Massive multi-transmit multi-receive system channel prediction method, system and medium

Abstract

This invention provides a method, system, and medium for channel prediction in a massive MIMO TDD system using a base-extended extrapolation model. The method includes: modeling the uplink channel using a base-extended model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating these unknown BEM coefficients; representing the estimated uplink channel as a linear combination of a discrete ellipsoidal BEM and its corresponding coefficients to reduce the complexity of downlink channel prediction; predicting the coefficients of a downlink time-varying DPS-BEM, and using the predicted DPS-BEM coefficients for downlink channel recovery. This invention alleviates the channel aging problem in massive MIMO TDD systems.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, and in particular to a method, system, and medium for channel prediction in a large-scale multiple-transmit multiple-receive system using base extension extrapolation. Background Technology

[0002] Massive multiple-input multiple-output (MIMO) is widely considered a key technology for 5G and 6G because it effectively utilizes spatial multiplexing gain, significantly improving channel spectral efficiency. Time division duplex (TDD) is a widely used duplexing scheme due to its advantages such as high spectral efficiency, flexible modulation of uplink and downlink traffic, low operating costs, and channel reciprocity. However, in high-mobility scenarios, the channel exhibits rapidly changing characteristics due to Doppler spread, and downlink channel aging may occur, rendering channel reciprocity in TDD systems unusable. Furthermore, estimating the downlink channel requires a large amount of reference signal, leading to a decrease in system spectral efficiency, and mobile terminals are often insufficient to support the enormous computational overhead of channel estimation. Summary of the Invention

[0003] The main objective of this invention is to provide a method, system, and medium for channel prediction in a large-scale MIMO TDD system based on base extension extrapolation, which aims to alleviate the channel aging problem in such systems.

[0004] To achieve the above objectives, this invention proposes a base-extended extrapolation channel prediction method for large-scale multiple-send multiple-receive systems, the method comprising the following steps:

[0005] The uplink channel is modeled using a base extension model, which represents the uplink channel as a small number of unknown BEM coefficients. The uplink channel is then reconstructed by estimating these small number of unknown BEM coefficients.

[0006] The estimated uplink channel is represented as a linear combination of a discrete ellipsoidal BEM and its corresponding coefficients to reduce the complexity of downlink channel prediction.

[0007] The coefficients of the downlink time-varying DPS-BEM are predicted, and the predicted DPS-BEM coefficients are used for downlink channel recovery.

[0008] A further technical solution of the present invention is that, in the step of modeling the uplink channel using a base extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating the BEM coefficients, a multi-user massive MIMO TDD system is considered, equipped with Base station service with root antenna For a user equipped with a single antenna, different time resources are allocated to uplink and downlink. Uplink transmission blocks include reference signals and data, while downlink only transmits data. The mobile device transmits data at a certain speed. It maintains relative movement with the base station.

[0009] A further technical solution of the present invention is that, in the step of modeling the uplink channel using a basis extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating the BEM coefficients, an orthogonal frequency division multiplexing modulation scheme is considered, and the frequency domain transmitted signal is... ,in This represents the number of subcarriers; after inverse discrete Fourier transform, the time-domain transmitted signal is obtained. ,in The discrete Fourier transform matrix is ​​used; inter-symbol interference is avoided by adding a sufficiently long cyclic prefix to the time-domain transmitted signal; at the receiver, after removing the cyclic prefix, the received frequency-domain signal can be represented as:

[0010] (1)

[0011] in Indicates the first The received signal from the root receiving antenna, Indicates the first The root receiving antenna has a mean of 0 and a variance of 0. Additive complex white Gaussian noise, Indicates the first The time-domain channel corresponding to the root receiving antenna is a pseudo-circular matrix, which can be represented as:

[0012] (2)

[0013] In high-mobility scenarios, the channel is time-varying, and the channel matrix... The number of unknown parameters is ,in Indicates the number of paths in the channel.

