Intelligent reconfigurable surface cascaded channel estimation method and system based on dimension reconstruction

Abstract

The method comprises the following steps: constructing a three-point cooperative intelligent reconfigurable surface communication model comprising three communication links; introducing an auxiliary configuration matrix on the three-point cooperative intelligent reconfigurable surface communication model; performing preliminary channel training on a single link to obtain a dimensionally reconstructed cascaded channel model; performing preliminary estimation on the three links through the dimensionally reconstructed cascaded channel model; transmitting a signal in an auxiliary transceiver to a base station; and decoupling and estimating a matrix of each segmented reflection channel by the base station on the estimated channel state information matrix. The application can be well applied in channel estimation of an ultra-large-scale intelligent reconfigurable surface cascaded channel, and can compensate for the loss of channel estimation accuracy caused by the increase in the scale of the intelligent reconfigurable surface.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of intelligent reconfigurable surface cascaded channel estimation, and particularly relates to an intelligent reconfigurable surface cascaded channel estimation method and system based on dimension reconstruction. BACKGROUND

[0002] Disasters such as earthquake mudslides destroy base stations in disaster areas, thereby interrupting wireless communication. Existing solutions mainly focus on arranging standby base stations in advance at important places. However, when disasters occur, a large number of disaster victims are scattered everywhere, so the base stations arranged in advance can only cover a small number of disaster victims. In order to further expand the coverage range of wireless communication in disasters, unmanned aerial vehicle-based emergency communication systems have been gradually valued in recent years.

[0003] With the continuous development of new generation wireless communication technology, intelligent reconfigurable surfaces have attracted widespread attention. They are low in price, easy to deploy and small in energy consumption, and can effectively transmit information that needs to be transmitted in disaster areas to the base station outside the disaster area through a reflection channel. Considering that the distance from the disaster area to the base station outside the disaster area is large and the communication system of the intelligent reconfigurable surface has a multiplicative attenuation path loss problem, an unmanned aerial vehicle carrying a super large intelligent reconfigurable surface is more practical in disaster communication.

[0004] Reliable intelligent reconfigurable surface reflection beam forming needs to rely on accurate channel state information estimation. In the channel state information estimation of the intelligent reconfigurable surface cascaded channel, the main difficulty lies in the mutual coupling of the cascaded channel state information and the need for a large number of parameters in estimation. The acquisition of the cascaded channel state information can mainly be divided into two kinds. One is implicit channel state information acquisition, also known as beam training, which aims to allow both sides of the communication to find the optimal one in the multiple direction beams predefined in the codebook. The other is explicit channel state information acquisition, also known as channel estimation, which aims to train the channel through the pilot signal known in advance by both the transmitter and the receiver and obtain the complete state information of the channel, including phase, amplitude and direction information.

[0005] In beam training, the prior art assumes that the optimal path between the intelligent reconfigurable surface and the base station has been determined, and based on the traditional beam training method in MIMO, the intelligent reconfigurable surface is divided into multiple subarrays, and the beam alignment between the base station and the user is completed. The traditional hierarchical multi-resolution beam search method is applied to the intelligent reconfigurable surface. The base station, the intelligent reconfigurable surface and the user each obtain a multi-layer beam forming codebook for joint spatial scanning to obtain the best directional beam of the RIS cascaded channel. The prior art is mainly devoted to reducing the complexity and training overhead of the search. In channel estimation, the traditional least squares method is applied and certain achievements are made, the deep reinforcement learning output continuous beam forming matrix is used for phase shift of BS and RIS, and the intelligent reconfigurable surface with a small number of active elements is designed and the neural network is used for estimation of the cascaded channel; and the method of channel estimation by means of decomposition of the parallax tensor is also devoted.

[0006] Defects and deficiencies of the prior art:

[0007] The methods in the cited documents are only applicable to small-scale intelligent reconfigurable surfaces;

[0008] The super-large-scale intelligent reconfigurable surface leads to an increase in the Rayleigh distance, which makes the design of the encoding codebook need to add the theory related to the near-field channel model, but the methods in the above documents are all based on the far-field theory;

[0009] The channel estimation methods at the present stage are mainly concentrated on intelligent reconfigurable surfaces of 16x16 and below. Although the intelligent reconfigurable surface of 64x64 is studied, the estimation accuracy also decreases obviously with the increase of the size of the intelligent reconfigurable surface;

[0010] The active element intelligent reconfigurable surface mentioned in the prior art has a large energy consumption, and also needs a large amount of training in the offline stage, which has a large time overhead. SUMMARY

[0011] The purpose of the present application is to provide an intelligent reconfigurable surface cascaded channel estimation method and system based on dimension reconstruction, to solve the above problems.

