gain and angle estimation under resource-constrained conditions.

[0094] Specifically, define the precoding matrix ,in, express The zero matrix of It can be expressed as:

[0095] .

[0096] After the far-field beam is scanned, the 、…、 、…、 The received signal matrix No. 1 column vectors 、 , we can get the codeword index with the largest beam gain of the first and last modules 、 ,in, 、…、 、…、 Respectively represent the user based on the 1st, ..., 、…、 Combiners 、…、 、…、 The received signal, express The zero vector of express The matrix of .

[0097] Step 4: Based on the codeword index with the largest beam gain of the first and last modules, estimate the codeword index with the largest beam gain of the entire array;

[0098] Taking into account the geometric relationship between the scattering ring power position spectrum and the modules, the codeword index with the maximum beam gain of the entire array can be roughly estimated by the codeword index with the maximum beam gain of the first and last modules.

[0099] Sine of the actual angle at which beam gain is maximum The codeword index with the largest beam gain The relationship between Given. The codeword index with the largest beam gain through the first and last modules 、 The sine value of the actual angle at which the beam gain of the head and tail modules is maximum can be derived 、 .

[0100] The sine of the actual angle at which the beam gain is maximum for the entire array It can be expressed as:

[0101] .

[0102] Given that the total number of antennas in the entire array is , for the entire array, the codeword index with the largest beam gain It can be expressed as:

[0103] ;

[0104] in, Indicates rounding.

[0105] Step 5: Taking the codeword index with the largest beam gain of the entire array as the center, select several candidate angles to form a candidate angle index set;

[0106] Due to the existence of power fluctuation, receiving noise and small-scale fading, step 4 is The estimation of may not be accurate enough. To solve this problem, this embodiment adopts a new angle selection scheme.

[0107] Estimated in step 4 As the center, select Candidate angles form a candidate angle index set ,in, Indicates rounding down. Indicates the total number of candidate angles.

[0108] Step 6: For each candidate angle in the candidate angle index set, perform non-uniform distance sampling using the polar codebook to generate candidate sampling points, and form a candidate sampling point index set from all candidate sampling points;

[0109] This embodiment provides a customized polar beam training method, namely, effective distance estimation, for estimating the effective distance of a scattering ring.

[0110] Specifically, a polar domain codebook is used, and each codeword in the polar domain codebook corresponds to a specific angle distance pair.

[0111] For each candidate angle in the candidate angle index set, a polar field codebook is used to perform non-uniform distance sampling to generate candidate sampling points, and a candidate sampling point index set is formed by all candidate sampling points.

[0112] The sampling distance corresponding to each candidate sampling point is:

[0113] ;

[0114] in, Indicates the Candidate angles No. The sampling distance corresponding to the candidate sampling points, Indicates the total number of candidate sampling points corresponding to each candidate angle, represents the threshold distance that limits the column coherence between near-field steering vectors, ,in, represents the correlation coefficient between codewords, represents the wavelength of the spherical wave, Indicates the dimensions of the array.

[0115] For each candidate angle, generate candidate sampling points, The candidate sampling points are arranged in ascending order according to the sampling distance to form a candidate sampling point index subset.

[0116] for candidate angles, generating candidate sampling points, forming a total of A subset of candidate sampling point indices.

[0117] According to the order of candidate angles from small to large, the candidate sampling points with the smallest sampling distance are selected from the candidate sampling point index subset corresponding to the candidate angles and arranged, and all candidate sampling points are selected in a non-repeated cycle to form a candidate sampling point index set. ,in, express Middle candidate sampling points.

[0118] Step 7: Assign each candidate sampling point in the candidate sampling point index set to the collaborative module for training, and determine the energy of each candidate sampling point;

[0119] This embodiment addresses the problem of polar sampling by using a modular effective distance estimation method. Due to the reduced near-field effect, users are located in the module's far field, making traditional near-field beam scanning methods inapplicable. This embodiment uses two modules to transmit angular codewords to candidate sampling points and uses the received signals to determine the energy of each candidate sampling point.