[0014] A further technical solution of the present invention is that the step of modeling the uplink channel using a radix extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating the BEM coefficients includes:

[0015] A radix extension model is used to model the time-varying channel:

[0016] Let the first The first antenna corresponding to the The time-varying channel of the path is The extended model using the complex exponential basis is expressed as follows:

[0017] (3)

[0018] in for The basis matrix of the extended model of the complex exponential basis. As the base expansion factor, To model the error, each column of the basis extension matrix is... ;generally In other words, the channel Only The signal can be approximated by a single coefficient; substituting equation (3) into equation (1), the base station can obtain the following received signal model:

[0019] (4)

[0020] in Denotes a permutation matrix, if It is achieved by applying an identity matrix Shift the column left circularly If you get it the next time, ,on the contrary. It is a set of discrete Fourier matrices. Submatrix composed of columns .

[0021] A further technical solution of the present invention is that the step of modeling the uplink channel using the radix extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating these BEM coefficients further includes:

[0022] Pilot-assisted channel estimation is used in the transmitted signal Insert One effective pilot, with each effective pilot deployed before and after it. A protection pilot is used to prevent the received signal from inter-carrier interference:

[0023] Assume the set of valid pilot indices is The received signal can then be expressed as:

[0024] (5)

[0025] in ,gather ;

[0026] Equation (5) can be rewritten in a combined form as follows:

[0027] (6)

[0028] The received signals from all receiving antennas at the base station can be represented as:

[0029] (7)

[0030] in In order to receive signals, The complex exponential basis expansion coefficient, This is the noise matrix.

[0031] A further technical solution of the present invention is that the step of modeling the uplink channel using the radix extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating these BEM coefficients further includes:

[0032] Using a parameterized model to model large-scale MIMO channels:

[0033] Among them, the The channel path and the channels corresponding to all receiving antennas can be represented as:

[0034] (8)

[0035] in , The number of activated clusters, For the first The number of rays in each activated ray cluster Indicates the first The complex gain of each ray, For the first The maximum Doppler frequency shift of each ray This represents the sampling period. Assuming the base station uses a uniform linear array, the steering vector... It can be represented as:

[0036] (9)

[0037] in Let AoA be the angle of arrival (Angle of Arrival) of the k-th activation cluster of the i-th ray. For the signal wavelength, For any two antennas, the angle of arrival of each active cluster can be expressed as: ,here This represents the angle of arrival at the center of the k-th active cluster. This represents an angular offset relative to the central angle and satisfies... ,in For the angular expansion of the k-th activated cluster, due to the limited scattering environment of the mobile device, the angular expansion... It is usually a relatively small value.

[0038] A further technical solution of the present invention is that the step of modeling the uplink channel using the radix extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating these BEM coefficients further includes:

[0039] Spatial domain channel is modeled using a spatial basis extension model. It can be represented as:

[0040] (10)

[0041] in Represents the spatial basis extension model. For spatial basis expansion coefficients, To model the error, we choose the generalized complex exponential basis extension model to construct the spatial basis extension model. The generalized complex exponential basis extended model is defined as follows: ,in , For the modeling frequency of the complex exponential basis extended model, Let the order be the complex exponential basis extended model. To model the resolution parameters of the frequency, typically ;

[0042] Substituting equation (10) into equations (3) and (7), we can obtain a new received signal model:

[0043] (11)

[0044] in For the measurement matrix, For the GCE-BEM coefficients to be estimated, the linear minimum mean square error (LMMSE) estimator can be used to obtain the solution for the upward GCE-BEM coefficients:

[0045] (12)

[0046] in For the estimated upward GCE-BEM coefficient, For matrix The noise variance of each column can be obtained through the uplink channel. Restore to .

