Broadband XL-MIMO-oriented near-field CSI distributed estimation method

By dividing the base station antenna units into subarrays and performing distributed estimation in a broadband XL-MIMO system, the calculation complexity and pilot overhead of full-dimensional near-field CSI under a hybrid architecture are solved, and high-precision channel reconstruction is achieved.

CN120434081AActive Publication Date: 2025-08-05SOUTHEAST UNIV

Patent Information

Application Number
CN202510507522.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-04-22
Publication Date
2025-08-05
Estimated Expiration
2045-04-22

AI Technical Summary

Technical Problem

In broadband XL-MIMO systems, the acquisition of full-dimensional near-field CSI under hybrid architecture has problems such as high computational complexity, large pilot overhead and insufficient accuracy, which is difficult to meet the estimation needs of low complexity and high precision.

Method used

The distributed estimation method is adopted to divide the base station antenna unit array into subarrays. Each subarray is connected to the local processing unit through independent radio frequency chains, and the central processing unit performs parameter fusion and independent estimation to realize linear decoupling and reconstruction of near-field channel parameters.

Benefits of technology

It reduces the computational complexity, reduces pilot overhead, and improves the accuracy of near-field channel reconstruction, meeting the low complexity and high-precision estimation requirements of broadband XL-MIMO systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120434081A_ABST
    Figure CN120434081A_ABST
Patent Text Reader

Abstract

The invention discloses a near-field CSI distributed estimation method for broadband XL-MIMO, and the method comprises the steps: firstly, enabling a center LPU to estimate a path time delay parameter of a center sub-array channel, enabling other LPUs to rapidly extrapolate the path time delay parameters of other sub-array channels, and transmitting an estimation result to a CPU; then, based on near-field wave beam squint and parameter symmetry characteristics brought by a large array and a large bandwidth, the CPU fuses time delay parameters estimated by different sub-arrays, and decouples and independently estimates near-field channel parameters including angle, distance and range parameters; then, the CPU transmits the estimated near-field channel parameters back to the LPUs, and each LPU independently estimates the path gain of the sub-array channel; and finally, transmitting the estimated path gain to a CPU (Central Processing Unit) by the LPUs, and further reconstructing a full-dimensional broadband near-field channel. The problems that the CSI acquisition calculation complexity is too high and the reconstruction precision is low in a large-array and large-bandwidth scene are solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a near-field CSI distributed estimation method for broadband XL-MIMO, and belongs to the technical field of near-field channel state information estimation. Background Art

[0002] Ultra-large-scale multiple-input, multiple-output (XL-MIMO) is considered a promising technology for sixth-generation (6G) wireless communication systems, addressing the rapidly growing demands for spectrum efficiency and capacity. By deploying a significantly increased number of antenna elements (thousands or even more) at base stations (BSs), future wireless communications can be provided with enhanced coverage, higher data rates, and improved quality of service for user equipment (UE). Furthermore, higher-frequency spectrum bands, such as the upper half of 6 GHz (U6G), from 6.425 GHz to 7.125 GHz, are emerging as a strategic choice for 6G international mobile communications services. These bands offer abundant spectrum resources, while higher frequencies allow for further reductions in antenna size, enabling the deployment of XL-MIMO within limited areas. Therefore, wideband XL-MIMO systems hold great potential for future wireless communications.

[0003] To fully utilize the beamforming gain of wideband XL-MIMO, accurate channel state information (CSI) is required. However, the large array and wide bandwidth deployment in wideband XL-MIMO results in near-field beam squint in the channel, which no longer exhibits spatial sparsity, making traditional compressed sensing-based algorithms inapplicable. Furthermore, the need to estimate more parameters in the near-field channel, coupled with the high-dimensional received signal resulting from the large array and wide bandwidth, results in unacceptable computational complexity for CSI acquisition. Furthermore, to reduce hardware costs, XL-MIMO typically deploys a hybrid precoding architecture, connecting all antenna elements via a small number of RF chains. This hybrid precoding architecture results in a more complex received signal format, and the UE needs to send multiple uplink pilots to compensate for the dimensionality of the received signal. This results in excessive pilot overhead and reduced accuracy for high-dimensional near-field CSI acquisition under the hybrid precoding architecture. Therefore, low-complexity estimation of wideband near-field CSI under the hybrid architecture has become an important and pressing challenge for XL-MIMO systems.