[0012] To achieve the above purpose, the present application adopts the following technical solutions:

[0013] In a first aspect, the present application provides an intelligent reconfigurable surface cascaded channel estimation method based on dimension reconstruction, comprising the following steps:

[0014] Constructing a three-point cooperative intelligent reconfigurable surface communication model comprising three communication links;

[0015] Introducing auxiliary configuration matrix on the three-point cooperative intelligent reconfigurable surface communication model, simplifying the configuration process of intelligent reconfigurable surface;

[0016] Preliminary channel training is carried out on a single link to obtain a dimensionally reconstructed cascade channel model;

[0017] The three links are preliminarily estimated through the dimensionally reconstructed cascade channel model, the signal in the auxiliary transceiver is sent to the base station, and the base station decouples the estimated channel state information matrix and estimates the matrix of each segmented reflection channel.

[0018] Optionally, the three communication links include link one: user-intelligent reconfigurable surface-base station; link two: auxiliary transceiver-intelligent reconfigurable surface-base station; link three: user-intelligent reconfigurable surface-auxiliary transceiver; the channel models of the three links are Y1=HΦ l GX1+Z1, Y2=HΦ l J T X2+Z2, Y3=JΦ l GX3+Z3, wherein the matrix Z is noise, Y and X are received signal and transmitted signal respectively, Φ is the phase configuration matrix of the intelligent reconfigurable surface, and is a diagonal matrix; the related matrix dimensions are: wherein N t is the number of antennas of the transmitter user, N r is the number of antennas of the receiver base station, N a is the number of antennas of the auxiliary transceiver, N s ×N s is the number of arrays of the intelligent reconfigurable surface, and P is the number of time slots of the pilot signal; in the model, the auxiliary transceiver mainly receives the pilot signal of link 3 and preliminarily estimates the channel, and in link two, the auxiliary transceiver serves as a transmitting device to facilitate the base station to estimate the channel of the link, and the auxiliary transceiver sends the estimated channel state information to the base station.

[0019] Optionally, an auxiliary configuration matrix Ψ a and Ψ b are introduced, wherein Ψ a is a pre-aid matrix, and Ψ b is a post-aid matrix; the total configuration matrix of the intelligent reconfigurable surface after introducing the auxiliary matrix is Ψ b Φ l Ψ a .

[0020] Optionally, the preliminary channel training method for a single link includes sending multiple pilot signals, configuration transformation of the intelligent reconfigurable surface, and preliminary acquisition of channel state information at the base station side;

[0021] The transmitted pilot signals need to pass through multiple different configured smart reconfigurable surfaces and a series of received signals are obtained at the base station; the original received signals are integrated into a dimensionally reconstructed received signal matrix, which, after decomposition, obtains a dimensionally reconstructed cascaded channel model:

[0022]

[0023] where N is the number of transmitted pilot signals X, [Ψ0, Ψ1, …, Ψ N-1 ] and [Ψ0, Ψ1, …, Ψ M-1 ] are the pre- and post-aiding matrices, H config = ([(HΨ0) T ,(HΨ1) T ,…,(HΨ M-1 ) T ] T is called the dimensionally reconstructed H channel, G config = [Ψ0G, Ψ1G, …, Ψ N-1 G] is called the dimensionally reconstructed G channel; in the estimation, the channel matrix and the dimensionally reconstructed channel matrix are split into amplitude matrices, denoted as A ∝ , phase matrices denoted as A p , and direction matrices denoted as A d , that is, for the channel matrix denoted as A, the following split is made:

[0024] (1) the transmitting side—smart reconfigurable surface: A = A ∝ A p A d ; (2) the receiving side—smart reconfigurable surface: A = A d A p A ∝ , and the direction matrix is phase calibrated by an optimization method so that the phase and direction information are estimated together; the preliminary training finally obtains a channel state information estimation matrix for the dimensionally reconstructed direction matrices H config,dp ,G config,d of the channels H and G and the coupled gain matrix: the estimation matrix of the product H config,α ·G config,α .