[0047] A further technical solution of the present invention is that the step of predicting the coefficients of the downlink time-varying DPS-BEM and using the predicted DPS-BEM coefficients for downlink channel recovery includes:

[0048] Assuming the uplink contains within a single frame Given OFDM symbols, the estimated uplink GCE-BEM coefficients for the j-th symbol are: It can be expanded into vector form: ,but The uplink GCE-BEM coefficients corresponding to each OFDM symbol can be expressed as: To predict the downlink GCE-BEM coefficient based on the uplink GCE-BEM coefficient, the uplink GCE-BEM coefficient is first approximated using Legendre polynomials to obtain a set of fitting parameters. Then, the downlink GCE-BEM coefficient is predicted by iteratively extrapolating the fitting parameters.

[0049] Legendre polynomials can be obtained through the following recurrence relation:

[0050] (13)

[0051] in , The upward GCE-BEM coefficients fitted using Legendre polynomials can be expressed as:

[0052] (14)

[0053] in for The j-th row and m-th column, Let be the order of the discrete Legendre polynomial. For discrete Legendre polynomial coefficients, The fitting error for the discrete Legendre polynomial. The noise is additive Gaussian noise in the downlink; Equation (14) can be written in matrix-vector form as follows:

[0054] (15)

[0055] in Let be the discrete Legendre polynomial basis matrix, and , Let be the coefficient matrix of the discrete Legendre polynomial, and , This represents the total error term;

[0056] The fitting coefficient for the k-th iteration can be calculated as follows:

[0057] (16)

[0058] in For the weighted error matrix, when there is no prior error information Assume the step size for each extrapolation is... ,but The discrete Legendre polynomial extrapolation model of the step can be expressed as:

[0059] (17)

[0060] In the (k+1)th iteration, the predicted downlink GCE-BEM coefficients are added to the existing GCE-BEM coefficient sample, and updated as follows:

[0061] (18)

[0062] The discrete Legendre polynomial matrix also needs to be updated:

[0063] (19)

[0064] The predicted first The downlink GCE-BEM coefficient can be expressed as:

[0065] (20)

[0066] Thus predicting the first The step downlink channel can be represented as:

[0067] (twenty one).

[0068] To achieve the above objectives, the present invention also proposes a base-extended extrapolation large-scale multiple-receiver system channel prediction system. The system includes a memory, a processor, and a base-extended extrapolation large-scale multiple-receiver system channel prediction program stored on the processor. The base-extended extrapolation large-scale multiple-receiver system channel prediction program is executed by the processor to perform the steps of the method described above.

[0069] To achieve the above objectives, the present invention also proposes a computer-readable storage medium, characterized in that the computer-readable storage medium stores a base-extended extrapolation large-scale multiple-receiver system channel prediction program, wherein the base-extended extrapolation large-scale multiple-receiver system channel prediction program is executed by a processor to perform the steps of the method described above.

[0070] The beneficial effects of the present invention regarding the basis extension extrapolation channel prediction method, system, and medium for large-scale multiple-receiver systems are as follows: The present invention utilizes a basis extension model to model the uplink channel, representing it as a small number of unknown BEM coefficients. The uplink channel is reconstructed by estimating these unknown BEM coefficients. The estimated uplink channel is represented as a linear combination of a discrete ellipsoidal BEM and its corresponding coefficients, reducing the complexity of downlink channel prediction. The coefficients of the downlink time-varying DPS-BEM are predicted, and the predicted DPS-BEM coefficients are used for downlink channel recovery, alleviating the channel aging problem in large-scale MIMO TDD systems. Attached Figure Description

[0071] Figure 1 This is a flowchart illustrating a preferred embodiment of the channel prediction method for large-scale multiple-send multiple-receive systems based on the extended extrapolation of the present invention.

[0072] Figure 2 This is a schematic diagram of a system model for a large-scale MIMO TDD system in a high-mobility scenario;

[0073] Figure 3 This is a flowchart of uplink channel estimation and downlink channel prediction for a TDD massive MIMO-OFDM system;

[0074] Figure 4 This is a schematic diagram illustrating the performance of the channel prediction algorithm in predicting the downlink channel.