[0120] In this embodiment, a combination of two modules for determining the energy of candidate sampling points is defined as a collaborative module.

[0121] Specifically, the first modules and The modules consist of Collaborative modules, ,in, Indicates the total number of collaborative modules, ,in, Indicates the total number of modules in the array, Indicates rounding down.

[0122] In this embodiment, Each candidate sampling point in the cyclic allocation is assigned to the first,... The collaborative modules are trained to determine the energy of each candidate sampling point. When the energy is allocated to the collaborative module, it is allocated to two modules in the collaborative module at the same time to perform effective distance scanning.

[0123] The candidate sampling point index set assigned to each collaborative module is:

[0124] ;

[0125] in, Indicates the The candidate sampling point index set assigned to the collaborative module, express Middle candidate sampling points, i.e. The collaborative module is assigned to candidate sampling points. The collaborative module is assigned to Candidate sampling points are Index in .

[0126] In this embodiment, determining the energy of each candidate sampling point specifically includes the following steps:

[0127] Step 1: Calculate the far-field codeword indexes of the two modules in the collaborative module corresponding to each candidate sampling point;

[0128] In this embodiment, calculating the far-field codeword indexes of two modules in the collaborative module corresponding to each candidate sampling point specifically includes the following steps:

[0129] Step ①: Based on the trigonometric function relationship, calculate the sine values ​​of the two modules in the collaborative module corresponding to each candidate sampling point;

[0130] In this embodiment, based on the candidate sampling point index set allocated to each collaborative module, the sine values ​​of the two modules in the collaborative module corresponding to each candidate sampling point are calculated respectively using a trigonometric function relationship.

[0131] The calculation formula for the sine values ​​of the two modules in the collaborative module corresponding to each candidate sampling point is:

[0132] ;

[0133] in, 、 Indicates the Candidate angles No. The collaborative module corresponding to the candidate sampling point ( The sine values ​​of two modules in a collaborative module, Indicates the Candidate angles No. The angle of the candidate sampling point relative to the center antenna of the entire array, 、 Indicates the The position coordinates of the central antennas of the two modules in the collaborative module, Indicates the spacing distance between adjacent antennas.

[0134] Step ②: Based on the sine values ​​of the two modules in the cooperative module corresponding to each candidate sampling point, calculate the far-field codeword indexes of the two modules in the cooperative module corresponding to each candidate sampling point.

[0135] In this embodiment, the calculation formula for the far-field codeword indexes of the two modules in the collaborative module corresponding to each candidate sampling point is:

[0136] ;

[0137] in, 、 Indicates the Candidate angles No. The collaborative module corresponding to the candidate sampling point ( The far-field codeword index of the two modules in the collaborative module), 、 Indicates that for For two modules in a collaborative module, to Divide evenly into After the The sine of an angle.

[0138] Step 2: Based on the far-field codeword indexes of the two modules in the collaborative module corresponding to each candidate sampling point, control the two modules in the collaborative module corresponding to each candidate sampling point to send pilot signals respectively, and superimpose the user's received signals based on the collaborative module corresponding to each candidate sampling point to obtain the energy of each candidate sampling point.

[0139] In this embodiment, two modules in the collaborative module corresponding to each candidate sampling point are controlled to send pilot signals respectively. , then the user's received signal based on the collaborative module corresponding to each candidate sampling point is:

[0140] ;

[0141] in, Indicates when hour The value of Indicates when hour The value of 、 Indicates that the user is based on Candidate angles No. The collaborative module corresponding to the candidate sampling point ( The received signals of the two modules in the collaborative module are 、 Indicates the The channel between two modules and users in a collaborative module, 、 Respectively express about 、 The steering vector, 、 Indicates the Additive Gaussian white noise of two modules in a collaborative module.