[0025] Optionally, preliminary estimation of the three links

[0026] According to the preliminary training method, the three links are preliminarily estimated for the channel state information, and the dimensionally reconstructed cascaded channel model is:

[0027] Link 2: where H config is the same as in formula (1), and J Tconfig = [Ψ0J T , Ψ1JT ,…,Ψ N′-1 J T ], N a is the number of antennas of the secondary transceiver;

[0028] Link 3: where J config = ([(JΨ0) T ,(JΨ1) T ,…,(jΨ N′-1 ) T ]) T and G config is the same as in equation (1), it is proved that J coNfig and J Tconfig are transposes of each other;

[0029] The cascaded channel model of link 1 dimension reconstruction is shown in equation (1);

[0030] Finally, the partial channel state information of links 1 and 2 is estimated at the base station: the H channel direction matrix of link 1 dimension reconstruction after phase calibration: The G channel direction matrix of link 1 dimension reconstruction: The H and G channel coupling gain matrix of link 1 dimension reconstruction: The H channel direction matrix of link 2 dimension reconstruction after phase calibration: The J T channel direction matrix of link 2 dimension reconstruction: The H and J T channel coupling gain matrix of link 2 dimension reconstruction The information related to link 3 is obtained in the secondary transceiver: the J channel direction matrix of link 3 dimension reconstruction after phase calibration: The G channel direction matrix of link 3 dimension reconstruction The J and G channel coupling gain matrix of link 3 dimension reconstruction:

[0031] Optionally, the transmission of the secondary transceiver to the base station:

[0032] The coupling gain matrix J of link 3 is transmitted to the base station; and is converted into a series of N a ×N a sub-matrices E0, E1, …, E N′-1 ; the received signal is represented as the following matrix after arrangement and is decomposed into the form of the dimension reconstruction channel model:

[0033]

[0034] Combining the channel state information acquired by the base station, the estimation formula (3) of the sub-matrix is obtained and the gain matrix of link 3 is recovered

[0035]

[0036] Optionally, the channel state information decoupling and final estimation at the base station side:

[0037] The base station obtains the directional matrix of each of the three segmented channels H, G and J, and decouples the gain matrix for coupling: According to the expansion of the three, the following equation group is obtained:

[0038]

[0039] Solving obtains the gain matrix of each channel and the estimated reconstructed channel:

[0040]

[0041] The matrix W in the above is a random disturbance term;

[0042] Take the first N r rows of , the first N s rows of , and the first B a rows of to obtain the final estimated channel matrix

[0043] In a second aspect, the present application provides a dimension reconstruction-based intelligent reconfigurable surface cascaded channel estimation system, characterized in that it comprises:

[0044] A model construction module is configured to construct a three-point cooperative intelligent reconfigurable surface communication model comprising three communication links.

[0045] An auxiliary configuration module is configured to introduce an auxiliary configuration matrix on the three-point cooperative intelligent reconfigurable surface communication model to simplify the configuration process of the intelligent reconfigurable surface.

[0046] A training module is configured to perform preliminary channel training on a single link to obtain a dimension reconstruction-based cascaded channel model.

[0047] An estimation output module is configured to perform preliminary estimation on the three links through the dimension reconstruction-based cascaded channel model, send the signal in the auxiliary transceiver to the base station, and decouple the estimated channel state information matrix at the base station to estimate the matrix of each segmented reflection channel.

[0048] In a third aspect, the present application provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for intelligent reconfigurable surface cascade channel estimation based on dimension reconstruction when executing the computer program.

[0049] In a fourth aspect, the present application provides a computer readable storage medium storing a computer program, wherein the computer program implements the steps of the method for intelligent reconfigurable surface cascade channel estimation based on dimension reconstruction when executed by a processor.

[0050] Compared with the prior art, the present application has the following technical effects:

[0051] The present application can be well applied in channel estimation of ultra-large intelligent reconfigurable surface cascade channel, and can compensate for the loss of estimation accuracy caused by the increase of the size of the intelligent reconfigurable surface.

[0052] 1. The size of the intelligent reconfigurable surface increases, and the estimation accuracy rises. Unlike most existing intelligent reconfigurable surface cascade channel estimation schemes, the estimation accuracy in the present application rises with the increase of the size of the intelligent reconfigurable surface, indicating that this channel estimation method effectively compensates for the loss of estimation accuracy caused by the increase of the size of the intelligent reconfigurable surface.