[0075] Figure 5 This is a schematic diagram illustrating the variation of the normalized mean square error of downlink channel prediction with downlink prediction length.

[0076] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. Detailed Implementation

[0077] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0078] Please refer to Figure 1 This invention proposes a base-extended extrapolation channel prediction method for large-scale multiple-receiver systems. A preferred embodiment of this method includes the following steps:

[0079] Step S10: Model the uplink channel using the basis expansion model (BEM), representing the uplink channel as a small number of unknown BEM coefficients, and reconstruct the uplink channel by estimating the small number of unknown BEM coefficients.

[0080] Step S20: The estimated uplink channel is represented as a linear combination of a discrete ellipsoid BEM and its corresponding coefficients to reduce the complexity of downlink channel prediction.

[0081] Specifically, in step S20, the base station represents the estimated uplink channel as a linear combination of a discrete ellipsoid BEM and its corresponding coefficients to reduce the complexity of downlink channel prediction.

[0082] Step S30: Predict the coefficients of the downlink time-varying DPS-BEM and use the predicted DPS-BEM coefficients for downlink channel recovery.

[0083] Specifically, in this embodiment, a base-extended extrapolation channel predictor is designed to predict the coefficients of the downlink time-varying DPS-BEM, and the predicted DPS-BEM coefficients are used for downlink channel recovery.

[0084] In this embodiment, the step of modeling the uplink channel using a base extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating the BEM coefficients, considers a multi-user massive MIMO TDD system equipped with Base station service with root antenna For a user equipped with a single antenna, different time resources are allocated to uplink and downlink. Uplink transmission blocks include reference signals and data, while downlink only transmits data. The mobile device transmits data at a certain speed. It maintains relative movement with the base station. In this system, channel aging is mainly caused by transmission delay and processing delay. See the system model diagram below. Figure 2 In the picture This indicates the channel aging time.

[0085] In this embodiment, the step of modeling the uplink channel using a base-extended model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating the BEM coefficients, considers using orthogonal frequency division multiplexing (OFDM) modulation, with the frequency domain transmitted signal being... ,in This represents the number of subcarriers. After inverse discrete Fourier transform, the time-domain transmitted signal is obtained. ,in Let be the discrete Fourier transform matrix. Inter-symbol interference is avoided by adding a sufficiently long cyclic prefix to the transmitted signal in the time domain. At the receiver, after removing the cyclic prefix, the received frequency domain signal can be represented as:

[0086] (1)

[0087] in Indicates the first The received signal from the root receiving antenna, Indicates the first The root receiving antenna has a mean of 0 and a variance of 0. Additive complex white Gaussian noise, Indicates the first The time-domain channel corresponding to the root receiving antenna is a pseudo-circular matrix, which can be represented as:

[0088] (2)

[0089] In high-mobility scenarios, the channel is time-varying, and the channel matrix... The number of unknown parameters is ,in Indicates the number of paths in the channel.

[0090] In this embodiment, the step of modeling the uplink channel using the base extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating the BEM coefficients includes:

[0091] A basis expansion model (BEM) is used to model the time-varying channel. In this embodiment, to reduce the number of unknown channel parameters, a basis expansion model (BEM) is used to model the time-varying channel.

[0092] Specifically, let the first The first antenna corresponding to the The time-varying channel of the path is The complex exponential basis extension model (CE-BEM) is used to represent it as follows:

[0093] (3)

[0094] in for The basis matrix of the extended model of the complex exponential basis. As the base expansion factor, To model the error, each column of the basis extension matrix is... ;generally In other words, the channel Only The signal can be approximated by a single coefficient; substituting equation (3) into equation (1), the base station can obtain the following received signal model:

[0095] (4)

[0096] in Denotes a permutation matrix, if It is achieved by applying an identity matrix Shift the column left circularly If you get it the next time, ,on the contrary. It is a set of discrete Fourier matrices. Submatrix composed of columns .