[0004] Preliminary research has been conducted on acquiring full-dimensional near-field CSI in hybrid-architecture wideband XL-MIMO systems. A widely adopted strategy currently analyzes the beam squint characteristics of wideband near-field signals, constructs a frequency-dependent near-field parameter codebook, and then detects near-field channel parameters for multiple paths. However, in practical wideband XL-MIMO systems, the number of antenna elements and subcarriers is too large, and using codebooks to estimate parameters incurs significant computational complexity. Furthermore, users need to send multiple uplink pilots to compensate for the dimensionality of the received signal, resulting in significant pilot overhead. Furthermore, the new parameter characteristics introduced by the large array and wide bandwidth are not fully utilized. Consequently, the estimation accuracy of full-dimensional near-field CSI is limited, and the overhead is excessive. This approach cannot meet the low computational complexity and high-precision full-dimensional near-field CSI estimation requirements of practical hybrid-architecture wideband XL-MIMO systems.

[0005] In summary, how to obtain high-precision hybrid architecture broadband XL-MIMO full-dimensional near-field CSI with low computational complexity has become a difficult problem that needs to be overcome in the sixth generation of mobile communications. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a distributed near-field CSI estimation method for broadband XL-MIMO, breaking through the bottleneck of full-dimensional near-field CSI acquisition in hybrid-architecture broadband ultra-large-scale MIMO systems. The method can achieve full-dimensional near-field CSI reconstruction in hybrid-architecture broadband XL-MIMO systems with relatively low hardware cost, extremely low computational complexity, and single pilot overhead, while ensuring the high-precision characteristics of the reconstructed channel.

[0007] The present invention adopts the following technical solutions to solve the above technical problems:

[0008] A near-field CSI distributed estimation method for broadband XL-MIMO. In a broadband XL-MIMO wireless transmission system, the base station is equipped with a uniformly distributed antenna array. The entire antenna array is divided into K sub-arrays, where K is an even number. All antenna elements in each sub-array are connected to G independent RF chains through analog phase shifters, where 1≤G≤N. s , N s is the number of antenna elements in each subarray, G radio frequency chains process received signals through a local processing unit, K local processing units are connected to the same central processing unit, and information is exchanged between the K local processing units; the method comprises the following steps:

[0009] Step 1: The user sends an uplink broadband pilot signal in the U6G or higher frequency band. The local processing unit corresponding to the center subarray of the K subarrays uses the received uplink broadband pilot signal to estimate the delay parameters of the center subarray channel. The other local processing units then extrapolate the delay parameters of the corresponding subarray channels.

[0010] Step 2: Each local processing unit transmits the estimated delay parameters of the corresponding subarray to the central processing unit. The central processing unit uses the symmetric characteristics of the broadband near-field channel parameters to linearly fuse the delay parameters, decouple and independently estimate the angular parameter information of the near-field channel;

[0011] Step 3: Based on the estimated angle and the delay parameter information of each subarray, the central processing unit uses the symmetric characteristics of the broadband near-field channel parameters to linearly fuse the delay parameters, decouple and independently estimate the distance parameter information of the near-field channel;

[0012] Step 4: Based on the estimated angle, distance, and delay parameter information of each subarray, the central processing unit estimates the range parameter information of the near-field channel and transmits all estimated near-field channel parameters, including angle, distance, and range parameter information, to each local processing unit. Each local processing unit independently estimates the path gain of the corresponding subarray, transmits the estimated path gain back to the central processing unit, and removes the detected path information from the received signal.