[0053] 2. The present application has strong universality and can be used as preparation work for other channel estimation schemes. After sending the pilot signal according to the agreed pilot signal-intelligent reconfigurable surface configuration correspondence diagram and performing the specified receiving signal processing and integration work at the receiving end, the dimension reconstruction cascade channel model proposed in the present application can be constructed. The only difference between this model and the actual cascade channel model is the dimension of the channel matrix, and the dimension reconstruction channel therein can be estimated by other cascade channel estimation schemes;

[0054] 3. In the present application, the intelligent reconfigurable surface used only consists of a large number of passive units, which is low in cost and convenient to deploy;

[0055] 4. The idea of auxiliary matrix is proposed. If the complex intelligent reconfigurable surface configuration change is not decomposed and directly listed, it will be a great burden on the control circuit. In order to reduce the burden and simplify such configuration transformation, the present application proposes pre- and post- auxiliary configuration matrices, which can obtain a series of complex phase configurations combined with the original configuration matrix of the intelligent reconfigurable surface. In addition, the present application gives a sufficient condition for the construction of the auxiliary matrix meeting the system requirements and designs a binary reconstruction method according to this condition.

[0056] 5. A signal transmission scheme for assisting the transceiver to the base station is proposed. This scheme aims to transmit certain specific signals before channel estimation is completed, and recover the signals transmitted by the assisting transceiver in the base station based on the agreed sub-matrix—the intelligent reconfigurable surface configuration correspondence diagram—and the estimated partial channel state information. Attached Figure Description

[0057] Figure 1 It is a three-point collaborative intelligent reconfigurable surface communication model;

[0058] Figure 2 This is a diagram of a single-link communication model structure;

[0059] Figure 3 This is a diagram showing the correspondence between pilot signals and intelligent reconfigurable surface configuration matrices in the preliminary channel training method of Link 1, and also indicating the received signals of the base station under each correspondence.

[0060] Figure 4 This is a diagram showing the correspondence between pilot signals and intelligent reconfigurable surface placement matrices in the preliminary channel training methods for links 2 and 3. Figure 4 (a) is link 2, (b) is link 3;

[0061] Figure 5 (a) is a schematic diagram of matrix splitting in the transmission scheme from the auxiliary transceiver to the base station, and (b) is a diagram showing the correspondence between the sub-matrix and the intelligent reconfigurable surface configuration matrix during transmission.

[0062] Figure 6 It estimates the normalized mean square error of the system under different signal-to-noise ratios with different pilot numbers.

[0063] Figure 7 It estimates the normalized mean square error of the channel at different signal-to-noise ratios for intelligent reconfigurable surfaces of different sizes. Detailed Implementation

[0064] This invention provides a channel estimation scheme for ultra-large-scale intelligent reconfigurable surface cascaded channels, where the ultra-large-scale intelligent reconfigurable surface (Ns×Ns) mainly refers to an intelligent reconfigurable surface where the number of arrays Ns is greater than the number of transceiver antennas. This invention introduces an auxiliary simplified signal transceiver to construct a three-point collaborative intelligent reconfigurable surface communication model on top of the traditional transmitter-intelligent reconfigurable surface-receiver communication system (hereinafter referred to as the traditional RIS communication model). Furthermore, relevant channel state information is obtained by repeatedly sending pilot signals and adjusting the configuration of the intelligent reconfigurable surfaces during the channel training phase. This process includes a preliminary channel training method for the traditional RIS communication model, a signal transmission scheme from the simplified transceiver device to the base station, and also details the method for configuring the intelligent reconfigurable surfaces during the channel training phase.

[0065] A three-point collaborative intelligent reconfigurable surface communication model: Figure 1 This structure is illustrated. For ease of subsequent description, we divide the communication links in this structure into three: Link 1 is User – Smart Reconfigurable Surface – Base Station; Link 2 is Auxiliary Transceiver – Smart Reconfigurable Surface – Base Station; Link 3 is User – Smart Reconfigurable Surface – Auxiliary Transceiver. The relevant channel matrices have been labeled. Figure 1 In the diagram, the channel models for the three links are as follows: Y1 = HΦ l GX1+Z1(link 1), Y2=HΦ l J T X2+Z2(link 2), Y3=JΦ l GX3+Z3 (Link 3), where matrix Z represents noise, Y and X represent the received and transmitted signals respectively, and Φ is the phase configuration matrix of the intelligent reconfigurable surface, which is a diagonal matrix. The relevant matrix dimensions are: N t N represents the number of antennas of the transmitting (user) side. r N represents the number of antennas at the receiver (base station). a To assist in the number of antennas for both the transmitter and receiver, N s ×N s Let P be the number of arrays of intelligent reconfigurable surfaces, and P be the number of time slots for the pilot signal. In this model, the auxiliary transceiver mainly serves to receive the pilot signal from link 3 and perform preliminary channel estimation. Simultaneously, it acts as a transmitter in link 2 to facilitate channel estimation for the base station. Furthermore, the auxiliary transceiver transmits the estimated channel state information to the base station.

