A scene scale unknown super large scale MIMO mixed field channel estimation method based on an OMP algorithm

Abstract

The application relates to a scene-proportion-unknown super-large-scale MIMO mixed field channel estimation method based on an OMP algorithm, which comprises the following steps: constructing a signal model from a sending end to a receiving end; calculating Rayleigh distance according to a sending end antenna array and a receiving end antenna array; constructing a far-field channel model and a near-field channel model between the sending end and the receiving end according to the calculated Rayleigh distance; constructing a mixed field channel model between the sending end and the receiving end according to the constructed far-field channel model and the near-field channel model; and solving the mixed field channel model by using a scene-proportion-unknown OMP algorithm to estimate the mixed field channel between the sending end and the receiving end. The application can effectively estimate a signal in a scene where the number proportion of far-field and near-field paths is unknown.

Description

Technical Field

[0001] This invention belongs to the field of communication technology, and in particular relates to a method for estimating the channel of a mixed field in ultra-large-scale MIMO with unknown scene proportions based on the OMP algorithm. Background Technology

[0002] With the continuous growth of wireless communication demands, next-generation communication systems such as 5G and 6G place higher requirements on spectral efficiency, system capacity, and quality of service. XL-MIMO technology, by deploying a large number of antenna elements in the communication system, can significantly improve system capacity and coverage, and is therefore considered one of the core technologies of future communication. However, the far-field and near-field channel components in XL-MIMO systems bring complex propagation characteristics, making channel modeling and estimation challenging. Traditional channel estimation methods are mostly based on uniform linear array models and usually assume that the ratio of far-field and near-field path components in the communication environment is known. However, the propagation environment in real-world scenarios is complex and variable. Factors such as building obstruction, obstacle reflection, user density, and propagation distance all affect the ratio of far-field and near-field components, making this assumption difficult to meet in practical applications.

[0003] Furthermore, existing XL-MIMO channel estimation methods are mostly designed for single-antenna users, and this single channel model is difficult to capture the complex propagation characteristics of multi-antenna users in real-world scenarios. Some current research attempts to improve channel estimation in scenarios where the proportions are known by introducing techniques such as sparse representation and compressed sensing. However, when the proportions of far-field and near-field paths are unknown, the performance of these methods is often unstable and difficult to adapt to the channel estimation requirements of mixed fields with unknown proportions in real-world communication environments.

[0004] Orthogonal matching pursuit (OMP) is a classic sparse signal reconstruction method with efficient iterative search capabilities, gradually selecting the sparse component that best matches the observed data to achieve high-precision channel estimation. However, in the problem of channel estimation in mixed fields with unknown proportions, traditional OMP and other sparse reconstruction algorithms have not been fully optimized and applied. XL-MIMO systems based on the Saleh-Valenzuela model for channel modeling possess high flexibility and can accurately describe multipath propagation characteristics; therefore, combining it with the OMP algorithm has application potential in scenarios with unknown proportions.

[0005] Therefore, it is necessary to propose a novel method to achieve mixed-field channel estimation in XL-MIMO systems when the ratio of far-field to near-field communication is unknown, through adaptive channel modeling and efficient sparse channel estimation. This method should not only possess strong channel reconstruction capabilities but also exhibit good robustness and adaptability in complex and ever-changing communication scenarios to meet the channel estimation requirements of future high-density and complex communication environments. Summary of the Invention

[0006] To address the problems existing in the background art and achieve the objectives mentioned therein, this invention provides a method for estimating the channel in a mixed-field environment of ultra-large-scale MIMO with unknown scene proportions based on the OMP algorithm, comprising:

[0007] S1: Construct a signal model from the transmitter to the receiver;

[0008] S2: Calculate the Rayleigh distance based on the transmitting antenna array and the receiving antenna array;

[0009] S3: Construct far-field and near-field channel models between the transmitter and receiver based on the calculated Rayleigh distance;

[0010] S4: Construct a hybrid field channel model between the transmitter and receiver based on the constructed far-field channel model and near-field channel model;

[0011] S5: Solve the mixed field channel model using the OMP algorithm with unknown scene proportions to estimate the mixed field channel between the transmitter and receiver.