[0142] The energy of each candidate sampling point is:

[0143] ;

[0144] in, Indicates the Candidate angles No. The energy of the candidate sampling points.

[0145] Step 8: Determine the optimal polar near-field codeword index based on the energy of each candidate sampling point, and complete the modular XL-MIMO near-field beam training.

[0146] In this embodiment, the optimal polar near-field codeword index is:

[0147] ;

[0148] in, The best angle among the corresponding candidate angles, The best sampling point among the corresponding candidate sampling points.

[0149] To verify the modular XL-MIMO near-field beam training method provided in an embodiment of the present invention, the modular XL-MIMO near-field beam training method provided in an embodiment of the present invention was applied to a modular XL-MIMO narrowband downlink communication system for simulation, and the performance of the beam training was evaluated using the average achievable rate (ESE).

[0150] Specifically, the array is divided into 64 modules, each module has 6 antennas, and the base station has a total of 384 antennas. Each module serves a single-antenna user through a radio frequency chain (RF chain). The base station communicates with single-antenna users through a total of 64 RF chains and transmits signals to single-antenna users.

[0151] The modular XL-MIMO near-field beam training method provided by the embodiment of the present invention has a performance comparison under different signal-to-noise ratios compared with the existing mixed-field beam training method and fast beam training method. Figure 1 As shown in the figure, the performance comparison under different overhead constraints is as follows: Figure 2As shown in the figure, the performance comparison of the case where the number of antennas in each module is fixed and the number of modules is changed is as follows: Figure 3 shown.

[0152] Depend on Figures 1 to 3 It can be seen that the modular XL-MIMO near-field beam training method provided by the embodiment of the present invention reduces the training overhead of traditional near-field beam training while maintaining comparable beamforming performance, compared to the existing mixed-field beam training method and fast beam training method.

[0153] An embodiment of the present invention provides a computer device, including:

[0154] Storage medium for storing computer programs;

[0155] A processor is configured to execute a computer program to implement the modular XL-MIMO near-field beam training method provided by any embodiment of the present invention.

[0156] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the modular XL-MIMO near-field beam training method provided by any embodiment of the present invention is implemented.

[0157] An embodiment of the present invention provides a computer program product, including a computer program. When the computer program is executed by a processor, the computer program implements the modular XL-MIMO near-field beam training method provided by any embodiment of the present invention.

[0158] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0159] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0160] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0161] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0162] The above are only preferred embodiments of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A modular XL-MIMO near-field beam training method is applied to a modular XL-MIMO system with a base station deploying a uniform linear array. The array is divided into several modules, each module has several antennas, and each module serves users through a radio frequency chain. The method is characterized by: include: Activate the antennas of the first and last modules in the array to perform far-field beam scanning and determine the codeword index with the largest beam gain for the first and last modules; Estimate the codeword index with the largest beam gain of the entire array based on the codeword index with the largest beam gain of the first and last modules; Taking the codeword index with the largest beam gain of the entire array as the center, several candidate angles are selected to form a candidate angle index set; For each candidate angle in the candidate angle index set, a polar codebook is used to perform non-uniform distance sampling to generate candidate sampling points, and a candidate sampling point index set is formed from all candidate sampling points; Assign each candidate sampling point in the candidate sampling point index set to the collaborative module for training, and determine the energy of each candidate sampling point; Based on the energy of each candidate sampling point, the optimal polar near-field codeword index is determined to complete modular XL-MIMO near-field beam training; The codeword index with the largest beam gain of the entire array is: ； in, Indicates the codeword index with the largest beam gain of the entire array, Indicates rounding. Indicates the total number of modules in the array, Indicates the number of antennas per module, The sine of the actual angle where the beam gain of the entire array is maximum, ,in, 、 Represents the codeword index with the largest beam gain of the first and last modules respectively; The collaborative module consists of the modules and modules, ,in, Indicates the total number of modules in the array, Indicates rounding down; When each candidate sampling point in the candidate sampling point index set is assigned to the collaborative module for training, it is also assigned to two modules in the collaborative module; Determining the energy of each candidate sampling point includes: Calculate the far-field codeword indexes of the two modules in the collaborative module corresponding to each candidate sampling point; Based on the far-field codeword indexes of the two modules in the collaborative module corresponding to each candidate sampling point, the two modules in the collaborative module corresponding to each candidate sampling point are controlled to send pilot signals respectively. The received signals of the user based on the collaborative module corresponding to each candidate sampling point are superimposed to obtain the energy of each candidate sampling point.