[0097] In this embodiment, the step of modeling the uplink channel using the base extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating these BEM coefficients further includes:

[0098] Pilot-assisted channel estimation is used in the transmitted signal Insert One effective pilot, with each effective pilot deployed before and after it. A protection pilot is used to prevent the received signal from being affected by inter-carrier interference.

[0099] Specifically, assume the set of indices for the effective pilots is as follows: The received signal can then be expressed as:

[0100] (5)

[0101] in ,gather ;

[0102] Equation (5) can be rewritten in a combined form as follows:

[0103] (6)

[0104] The received signals from all receiving antennas at the base station can be represented as:

[0105] (7)

[0106] in In order to receive signals, The complex exponential basis expansion coefficient, This is the noise matrix.

[0107] In this embodiment, the step of modeling the uplink channel using the base extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating these BEM coefficients further includes:

[0108] A parametric model is used to model large-scale MIMO channels.

[0109] Among them, the The channel path and the channels corresponding to all receiving antennas can be represented as:

[0110] (8)

[0111] in , The number of activated clusters, For the first The number of rays in each activated ray cluster Indicates the first The complex gain of each ray, For the first The maximum Doppler frequency shift of each ray This represents the sampling period. Assuming the base station uses a uniform linear array, the steering vector... It can be represented as:

[0112] (9)

[0113] in Let AoA be the angle of arrival (Angle of Arrival) of the k-th activation cluster of the i-th ray. For the signal wavelength, For any two antennas, the angle of arrival of each active cluster can be expressed as: ,here This represents the angle of arrival at the center of the k-th active cluster. This represents an angular offset relative to the central angle and satisfies... ,in For the angular expansion of the k-th activated cluster, due to the limited scattering environment of the mobile device, the angular expansion... It is usually a relatively small value.

[0114] In this embodiment, the step of modeling the uplink channel using the base extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating these BEM coefficients further includes:

[0115] Spatial domain channel is modeled using a spatial basis extension model. It can be represented as:

[0116] (10)

[0117] in Represents the spatial basis extension model. For spatial basis expansion coefficients, To model the error, we choose the generalized complex exponential basis extension model to construct the spatial basis extension model. The generalized complex exponential basis extended model is defined as follows: ,in , For the modeling frequency of the complex exponential basis extended model, Let the order be the complex exponential basis extended model. To model the resolution parameters of the frequency, typically ;

[0118] Substituting equation (10) into equations (3) and (7), we can obtain a new received signal model:

[0119] (11)

[0120] in For the measurement matrix, For the GCE-BEM coefficients to be estimated, the linear minimum mean square error (LMMSE) estimator can be used to obtain the solution for the upward GCE-BEM coefficients:

[0121] (12)

[0122] in For the estimated upward GCE-BEM coefficient, For matrix The noise variance of each column can be obtained through the uplink channel. Restore to .

[0123] In this embodiment, unlike existing prediction algorithms that directly predict channel gain during the downlink prediction stage, this invention proposes to indirectly achieve downlink channel prediction by predicting the downlink BEM coefficient.

[0124] Specifically, the step of predicting the coefficients of the downlink time-varying DPS-BEM and using the predicted DPS-BEM coefficients for downlink channel recovery includes:

[0125] Assuming the uplink contains within a single frame Given OFDM symbols, the estimated uplink GCE-BEM coefficients for the j-th symbol are: It can be expanded into vector form: ,but The uplink GCE-BEM coefficients corresponding to each OFDM symbol can be expressed as: To predict the downlink GCE-BEM coefficient based on the uplink GCE-BEM coefficient, the uplink GCE-BEM coefficient is first approximated using Legendre polynomials to obtain a set of fitting parameters. Then, the downlink GCE-BEM coefficient is predicted by iteratively extrapolating the fitting parameters.