[0013] Step 5: Repeat steps 1 to 4 L times. The near-field channel parameters of the L paths and the path gain of each subarray are detected in sequence from the received signal of the local processing unit. The central processing unit reconstructs the full-dimensional broadband near-field channel.

[0014] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0015] 1. This invention explores the new degree of freedom provided by the near-field beam squint effect in large-array, large-bandwidth systems. By utilizing the parameter symmetry between the difference values of the path delay parameters observed by different antenna units, the delay parameters observed by different antenna units are linearly fused, and the near-field channel parameters are decoupled and independently estimated, thus overcoming the challenges posed by the near-field beam squint effect to CSI acquisition.

[0016] 2. The present invention utilizes a distributed estimation strategy to separate the full-dimensional signal processing process into multiple independent sub-array signal processing processes. The central processing unit (CPU) only needs to perform a simple linear combination of the delay parameters estimated by different local processing units (LPUs), greatly reducing the computational complexity of broadband near-field channel CSI acquisition.

[0017] 3. Although each subarray deploys only a single RF chain, the proposed method obtains delay parameters by observing multi-subcarrier signals and estimates near-field channel parameters by linearly combining the delay parameters of different subarrays. This eliminates the need to obtain the received signal in the antenna domain within the subarray. Therefore, users only need to send a single uplink pilot signal, effectively reducing the pilot overhead for full-dimensional near-field CSI acquisition in hybrid-architecture wideband XL-MIMO systems.

[0018] 4. The present invention estimates angle, distance and range parameters based on the parameter symmetry characteristics of the broadband near-field channel, avoiding the challenges of high computational complexity and low estimation accuracy faced by the codebook-based estimation algorithm. The estimation result is a continuous value off-grid, which can achieve improved broadband near-field channel reconstruction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of a near-field CSI distributed estimation method for broadband XL-MIMO according to the present invention;

[0020] Figure 2 This is a model diagram of a distributed hybrid architecture broadband XL-MIMO system according to an embodiment of the present invention;

[0021] Figure 3 This is a flow chart of the cooperation between the CPU and LPUs according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be interpreted as limiting the present invention.

[0023] like Figure 1 As shown, the present invention proposes a low-complexity near-field CSI distributed estimation method for broadband XL-MIMO. In a broadband XL-MIMO wireless transmission system, the base station divides a large array into multiple subarrays. Each subarray deploys an independent RF chain and processes the received signal through an independent landing unit (LPU). After the central LPU estimates the channel delay parameters of the central subarray, other LPUs quickly extrapolate the channel delay parameters of other subarrays. The CPU utilizes the symmetric characteristics of the broadband near-field channel parameters to perform linear data fusion on the delay parameters of different subarrays, decouple the near-field channel parameters, and independently estimate them. The LPUs calculate the path gain of each subarray separately, and then the CPU reconstructs the full-dimensional broadband near-field channel.

[0024] The array unit distribution forms of the base station include but are not limited to the following forms: uniform linear array, uniform planar array and uniform circular array.

[0025] The base station subarray division strategies include but are not limited to the following: evenly dividing the entire array into several subarrays of the same dimension, dividing the entire array into several subarrays of different dimensions, and treating each antenna unit as an independent subarray.

[0026] The hardware architecture of a single subarray of a base station includes but is not limited to the following: different antenna elements deploying independent RF chains, all antenna elements connected to a small number of RF chains, and all antenna elements connected to the same RF chain.

[0027] The near-field channel is composed of one or more propagation paths, the number of near-field propagation paths is L, and L satisfies that L is greater than or equal to 1.

[0028] The base station serves one or more users simultaneously. The number of users is T, and T is greater than or equal to 1. Different users send orthogonal broadband pilot signals.

[0029] The specific steps include:

[0030] (1) The user equipment (UE) sends an uplink broadband pilot signal. The array center LPU first estimates the path delay parameters of the center sub-array channel, and then uses other LPUs to quickly extrapolate the path delay parameters of other sub-array channels.