2. The modular XL-MIMO near-field beam training method according to claim 1, characterized in that: The candidate angle index set is: ； in, represents the candidate angle index set, Indicates the codeword index with the largest beam gain of the entire array, Indicates rounding down. Indicates the total number of candidate angles.

3. The modular XL-MIMO near-field beam training method according to claim 1, characterized in that: The sampling distance corresponding to each candidate sampling point is: ； in, Indicates the Candidate angles No. The sampling distance corresponding to the candidate sampling points, Indicates the total number of candidate sampling points corresponding to each candidate angle, represents the threshold distance that limits the column coherence between near-field steering vectors, ,in, represents the correlation coefficient between codewords, represents the wavelength of the spherical wave, Indicates the dimensions of the array.

4. The modular XL-MIMO near-field beam training method according to claim 1, wherein: Calculating the far-field codeword indexes of the two modules in the collaborative module corresponding to each candidate sampling point includes: Based on the trigonometric function relationship, the sine values ​​of the two modules in the collaborative module corresponding to each candidate sampling point are calculated; Calculate the far-field codeword indexes of the two modules in the collaborative module corresponding to each candidate sampling point based on the sine values ​​of the two modules in the collaborative module corresponding to each candidate sampling point; The calculation formula for the sine values ​​of the two modules in the collaborative module corresponding to each candidate sampling point is: ； in, 、 Indicates the Candidate angles No. The sine values ​​of the two modules in the collaborative module corresponding to the candidate sampling points, Indicates the Candidate angles No. The angle of the candidate sampling point relative to the center antenna of the entire array, Indicates the Candidate angles No. The sampling distance corresponding to the candidate sampling points, 、 Indicates the The position coordinates of the central antennas of the two modules in the collaborative module, Indicates the number of antennas per module, Indicates the spacing distance between adjacent antennas; The calculation formula for the far-field codeword index of the two modules in the collaborative module corresponding to each candidate sampling point is: ； in, 、 Indicates the Candidate angles No. The far-field codeword indexes of the two modules in the collaborative module corresponding to the candidate sampling points, 、 Indicates that for For two modules in a collaborative module, to Divide evenly into After the The sine of the angle; The received signal of the user based on the collaborative module corresponding to each candidate sampling point is: ； in, Indicates when hour The value of Indicates when hour The value of 、 Indicates that the user is based on Candidate angles No. The received signals of the two modules in the collaborative module corresponding to the candidate sampling points, 、 Indicates the The channel between two modules and users in a collaborative module, represents the transposed conjugate, 、 Respectively express about 、 The steering vector, Indicates the pilot signals sent by the two modules in each collaborative module, 、 Indicates the Additive Gaussian white noise of two modules in a collaborative module; The energy of each candidate sampling point is: ； in, Indicates the Candidate angles No. The energy of the candidate sampling points.

5. The modular XL-MIMO near-field beam training method according to claim 1, wherein: The best extreme near-field codeword index is: ； in, The best angle among the corresponding candidate angles, The best sampling point among the corresponding candidate sampling points, Indicates the Candidate angles No. The energy of the candidate sampling points.

6. A computer device, characterized in that: include: Storage medium for storing computer programs; A processor, configured to execute the computer program to implement the modular XL-MIMO near-field beam training method according to any one of claims 1 to 5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the modular XL-MIMO near-field beam training method according to any one of claims 1 to 5 is implemented.

Images