[0126] Legendre polynomials can be obtained through the following recurrence relation:

[0127] (13)

[0128] in , The upward GCE-BEM coefficients fitted using Legendre polynomials can be expressed as:

[0129] (14)

[0130] in for The j-th row and m-th column, Let be the order of the discrete Legendre polynomial. For discrete Legendre polynomial coefficients, The fitting error for the discrete Legendre polynomial. The noise is additive Gaussian noise in the downlink; Equation (14) can be written in matrix-vector form as follows:

[0131] (15)

[0132] in Let be the discrete Legendre polynomial basis matrix, and , Let be the coefficient matrix of the discrete Legendre polynomial, and , This represents the total error term;

[0133] The fitting coefficient for the k-th iteration can be calculated as follows:

[0134] (16)

[0135] in For the weighted error matrix, when there is no prior error information Assume the step size for each extrapolation is... ,but The discrete Legendre polynomial extrapolation model of the step can be expressed as:

[0136] (17)

[0137] In the (k+1)th iteration, the predicted downlink GCE-BEM coefficients are added to the existing GCE-BEM coefficient sample, and updated as follows:

[0138] (18)

[0139] The discrete Legendre polynomial matrix also needs to be updated:

[0140] (19)

[0141] The predicted first The downlink GCE-BEM coefficient can be expressed as:

[0142] (20)

[0143] Thus predicting the first The step downlink channel can be represented as:

[0144] (twenty one).

[0145] in, This involves operations that convert a vector into a matrix. The pseudocode for the downlink channel prediction algorithm is summarized in Algorithm 1.

[0146]

[0147] The block diagram of the entire uplink channel estimation and downlink channel prediction method is as follows: Figure 3 As shown.

[0148] Experimental comparison

[0149] To verify the performance of the method of the present invention, simulation experiments were conducted.

[0150] Experiment 1

[0151] Assume the number of antennas at the base station is Number of subcarriers The order of the complex exponential basis extended model GCE-BEM order .from Figure 4 It can be observed that the proposed channel prediction algorithm can accurately predict the downlink channel, achieving a channel prediction error of -29 dB even at a speed of 240 km / h. Figure 5 It can be seen that the proposed channel prediction algorithm can effectively predict long downlink channels, such as a channel prediction error of -10 dB when predicting a downlink channel of 5 OFDM symbols.

[0152] The beneficial effects of the basis extension extrapolation channel prediction method for large-scale multiple-receiver systems of the present invention are as follows: The present invention utilizes the basis extension model to model the uplink channel, representing it as a small number of unknown BEM coefficients, and reconstructs the uplink channel by estimating these unknown BEM coefficients; the estimated uplink channel is represented as a linear combination of a discrete ellipsoidal BEM and its corresponding coefficients, thereby reducing the complexity of downlink channel prediction; the coefficients of the downlink time-varying DPS-BEM are predicted, and the predicted DPS-BEM coefficients are used for downlink channel recovery, alleviating the channel aging problem in large-scale MIMO TDD systems.

[0153] To achieve the above objectives, the present invention also proposes a base-extended extrapolation large-scale multiple-receiver system channel prediction system. The system includes a memory, a processor, and a base-extended extrapolation large-scale multiple-receiver system channel prediction program stored on the processor. When the base-extended extrapolation large-scale multiple-receiver system channel prediction program is run by the processor, the steps of the method described in the above embodiments are executed, and will not be repeated here.

[0154] To achieve the above objectives, the present invention also proposes a computer-readable storage medium storing a base-extended extrapolation large-scale multiple-receiver system channel prediction program, wherein the base-extended extrapolation large-scale multiple-receiver system channel prediction program is executed by a processor to perform the steps of the method described in the above embodiments.