[0031] Broadband signal bands include but are not limited to the following: upper 6 GHz (U6G), millimeter wave bands, sub-terahertz bands, terahertz bands and higher frequency bands.

[0032] (2) The LPUs transmit the estimated path delay parameters of different subarrays to the CPU. The CPU uses the symmetric characteristics of the broadband near-field channel parameters to perform linear data fusion, decouple, and independently estimate the angular parameter information of the near-field channel.

[0033] (3) Based on the estimated angle and multi-subarray delay parameter information, the CPU uses the symmetric characteristics of broadband near-field channel parameters to perform deep linear fusion of data, decouple and independently estimate the distance parameter information of the near-field channel;

[0034] (4) Based on the estimated angle, distance, and multi-subarray delay parameter information, the CPU estimates the range parameter information of the near-field channel and transmits all the estimated near-field channel parameters back to the LPUs; each LPU independently estimates the subarray path gain using the near-field channel parameters fed back by the CPU, removes the detected path information from the received signal, and transmits the path gain information to the CPU;

[0035] (5) The above (1)-(4) are repeated L times, each time detecting the near-field channel parameters of one path, and the near-field channel parameters of all L paths are detected L times. Finally, the full-dimensional broadband near-field channel is reconstructed by the CPU.

[0036] Example

[0037] like Figure 2As shown in the hybrid architecture broadband XL-MIMO system, it is assumed that the BS is equipped with a uniform linear array (ULA) with an antenna unit spacing of half a wavelength, the number of antenna units is N, and the entire array is evenly divided into K sub-arrays, K is an even number, and the number of antenna units in each sub-array is N. s All antenna elements of each subarray are combined into an independent RF chain through analog phase shifters, and the received signal is processed by an independent LPU. K LPUs are connected to the same CPU, and information can be exchanged between the K LPUs. The system carrier frequency is f c , the number of subcarriers is M, and the frequency interval between different subcarriers is Δf. Consider the U6G frequency band. Assume that a single user equipment (UE) uses a single antenna configuration. To perform data detection and transmission, the base station needs to obtain full-dimensional broadband near-field CSI and reconstruct the full-dimensional broadband near-field channel H. This embodiment will reconstruct H, including the following steps:

[0038] Step 1: The user equipment sends an uplink broadband all-one pilot signal. Each XL-MIMO subarray receives the uplink pilot signal through a hybrid precoding architecture, and the LPU processes the received signal. The phase of each analog phase shifter is randomly generated in the interval [0, 2π]. The signal model received by the kth subarray is:

[0039]

[0040] in, is the broadband signal received by the kth subarray, is the uplink channel between the UE and the kth subarray, To simulate the phase shift matrix, each unit satisfies the constant modulus and random phase characteristics, that is, |[f k ] n |=1 / N s ,n=1,…,N s , P is the signal transmission power, The mean is 0 and the variance is σ 2 The uplink full-dimensional broadband near-field channel has the following expression:

[0041]

[0042] Where L is the number of propagation paths, g l is the gain of the lth path in the channel, θ l is the azimuth angle of the incident signal in the lth path, r l represents the range parameter from the user to the scatterer in the lth path, d l It represents the distance parameter from the scatterer to the BS center in the lth path. The user in the line-of-sight (LoS) path is regarded as a scatterer. is the near-field steering vector of the antenna domain of the lth path, is the multi-subcarrier steering vector of the lth path, is the near-field beam squint vector of the l-th path, l=1,…,L.

[0043] Due to the presence of near-field beam squint, full-dimensional broadband near-field channels do not have spatial sparsity. However, since the entire array is divided into K subarrays, each with a small aperture, the internal beam squint effect can be ignored, while the beam squint effect between different subarrays still needs to be considered. The kth subarray channel has the following expression:

[0044]

[0045] in, is the antenna domain near-field steering vector of the lth path on the kth subarray, is the near-field beam squint effect on the kth subarray, is an all-one vector, and The azimuth and distance of the lth path are observed for the kth subarray respectively. Furthermore, the received signal of the kth subarray can be expressed as:

[0046]

[0047] in, The multi-subcarrier steering vector for the k-th subarray observation path l, represents the equivalent delay parameter of the kth subarray observing the lth path.