[0012] Preferably, the signal model from the transmitter to the receiver includes:

[0013]

[0014]

[0015]

[0016]

[0017]

[0018]

[0019]

[0020] in, This indicates the pilot signal received by the receiver. Represents a merged matrix. This represents the channel matrix between the transmitter and receiver; N represents the noise at the receiver. This indicates the pilot signal transmitted by the transmitter; M represents the number of time slots. Indicates the first Pilot signals received by the receiver in each time slot; Indicates the number of RF chains in the receiving antenna array at the receiving end; Indicates the first Pilot signals transmitted by each time slot transmitter; Represents a hybrid precoding matrix; express Received noise, For the first Noise power after time slot combining Indicates size is unit array; This indicates that the noise follows a complex Gaussian distribution. , Indicates size is unit array, It represents the Kronecker product.

[0021] Preferably, the Rayleigh distance includes:

[0022]

[0023] in, Indicates Rayleigh distance, and These represent the apertures of the transmitting antenna array and the receiving antenna array, respectively. Indicates wavelength.

[0024] Preferably, the far-field channel model between the transmitter and receiver includes: when the distance between the scatterer and the transmitting antenna array is greater than the Rayleigh distance or the distance between the scatterer and the receiving antenna array is greater than the Rayleigh distance, the channel should be modeled as a far-field channel, and modeled based on plane waves. The far-field channel model is constructed as follows:

[0025]

[0026]

[0027] in, For the complex gain of the far-field path, The number of far-field paths, and These are the steering vectors for the transmitting antenna array and the receiving antenna array, respectively. This indicates the conjugate transpose, and j represents the imaginary unit. and These represent the number of transmitting antennas and the number of receiving antennas, respectively. Represents the natural base. Represents pi (π). and These represent the antennas of the transmitting antenna array and the receiving antenna array at the [missing information - likely a specific location or point]. The angle at the path, and , and These represent the actual physical angles of the transmitting and receiving antenna arrays, respectively, and their values ​​range from [value range missing]. , , ;

[0028] The above non-sparse channel Using sparse corner domain channels Represented as:

[0029]

[0030]

[0031] in, It is a sparse far-field angular domain channel; and These are the Fourier transform matrices for the transmitting and receiving ends, respectively; , and These represent the sampling angles of the far-field steering vectors of the transmitting and receiving antenna arrays, respectively.

[0032] Preferably, the near-field channel model between the transmitter and receiver includes: when the distance between the scatterer and the transmitting antenna array is less than the Rayleigh distance or the distance between the scatterer and the receiving antenna array is less than the Rayleigh distance, the channel should be modeled as a near-field channel, and modeled based on spherical waves. The near-field channel model is constructed as follows:

[0033]

[0034]

[0035]

[0036]

[0037]

[0038]

[0039]

[0040] in, For the complex gain of the near-field path, The number of near-field paths, and These are the steering vectors for the transmitting antenna array and the receiving antenna array, respectively; Indicates conjugate transpose; and They represent the first The distance from each scatterer to the center of the transmitting and receiving antenna arrays Indicates the first The scatterer to the transmitting antenna array The distance between the antennas; Indicates the first The scatterer to the receiving antenna array The distance between the antennas; This indicates the spacing between antennas in the transmitting antenna array and the receiving antenna array; Indicates the transmitting antenna array number The distance between each antenna and the central array antenna; Indicates the receiving antenna array number The distance between each antenna and the central array antenna;

[0041] The above non-sparse channel Using sparse polar-domain channels Represented as:

[0042]

[0043]

[0044] in, For sparse polar region channels ; , ;matrix Each column element represents the angle. and distance Upsampled near-field array steering vector; Each column element represents the angle. and distance Upsampled near-field array steering vector, Indicates the sampling angle The number of sampling distances at each location, Indicates the sampling angle The number of sampling distances at each location.

[0045] Preferably, the hybrid field channel model includes:

[0046]

[0047] in, This represents a hybrid field channel model.

[0048] Preferably, solving the hybrid channel model includes: performing channel estimation for the far-field channel model and the near-field channel model respectively. When estimating the far-field channel model, it is assumed that the proportion of the far-field channel model is 1. When estimating the near-field channel model, the proportion is traversed with a certain step size until the residual between the received signal and the predicted signal of the current estimated channel model reaches a set threshold. Finally, the channel parameters are updated to obtain the estimation result of the hybrid channel model.