[0155] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural changes made based on the content of the present invention specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for channel prediction in a large-scale multiple-send multiple-receive system using base extension extrapolation, characterized in that, The method includes the following steps: The uplink channel is modeled using a base extension model, which represents the uplink channel as a small number of unknown BEM coefficients. The uplink channel is then reconstructed by estimating these small number of unknown BEM coefficients. The estimated uplink channel is represented as a linear combination of a discrete ellipsoidal BEM and its corresponding coefficients to reduce the complexity of downlink channel prediction. Predict the coefficients of the downlink time-varying DPS-BEM and use the predicted DPS-BEM coefficients for downlink channel recovery; In the step of modeling the uplink channel using a base-extended model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating the BEM coefficients, a multi-user massive MIMO TDD system is considered, equipped with a base-extended model. Base station service with root antenna For a user equipped with a single antenna, different time resources are allocated to uplink and downlink. Uplink transmission blocks include reference signals and data, while downlink only transmits data. The mobile device transmits data at a certain speed. Maintain relative movement with the base station; In the step of modeling the uplink channel using a basis extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating these BEM coefficients, an orthogonal frequency division multiplexing (OFDM) modulation scheme is considered, and the frequency domain transmitted signal is... ,in This represents the number of subcarriers; after inverse discrete Fourier transform, the time-domain transmitted signal is obtained. ,in The discrete Fourier transform matrix is ​​used; inter-symbol interference is avoided by adding a sufficiently long cyclic prefix to the time-domain transmitted signal; at the receiver, after removing the cyclic prefix, the received frequency-domain signal is represented as: (1) in Indicates the first The received signal from the root receiving antenna, Indicates the first The mean and variance of the root receiving antenna are 0. Additive complex white Gaussian noise, Indicates the first The time-domain channel corresponding to the root receiving antenna is a pseudo-circular matrix, represented as: (2) In high-mobility scenarios, the channel is time-varying, and the channel matrix... The number of unknown parameters is ,in Indicates the number of paths in the channel; The step of modeling the uplink channel using a base-extended model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating the BEM coefficients includes: A radix extension model is used to model the time-varying channel: Let the first The first antenna corresponding to the The time-varying channel of the path is The extended model using the complex exponential basis is expressed as follows: (3) in for The basis extension matrix of the complex exponential basis extension model. As the base expansion factor, To model the error, each column of the basis extension matrix is... ; In other words, the channel Only The signal can be approximated by a single coefficient; substituting equation (3) into equation (1), the base station obtains the following received signal model: (4) in Denotes a permutation matrix, if It is achieved by applying an identity matrix. Perform a left circular shift on the column. If you get it the next time, ,on the contrary, It is a set of discrete Fourier matrices. Submatrix composed of columns .

2. The method for channel prediction in a massive MIMO system according to claim 1, characterized in that, The step of modeling the uplink channel using a base extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating these BEM coefficients further includes: Pilot-assisted channel estimation is used in the transmitted signal Insert One effective pilot, with each effective pilot deployed before and after it. A protection pilot is used to prevent the received signal from inter-carrier interference: Assume the set of indices for effective pilots is The received signal is then represented as: (5) in ,gather ; Equation (5) can be rewritten in a combined form as follows: (6) The received signals from all receiving antennas at the base station are represented as follows: (7) in In order to receive signals, The complex exponential basis expansion coefficient, This is the noise matrix.

3. The method for channel prediction in a massive MIMO system based on base extension extrapolation according to claim 2, characterized in that, The step of modeling the uplink channel using a base extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating these BEM coefficients further includes: Using a parameterized model to model large-scale MIMO channels: Among them, the The channel path and the channel corresponding to all receiving antennas are represented as follows: (8) in , The number of activated clusters, For the first The number of rays in each activated ray cluster Indicates the first The complex gain of each ray, For the first The maximum Doppler frequency shift of each ray This represents the sampling period; assuming the base station uses a uniform linear array, the steering vector... Represented as: (9) in Let be the angle of arrival of the k-th activation cluster of the i-th ray. For the signal wavelength, Let the distance between any two antennas be denoted as , and the angle of arrival for each active cluster be denoted as . ,here This represents the angle of arrival at the center of the k-th active cluster. This represents an angular offset relative to the central angle and satisfies... ,in The angle expansion for the k-th activated cluster.