[0048] For ease of understanding, assume that L = 1. When L > 1, an iterative estimation strategy can be used to sequentially detect the near-field channel parameters of L paths. Taking the estimation process of a single broadband near-field path as an example, the path index subscript l is omitted and the center subarray index is set to The central LPU uses the received uplink pilot signal to perform inverse Fourier transform and detects the delay parameter based on the delay domain codebook. Since the difference in delay parameters of adjacent sub-arrays is less than the delay resolution 1 / M, The sub-array delay parameters can be extrapolated as in The value selection rule is to make the The correlation between the received signal of the subarray and the multi-subcarrier steering vector is the strongest. Similarly, The sub-array delay parameters can be extrapolated as Based on this, extrapolate the delay parameters of all sub-arrays in turn and save them as Where k = 1,…,K.

[0049] Step 2: If Figure 3 As shown in Figure 1, K LPUs transmit the estimated delay parameters of different subarrays to the CPU, which performs linear data fusion to decouple and estimate the near-field channel parameters, including angle, distance, and range parameters. In particular, the delay parameter of the kth subarray can be equivalently expressed as:

[0050]

[0051] in, c is the speed of light, s is the distance between adjacent antennas. k It has odd and even symmetry properties related to the antenna index k, that is, δ k =-δ K-k+1 ,as well as So we have:

[0052] η(K-k+1)-η(k)=2δ k N s sθ

[0053] By combining the delay parameters of the kth and K-k+1th subarrays, the angle parameters can be decoupled and estimated independently, k = 1, ..., K / 2. Therefore, by combining the delay parameters detected by K / 2 groups of different subarrays, the angle parameters can be estimated as:

[0054]

[0055] Among them, it means Pseudo-inverse operation.

[0056] Step 3: The CPU further performs deep linear fusion on the delay parameters of different sub-arrays, where the kth and The delay parameters of the sub-arrays satisfy:

[0057]

[0058] Therefore, by combining the kth and The delay parameters and distance parameters of the sub-arrays can be decoupled and estimated independently, k = 1, ..., K / 2. Furthermore, the above formula can be expressed as:

[0059]

[0060] By combining the time delay parameters detected by K / 2 groups of different sub-arrays and the angle parameters estimated in step 2, the distance parameter can be estimated as:

[0061]

[0062] Step 4: When the angle and distance parameters are accurately estimated, the delay parameters of the kth and K-k+1th subarrays can be recombined as:

[0063]

[0064] By combining the delay parameters of the kth and K-k+1th subarrays, k=1,…,K / 2, the range parameter can be linearly estimated as:

[0065]

[0066] The CPU will detect the near-field channel parameters Feedback is sent to the LPUs, and each LPU estimates the path gain of the subarray separately. The path gain of the k-th subarray can be calculated as:

[0067]

[0068] Among them, the coefficient matrix It can be expressed as:

[0069]

[0070] Where k = 1,…,K, The distance parameter of the incident signal observed by the kth subarray can be obtained by Based on this, the K LPUs remove the detected path information from the received signal and pass the estimated path gain back to the CPU. The received signal of the kth LPU is updated as:

[0071]

[0072] Step 5: The above steps are iterated L times, and the near-field channel parameters of L paths and the path gains of different sub-arrays are detected from the received signal in turn, and saved as Where l = 1, ..., L, k = 1, ..., K. The CPU uses the estimated angles, distances, ranges, and gains of the L paths to reconstruct the full-dimensional broadband near-field channel between the user and the BS as:

[0073]

[0074] in, is the reconstructed full-dimensional broadband near-field channel between the user and the BS.