[0049] The present invention has at least the following beneficial effects

[0050] This invention addresses scenarios where both the transceiver and receiver are equipped with antenna arrays and the ratio of far-field to near-field path numbers is unknown. It constructs far-field and near-field channel models separately, thereby effectively building a hybrid-field channel model. The OMP algorithm is used to perform channel estimation in three stages. The proposed OMP hybrid-field channel estimation algorithm consistently demonstrates superior estimation performance under all ratio values, achieving an estimation performance improvement of approximately 5.8 dB compared to the most traditional OMP algorithm and approximately 1.8 dB compared to the traditional OMP-1 algorithm. Attached Figure Description

[0051] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0052] Figure 2 This is a schematic diagram of the far-field and near-field distribution of the present invention;

[0053] Figure 3 This is a schematic diagram of the hybrid field channel model of the present invention;

[0054] Figure 4 This is a schematic diagram showing the experimental simulation comparison of the present invention. Detailed Implementation

[0055] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0056] Please see Figures 1-4 This invention provides a method for channel estimation in ultra-large-scale MIMO mixed-field scenarios with unknown scene proportions based on the OMP algorithm, comprising:

[0057] S1: Construct a signal model from the transmitter to the receiver;

[0058] Assume that the transmitting end and the receiving end are each equipped with root antenna and Since antenna arrays are typically implemented in a hybrid digital-analog configuration with a certain number of radio frequency (RF) chains, it is assumed that the number of RF chains in the transmitting and receiving antenna arrays is [number missing]. and ,set up If we represent the channel from the transmitter to the receiver, then the received signal at the receiver can be represented as:

[0059] (1)

[0060] in, They represent the first time. The pilot signals received in each time slot, the combining matrix, the mixing precoding matrix, and the transmitted pilot signals, express Received noise, For the first Noise power after combining time slots.

[0061] set up , The The element is the first one. The first time slot of the transmitting antenna array The transmitted signal from the antenna is then in The pilot signal received in each time slot can be represented as:

[0062] (2)

[0063] in, , , Consider the pilot signals transmitted in M ​​time slots, the received pilot signals, and the noise at the receiver. , Indicates size is The identity matrix, Size is The identity matrix, Let represent the Kronecker product, and let represent that the noise follows a complex Gaussian distribution. In the channel estimation problem, H needs to be estimated given P, W, and Y. In an XL-MIMO system, the number of antennas in the transmit antenna array... The value of is usually very large. Therefore, in order to reduce the pilot overhead in practical communication systems, a channel estimation scheme with low overhead should be used. ).

[0064] S2: Calculate the Rayleigh distance based on the transmitting antenna array and the receiving antenna array;

[0065] The electromagnetic radiation field in a wireless communication system can be divided into the far field and the near field; different fields lead to different channel models. The boundary between these two fields is the Rayleigh distance. Confirmed, among which These are the apertures of the transmitting antenna array and the receiving antenna array, respectively. For wavelength. The models for the far field and near field are as follows: Figure 2 As shown.

[0066] S3: Construct far-field and near-field channel models between the transmitter and receiver based on the calculated Rayleigh distance;

[0067] Constructing a far-field channel model:

[0068] When the distance between the scatterer and the transmitting and receiving antenna arrays is greater than the Ruili distance, the channel is located in the far-field region. The channel model should be based on plane wave modeling. The far-field channel model can be constructed as follows:

[0069] (3)

[0070] in, For the complex gain of the far-field path, The number of far-field paths, and These are the steering vectors for the transmitting antenna array and the receiving antenna array, respectively. This indicates the conjugate transpose, and j represents the imaginary unit. Represents the natural base. Represents pi (π). and These represent the antennas of the transmitting antenna array and the receiving antenna array at the [missing information - likely a specific location or point]. The angle at each path is represented as:

[0071] (4)

[0072] in, For the complex gain of the far-field path, The number of far-field paths, and These are the steering vectors for the transmitting antenna array and the receiving antenna array, respectively. This indicates the conjugate transpose, and j represents the imaginary unit. and These represent the number of transmitting antennas and the number of receiving antennas, respectively. Represents the natural base. Represents pi (π). and These represent the antennas of the transmitting antenna array and the receiving antenna array at the [missing information - likely a specific location or point]. The angle at the path, and , and These represent the actual physical angles of the transmitting and receiving antenna arrays, respectively, and their values ​​range from [value range missing]. , , ;