4. The base-extended extrapolation channel prediction method for large-scale multiple-send multiple-receive systems according to claim 3, characterized in that, The step of modeling the uplink channel using a base extension model, representing the uplink channel as a small number of unknown BEM coefficients, and reconstructing the uplink channel by estimating these BEM coefficients further includes: Spatial domain channel is modeled using a spatial basis extension model. , is represented as: (10) in Represents the spatial basis extension model. For spatial basis expansion coefficients, To mitigate modeling errors, a generalized complex exponential basis extension model is chosen to construct the spatial basis extension model. The generalized complex exponential basis extended model is defined as follows: ,in , For the modeling frequency of the complex exponential basis extended model, Let the order be the complex exponential basis extended model. To model the resolution parameters of the frequency. ; Substituting equation (10) into equations (3) and (7), we obtain a new received signal model: (11) in For the measurement matrix, For the GCE-BEM coefficients to be estimated, the linear minimum mean square error estimator is used to obtain the solution for the upward GCE-BEM coefficients: (12) in For the estimated upward GCE-BEM coefficient, For matrix The noise variance of each column, the uplink channel through Restore to .

5. The base-extended extrapolation channel prediction method for large-scale multiple-receiver systems according to claim 4, characterized in that, The step of predicting the coefficients of the downlink time-varying DPS-BEM and using the predicted DPS-BEM coefficients for downlink channel recovery includes: Assuming the uplink contains within a single frame The uplink GCE-BEM coefficients estimated by the j-th symbol are: (The original text contains several grammatical errors and inconsistencies. A more accurate translation would require the full context.) Expand into vector form: ,but The uplink GCE-BEM coefficients corresponding to each OFDM symbol are expressed as follows: To predict the downlink GCE-BEM coefficient based on the uplink GCE-BEM coefficient, the uplink GCE-BEM coefficient is first approximated using Legendre polynomials to obtain a set of fitting parameters. Then, the downlink GCE-BEM coefficient is predicted by iteratively extrapolating the fitting parameters. Legendre polynomials are obtained through the following recurrence relation: (13) in , The upward GCE-BEM coefficients fitted using Legendre polynomials are expressed as follows: (14) in for The j-th row and m-th column, Let be the order of the discrete Legendre polynomial. For discrete Legendre polynomial coefficients, The fitting error for the discrete Legendre polynomial. The downlink noise is additive Gaussian noise; Equation (14) can be written in matrix-vector form as follows: (15) in Let be the discrete Legendre polynomial basis matrix, and , Let be the coefficient matrix of the discrete Legendre polynomial, and , This represents the total error term; The fitting coefficients for the k-th iteration are calculated as follows: (16) in For the weighted error matrix, when there is no prior error information Assume the step size for each extrapolation is... ,but The discrete Legendre polynomial extrapolation model of the step is expressed as: (17) In the (k+1)th iteration, the predicted downlink GCE-BEM coefficients are added to the existing GCE-BEM coefficient sample, and updated as follows: (18) The discrete Legendre polynomial matrix also needs to be updated: (19) The predicted first The downlink GCE-BEM coefficient is expressed as follows: (20) Thus predicting the first The downlink channel is represented as follows: (21)。 6. A channel prediction system for a base-extended extrapolation large-scale multiple-send multiple-receive system, characterized in that, The system includes a memory, a processor, and a base-extended extrapolation large-scale multiple-input multiple-receiver (MLM) channel prediction program stored on the processor, the base-extended extrapolation MML channel prediction program being executed by the processor to perform the steps of the method as described in any one of claims 1 to 5.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a base-extended extrapolation channel prediction program for a massive MIMO system, which, when executed by a processor, performs the steps of the method as described in any one of claims 1 to 5.

Images