[0075] Based on the same inventive concept, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the aforementioned method for distributed near-field CSI estimation for wideband XL-MIMO are implemented.

[0076] Based on the same inventive concept, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the steps of the aforementioned near-field CSI distributed estimation method for wideband XL-MIMO.

[0077] It will be understood by those skilled in the art 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. 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-ROM, optical storage, etc.) containing computer-usable program code.

[0078] 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.

[0079] 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.

[0080] 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.

[0081] The above embodiments are only for illustrating the technical idea of the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the present invention.

Claims

1. Near-field CSI distributed estimation method for broadband XL-MIMO. In a broadband XL-MIMO wireless transmission system, the base station is equipped with a uniformly distributed antenna array. The entire antenna array is divided into K sub-arrays, where K is an even number. All antenna elements in each sub-array are connected to G independent RF chains through analog phase shifters, where 1≤G≤N. s , N s is the number of antenna elements in each subarray, G radio frequency chains process received signals through a local processing unit, K local processing units are connected to the same central processing unit, and information is exchanged between the K local processing units; it is characterized in that, The method comprises the following steps: Step 1: The user sends an uplink broadband pilot signal in the U6G or higher frequency band. The local processing unit corresponding to the center subarray of the K subarrays uses the received uplink broadband pilot signal to estimate the delay parameters of the center subarray channel. The other local processing units then extrapolate the delay parameters of the corresponding subarray channels. Step 2: Each local processing unit transmits the estimated delay parameters of the corresponding subarray to the central processing unit. The central processing unit uses the symmetric characteristics of the broadband near-field channel parameters to linearly fuse the delay parameters, decouple and independently estimate the angular parameter information of the near-field channel; Step 3: Based on the estimated angle and the delay parameter information of each subarray, the central processing unit uses the symmetric characteristics of the broadband near-field channel parameters to linearly fuse the delay parameters, decouple and independently estimate the distance parameter information of the near-field channel; Step 4: Based on the estimated angle, distance, and delay parameter information of each subarray, the central processing unit estimates the range parameter information of the near-field channel and transmits all estimated near-field channel parameters, including angle, distance, and range parameter information, to each local processing unit. Each local processing unit independently estimates the path gain of the corresponding subarray, transmits the estimated path gain back to the central processing unit, and removes the detected path information from the received signal. Step 5: Repeat steps 1 to 4 L times. The near-field channel parameters of the L paths and the path gain of each subarray are detected in sequence from the received signal of the local processing unit. The central processing unit reconstructs the full-dimensional broadband near-field channel.

2. The near-field CSI distributed estimation method for broadband XL-MIMO according to claim 1, characterized in that In step 1, the kth subarray channel expression is as follows: Among them, H k is the broadband near-field steering vector from the user to the kth subarray, L is the number of propagation paths, g l is the gain of the lth path in the channel, θ l is the azimuth angle of the incident signal in the lth path, r l is the range parameter from the user to the scatterer in the lth path, d l is the distance parameter from the scatterer to the center of the base station in the lth path, w k (θ l ,d l ) is the antenna domain near-field steering vector of the lth path on the kth subarray, p(d l ,r l ) is the multi-subcarrier steering vector of the lth path, is the near-field beam squint effect on the kth subarray, and The azimuth and distance of the lth path observed by the kth subarray are respectively, k = 1,…,K; The received signal of the kth subarray is expressed as: Among them, y k is the broadband signal received by the kth subarray, P is the signal transmission power, f k is the analog phase shift matrix, b(τ k,l ) is the multi-subcarrier steering vector of the k-th subarray observation path l, represents the equivalent delay parameter of the kth subarray observing the lth path, Δf is the frequency interval between adjacent subcarriers, c is the speed of light, and n k The mean is 0 and the variance is σ 2 Additive Gaussian white noise; Let the center subarray index be The local processing unit corresponding to the central sub-array performs inverse Fourier transform on the received uplink broadband pilot signal based on the delay domain codebook to calculate the delay parameter No. The delay parameters of the sub-arrays are extrapolated as follows: Among them, 1 / M is the delay resolution, The value rule is to make the The sub-array receives the signal and The subarray observes the multi-subcarrier steering vector of the lth path The correlation between them is the strongest, and the delay parameters of all sub-arrays are extrapolated in turn.