[0073] Due to the existence of finite scattering in the communication environment, the angular domain channel The far field is typically sparse. To reduce pilot overhead used for channel estimation, the aforementioned non-sparse channels... Sparse angular domain channels can be used Represented as;

[0074] (5)

[0075] (6)

[0076] in, It is a sparse far-field angular domain channel; and These are the Fourier transform matrices for the transmitting and receiving ends, respectively; , and These represent the sampling angles of the far-field steering vectors of the transmitting and receiving antenna arrays, respectively.

[0077] Constructing a near-field channel model:

[0078] When the distance between the scatterer and the transmitting and receiving antenna arrays is less than the Rayleigh distance, the channel is located in the near-field region. The channel model should be based on spherical waves, and the near-field channel model can be constructed as follows:

[0079] (7)

[0080] in, For the complex gain of the near-field path, The number of near-field paths, and These are the steering vectors for the transmitting antenna array and the receiving antenna array, respectively; and They represent the first The distances from each scatterer to the centers of the transmitting and receiving antenna arrays are expressed as follows:

[0081] (8)

[0082] in, Indicates the first The scatterer to the transmitting antenna array The distance between the antennas; Indicates the first The scatterer to the receiving antenna array The distance between the antennas; Indicates the transmitting antenna array number The distance between each antenna and the central array antenna; Indicates the receiving antenna array number The distance between each antenna and the central array antenna can be expressed as:

[0083] (9)

[0084] in, This represents the spacing between antennas in the transmitting and receiving antenna arrays; its expression is:

[0085] (10)

[0086] Unlike the far-field, the steering vector in the near-field has an additional distance parameter. Therefore, the Fourier transform (DFT) matrix cannot sparsify the original non-sparse channel matrix. Thus, the near-field polar transform matrix is ​​used to address this issue. The polar transform matrix can be expressed as:

[0087] (11)

[0088] Based on polar domain transformation matrix and The above non-sparse channels Sparse polar-domain channels can be used Represented as:

[0089] (12)

[0090] in, For sparse polar region channels ; , ;matrix Each column element represents the angle. and distance Upsampled near-field array steering vector; Each column element represents the angle. and distance Upsampled near-field array steering vector, Indicates the sampling angle The number of sampling distances at each location, Indicates the sampling angle The number of sampling distances at each location.

[0091] S4: Construct a hybrid field channel model between the transmitter and receiver based on the constructed far-field channel model and near-field channel model;

[0092] Building upon the far-field and near-field channel models, a hybrid-field channel model is constructed to consider that scatterers in real-world XL-MIMO scenarios may be located in either the near or far field. This model aims to more comprehensively describe the impact of different scatterers on channel characteristics during signal propagation. Hybrid-field XL-MIMO channel It can be represented as:

[0093] (13)

[0094] in, This represents a hybrid field channel model.

[0095] S5: Solve the mixed field channel model using the OMP algorithm with unknown scene proportions to estimate the mixed field channel between the transmitter and receiver.

[0096] Based on equations (5) and (12), equation (2) can be expressed as:

[0097] (14)

[0098] in, and These represent the receiving and transmitting sensing matrices in the far field, respectively. and Let represent the near-field receiving and transmitting sensing matrices, respectively. From the derivation of equations (5) and (12), we can obtain... and They are all sparse, which allows for the estimation of the entire mixed-field channel.

[0099] This section will elaborate on the specific implementation process of the proposed method. The mixed-field channel estimation algorithm for scenarios where the ratio of far-field and near-field components is unknown can be summarized as follows:

[0100] Input: Receive signal Merging matrices Pilot signal matrix Far-field transmitter transformation matrix Transformation matrix of far-field receiver Near-field transmitter transformation matrix Near-field receiver transformation matrix Total number of paths Near-field path ratio .

[0101] Output: Estimation results .

[0102] The specific process for estimating the mixed field channel between the transmitter and receiver is as follows:

[0103] initialization: , , residual matrix Actual number of iterations , .

[0104] Phase 1: Estimation of far-field path components in the angular domain.

[0105] Step 1: Command ;

[0106] Step 2: For each far-field path, calculate the correlation between the far-field sensing matrix and the residual matrix, and then... Find the column index that is most relevant to the far-field sensing matrix and the residual matrix as the far-field support index. This represents the conjugate transpose of the calculated far-field receiving sensing matrix. This represents the conjugate transpose of the far-field transmitting sensing matrix. Represent the L2 norm, and through Calculate the indices of the far-field transmitter and receiver. This indicates rounding down. This represents the modulo function, and updates the total far-field support set and the far-field component support sets. .

[0107] Step 3: Solve using the least squares method , Indicates in Far-field channel estimation matrix: First, calculate the far-field sparse matrix. , Indicates that the far-field transmission sensing matrix takes the first... OK, Indicates that the far-field receiving sensing matrix takes the first... The column is then calculated using the least squares method. , This represents the channel estimation vector corresponding to the far-field channel estimation matrix, which is finally obtained by vector-matrix conversion. .

[0108] Step 4: Update the far-field residual matrix: .

[0109] Step 5: Number of Update Iterations : ,if If the result is positive, repeat step 2 to continue iterating; otherwise, the iteration ends, and the latest iteration result is obtained. .

[0110] Phase 2: Estimation of near-field path components in polar coordinates.

[0111] Step 6: Order Minimum value of residual vector , This represents the residual vector corresponding to the residual matrix.

[0112] Step 7: The step size will Adjust from 1 to 0 to iterate through all possibilities: via , This indicates the extraction of the first to the second far-field support set. One element, This indicates taking the th element in the residual matrix. The column will correspond to Far-field support set and residual vector Stored separately and middle.

[0113] Step 8: For each near-field path, calculate the correlation between the near-field sensing matrix and the residual matrix, and then... Find the perception matrix and residual matrix The most relevant column index is used as the near-field supporting index, and through Calculate the indices of the near-field transmitter and receiver, and update the total near-field support set and the near-field component support set. .

[0114] Step 9: Least squares method , Indicates in Near-field estimation matrix at: First, calculate the near-field sparse matrix. , This indicates that the near-field transmission sensing matrix takes the first position. OK, This indicates that the near-field receiving sensing matrix takes the first position. The column is then calculated using the least squares method. , This represents the channel estimation vector corresponding to the near-field channel estimation matrix, which is then transformed into a vector matrix to obtain... .

[0115] Step 10: Update the near-field residual matrix and residual vector: , .

[0116] Step 11: Number of Update Iterations : ,if If the result is positive, repeat step 8 to continue iterating; otherwise, the iteration ends, and the latest iteration result is obtained. .

[0117] Step 12: If the value of the residual vector is less than or equal to the minimum value of the residual vector, that is... Then update the minimum residual vector. .

[0118] Phase 3: Output the final estimation results.

[0119] Step 13: Initialize the hybrid field channel matrix: .

[0120] Step 14: If the far-field support set If near-field support set .

[0121] Experimental simulation:

[0122] The simulation parameters are set as follows:

[0123] Carrier frequency: 30GHz

[0124] Carrier wavelength: 0.01m

[0125] Number of transmitting antennas: 2

[0126] Number of receiver antennas: 256

[0127] Total number of paths: 10

[0128] Pilot length: 1

[0129] SNR: -6dB

[0130] The normalized mean square error (NMSE) between the channel estimation matrix and the true channel matrix is ​​used as a performance metric, defined as follows:

[0131]

[0132] in, Represents the actual channel matrix, This represents the estimated channel matrix. It represents the mathematical expectation.

[0133] The proposed ultra-large-scale MIMO hybrid field channel estimation algorithm is compared with the hybrid field channel estimation algorithms of traditional OMP algorithms, as well as LS and MMSE. The comparison results are as follows: Figure 4 As shown:

[0134] from Figure 4As can be seen, the proposed OMP mixed-field channel estimation algorithm consistently achieves the lowest NMSE value across all γ values. Specifically, compared to the most traditional OMP algorithm, it improves performance by approximately 5.8 dB, and compared to the traditional OMP-1 algorithm, it improves performance by approximately 1.8 dB. Particularly noteworthy is the proposed algorithm's NMSE value reaching its lowest when γ=0, i.e., in the near-field scenario, demonstrating the best estimation performance. When γ=1, i.e., in the far-field scenario, the proposed algorithm achieves the second-lowest NMSE value.