3. The near-field CSI distributed estimation method for wideband XL-MIMO according to claim 2, characterized in that: In step 2, the delay parameter τ of the kth subarray is k The equivalent expression is: Where r is the range parameter from the user to the scatterer, d is the distance parameter from the scatterer to the center of the base station, s is the spacing between adjacent antennas, and θ is the azimuth angle of the incident signal observed by the entire XL-MIMO array; δ k It has odd and even symmetry properties related to the subarray index k, that is, δ k =-δ K-k+1 ,as well as Therefore: η(K-k+1)-η(k)=2δ k N s sth By combining the delay parameters of the kth and K-k+1th subarrays, k=1,…,K / 2, the angle parameters are decoupled and estimated independently as: in, is the estimated azimuth angle of the incident signal observed by the entire XL-MIMO array, are the delay parameters of the 1st and K-1+1th subarrays respectively, Respectively The delay parameters of the sub-arrays, Represents a pseudo-inverse operation.

4. The near-field CSI distributed estimation method for wideband XL-MIMO according to claim 3, characterized in that In step 3, the kth and The delay parameters of the sub-arrays satisfy: Then combine the kth and The delay parameters of the subarrays, k = 1, ..., K / 2, and the distance parameters are decoupled and estimated independently as: in, is the estimated range parameter of the incident signal observed by the entire XL-MIMO array, Respectively Delay parameters of K sub-arrays.

5. The near-field CSI distributed estimation method for wideband XL-MIMO according to claim 4, characterized in that: In step 4, the delay parameters of the kth and K-k+1th subarrays are recombined as follows: By combining the delay parameters of the kth and K-k+1th subarrays, k=1,…,K / 2, the range parameter is linearly estimated as: in, is the estimated range parameter of the incident signal observed by the entire XL-MIMO array; The central processing unit transmits the estimated angle, distance and range parameter information back to each local processing unit, and each local processing unit estimates the path gain of the corresponding subarray separately. The path gain of the kth subarray is as follows: Coefficient matrix Expressed as: in, is the antenna near-field steering vector of the kth subarray, is the delay domain steering vector of the kth subarray, The distance parameter of the incident signal observed by the kth subarray; The received signal of the kth local processing unit is updated as: Among them, y′ k is the updated received signal of the kth local processing unit.

6. The near-field CSI distributed estimation method for wideband XL-MIMO according to claim 5, characterized in that In step 5, when all L paths are detected, the full-dimensional broadband near-field channel between the user and the base station is reconstructed as: in, To reconstruct the full-dimensional broadband near-field channel between the user and the base station, are the gains of the lth path estimated for the 1st and Kth subarrays respectively, is the estimated full-dimensional antenna domain steering vector of the lth path, is the estimated delay domain steering vector of the lth path, is the estimated near-field beam squint effect matrix of the lth path.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: When the processor executes the computer program, the steps of the near-field CSI distributed estimation method for wideband XL-MIMO are implemented.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the near-field CSI distributed estimation method for wideband XL-MIMO are implemented.

Citation Information

Patent Citations

  • Sparse channel estimation method for super-large-scale MIMO system

    CN116032699A

  • Knowledge-driven deep learning-based TSN high-dimensional channel rapid estimation method

    CN118300927A

  • XL-MIMO near-field CSI estimation method based on spatial geometric distributed processing

    CN119363521A

  • Deep generative models for downlink channel estimation in FDD massive MIMO systems

    WO2022172198A1

Cited By

  • Distributed XL-MIMO near-field channel estimation method and system based on cooperation mechanism

    CN122137705A