[0135] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for channel estimation in a mixed-field ultra-large-scale MIMO scenario with unknown scene proportions based on the OMP algorithm, characterized in that, include: S1: Construct a signal model from the transmitter to the receiver; S2: Calculate the Rayleigh distance based on the transmitting antenna array and the receiving antenna array; The Rayleigh distance includes: in, Indicates Rayleigh distance, and These represent the apertures of the transmitting antenna array and the receiving antenna array, respectively. Indicates wavelength; S3: Construct far-field and near-field channel models between the transmitter and receiver based on the calculated Rayleigh distance; The far-field channel model between the transmitter and receiver includes: when the distance between the scatterer and the transmitting antenna array is greater than the Rayleigh distance or the distance between the scatterer and the receiving antenna array is greater than the Rayleigh distance, the channel should be modeled as a far-field channel, and modeled based on plane waves. The far-field channel model is constructed as follows: in, For the complex gain of the far-field path, The number of far-field paths, and These are the steering vectors for the transmitting antenna array and the receiving antenna array, respectively. This indicates the conjugate transpose, and j represents the imaginary unit. and These represent the number of transmitting antennas and the number of receiving antennas, respectively. Represents the natural base. Represents pi (π). and These represent the antennas of the transmitting antenna array and the receiving antenna array at the [missing information - likely a specific location or point]. The angle at the path, and , and These represent the actual physical angles of the transmitting and receiving antenna arrays, respectively, and their values ​​range from [value range missing]. , , ; The above non-sparse channel Using sparse corner domain channels Represented as: in, It is a sparse far-field angular domain channel; and These are the Fourier transform matrices for the transmitting and receiving ends, respectively; , and These represent the sampling angles of the far-field steering vectors of the transmitting antenna array and the receiving antenna array, respectively. The near-field channel model between the transmitter and receiver includes the following: when the distance between the scatterer and the transmitting antenna array is less than the Rayleigh distance or the distance between the scatterer and the receiving antenna array is less than the Rayleigh distance, the channel should be modeled as a near-field channel, and modeled based on spherical waves. The near-field channel model is constructed as follows: in, For the complex gain of the near-field path, The number of near-field paths, and These are the steering vectors for the transmitting antenna array and the receiving antenna array, respectively; Indicates conjugate transpose; and They represent the first The distance from each scatterer to the center of the transmitting and receiving antenna arrays Indicates the first The scatterer to the transmitting antenna array The distance between the antennas; Indicates the first The scatterer to the receiving antenna array The distance between the antennas; This indicates the spacing between antennas in the transmitting antenna array and the receiving antenna array; Indicates the transmitting antenna array number The distance between each antenna and the central array antenna; Indicates the receiving antenna array number The distance between each antenna and the central array antenna; The above non-sparse channel Using sparse polar-domain channels Represented as: in, For sparse polar region channels ; , ;matrix Each column element represents the angle. and distance Upsampled near-field array steering vector; Each column element represents the angle. and distance Upsampled near-field array steering vector, Indicates the sampling angle The number of sampling distances at each location, Indicates the sampling angle The number of sampling distances at each location; S4: Construct a hybrid field channel model between the transmitter and receiver based on the constructed far-field channel model and near-field channel model; S5: Solve the mixed field channel model using the OMP algorithm with unknown scene proportions to estimate the mixed field channel between the transmitter and receiver; The solution to the hybrid channel model includes: performing channel estimation for the far-field channel model and the near-field channel model respectively. When estimating the far-field channel model, it is assumed that the proportion of the far-field channel model is 1. When estimating the near-field channel model, the proportion is traversed with a certain step size until the residual between the received signal and the predicted signal of the current estimated channel model reaches a set threshold. Finally, the channel parameters are updated to obtain the estimation result of the hybrid channel model. OMP algorithms for scenes with unknown proportions include: Input: Receive signal Merging matrices Pilot signal matrix Far-field transmitter transformation matrix Transformation matrix of far-field receiver Near-field transmitter transformation matrix Near-field receiver transformation matrix Total number of paths Near-field path ratio ; Output: Estimation results ; The specific process for estimating the mixed field channel between the transmitter and receiver is as follows: initialization: , , residual matrix Actual number of iterations , ; Phase 1: Estimation of far-field path components in the angular domain; Step 1: Command ; Step 2: For each far-field path, calculate the correlation between the far-field sensing matrix and the residual matrix, and then... Find the column index that is most relevant to the far-field sensing matrix and the residual matrix as the far-field support index. This represents the conjugate transpose of the calculated far-field receiving sensing matrix. This represents the conjugate transpose of the far-field transmitting sensing matrix. Represent the L2 norm, and through Calculate the indices of the far-field transmitter and receiver. This indicates rounding down. This represents the modulo function, and updates the total far-field support set and the far-field component support sets. ; Step 3: Solve using the least squares method , Indicates in Far-field channel estimation matrix: First, calculate the far-field sparse matrix. , Indicates that the far-field transmission sensing matrix takes the first... OK, Indicates that the far-field receiving sensing matrix takes the first... The column is then calculated using the least squares method. , This represents the channel estimation vector corresponding to the far-field channel estimation matrix, which is finally obtained by vector-matrix conversion. ; Step 4: Update the far-field residual matrix: ; Step 5: Number of Update Iterations : ,if If the result is positive, repeat step 2 to continue iterating; otherwise, the iteration ends, and the latest iteration result is obtained. ; Phase Two: Estimation of near-field path components in polar coordinates; Step 6: Order Minimum value of residual vector , This represents the residual vector corresponding to the residual matrix; Step 7: The step size will Adjust from 1 to 0 to iterate through all possibilities: via , This indicates the extraction of the first to the second far-field support set. One element, This indicates taking the th element in the residual matrix. The column will correspond to Far-field support set and residual vector Stored separately and middle; Step 8: For each near-field path, calculate the correlation between the near-field sensing matrix and the residual matrix, and then... Find the perception matrix and residual matrix The most relevant column index is used as the near-field supporting index, and through Calculate the indices of the near-field transmitter and receiver, and update the total near-field support set and the near-field component support set. ; Step 9: Least squares method , Indicates in Near-field estimation matrix at: First, calculate the near-field sparse matrix. , This indicates that the near-field transmission sensing matrix takes the first position. OK, This indicates that the near-field receiving sensing matrix takes the first position. The column is then calculated using the least squares method. , This represents the channel estimation vector corresponding to the near-field channel estimation matrix, which is then transformed into a vector matrix to obtain... ; Step 10: Update the near-field residual matrix and residual vector: , ; Step 11: Number of Update Iterations : ,if If the result is positive, repeat step 8 to continue iterating; otherwise, the iteration ends, and the latest iteration result is obtained. ; Step 12: If the value of the residual vector is less than or equal to the minimum value of the residual vector, that is... Then update the minimum residual vector. ; Phase 3: Output the final estimation results; Step 13: Initialize the hybrid field channel matrix: ; Step 14: If the far-field support set If near-field support set .

2. The method for channel estimation in a mixed-field ultra-large-scale MIMO scenario with unknown scene proportions based on the OMP algorithm according to claim 1, characterized in that, The signal model from the transmitter to the receiver includes: in, This indicates the pilot signal received by the receiver. Represents a merged matrix. This represents the channel matrix between the transmitter and receiver; N represents the noise at the receiver. This indicates the pilot signal transmitted by the transmitter; M represents the number of time slots. Indicates the first Pilot signals received by the receiver in each time slot; Indicates the number of RF chains in the receiving antenna array at the receiving end; Indicates the first Pilot signals transmitted by each time slot transmitter; Represents a hybrid precoding matrix; express Received noise, For the first Noise power after time slot combining Indicates size is The identity matrix; This indicates that the noise follows a complex Gaussian distribution. , Indicates size is The identity matrix, It represents the Kronecker product.

3. The method for channel estimation in a mixed-field ultra-large-scale MIMO scenario with unknown scene proportions based on the OMP algorithm according to claim 1, characterized in that, The hybrid field channel model includes: in, This represents a hybrid field channel model.

Images