[0066] Intelligent Reconfigurable Surface Configuration Matrix: Considering the relatively complex configuration changes of intelligent reconfigurable surfaces involved in this scheme, an auxiliary configuration matrix Ψ is introduced to further simplify the configuration process. a and Ψ b Ψ a For the preceding auxiliary matrix, Ψ b As auxiliary matrices, these two matrices are set as diagonal matrices with diagonal elements containing only ±1. After introducing the auxiliary matrices, the total configuration matrix of the intelligent reconfigurable surface is Ψ. b Φ l Ψ a Furthermore, the channel model has changed to some extent. Taking link 1 as an example, its channel model is now Y1 = HΨ. b Φ l Ψ a GX1+Z1.

[0067] Preliminary channel training method for a single link: See the single-link communication model structure diagram. Figure 2The preliminary training of the model channel includes the transmission of multiple pilot signals, the configuration transformation of the smart reconfigurable surface, and the preliminary acquisition of the channel state information at the base station side. In the present scheme, the transmitted pilot signals need to pass through multiple different configurations of the smart reconfigurable surface (see detailed correspondence Figure 3 ) and obtain a series of received signals at the base station. In order to solve the channel state information from the received signals, the original received signals need to be processed in a series of processes and finally integrated into a dimensionally reconstructed received signal matrix (see processing process Figure 3 ), which can obtain a channel model similar to three links but with changed dimension of the channel matrix after decomposition. We call this decomposition as the dimensionally reconstructed cascaded channel model:

[0068]

[0069] where N is the number of transmitted pilot signals X, [Ψ0, Ψ1, …, Ψ N-1 ] and [Ψ0, Ψ1, …, Ψ M-1 ] are the pre- and post-aiding matrices, H config = ([(HΨ0) T ,(HΨ1) T ,…,(HΨ M-1 ) T ] T is called the dimensionally reconstructed H channel, and G config = [Ψ0G, Ψ1G, …, Ψ N-1 G] is called the dimensionally reconstructed G channel. The detailed steps of the base station's estimation of the dimensionally reconstructed H and G channels in model (1) are shown in the “dimensionally reconstructed cascaded channel estimation algorithm” in the “specific implementation” section. It needs to be noted that in the estimation, the channel matrix and the dimensionally reconstructed channel matrix are split into amplitude matrices (denoted as A ∝ ), phase matrices (denoted as A p ), and direction matrices (denoted as A d ), that is, for the channel matrix (denoted as A), the following split can be made: (1) the transmitting side, the smart reconfigurable surface: A = A ∝ A p A d ; (2) the smart reconfigurable surface, the receiving side: A = A d A p A ∝ . Considering that the phase information has less impact than the direction and gain information, in order to simplify the estimation steps, the direction matrix is phase-calibrated through an optimization method so that the phase and direction information are estimated together. In addition, the channel state information estimation matrix finally obtained by the preliminary training is the dimensionally reconstructed direction matrix H config,dp ,G config,d of the channels H and G and the coupled gain matrix: product Hconfig,α • G config,α estimation matrix.

[0070] Sufficient condition and construction method of auxiliary configuration matrix:

[0071]

Sufficient condition

[0072] According to the requirement of the rank of matrix in the scheme, it can be proved that the number of matrix N and M in the pre- and post- auxiliary matrix [Ψ0,Ψ1,…,Ψ N-1 ] and [Ψ0,Ψ1,…,Ψ M-1 ] should meet the condition: N·N t ≥N s ,M·N r =N s , and N t , N s and N r need to meet the multiple relationship. In the scheme, we take After adding the auxiliary transceiver, we take

[0073]

Construction method

[0074] The auxiliary matrix Ψ i is a diagonal matrix composed of ±1, which can be expressed as When the diagonal elements in it satisfy the sufficient condition above, we call such a construction method as the binary construction method.

[0075] Three-point cooperative channel estimation scheme: This scheme consists of three parts. The first is the preliminary estimation of the three links and the estimation results. The second is the transmission scheme of sending the signal in the auxiliary transceiver to the base station. The last is the decoupling of the estimated channel state information matrix by the base station and the estimation of the matrix of each segmented reflection channel.

[0076]

Preliminary estimation of three links

[0077] According to the preliminary training method mentioned in 3, the preliminary channel state information estimation of links 1, 2 and 3 is carried out. The corresponding relationship between the pilot signal of link 1 and the intelligent reconfigurable surface configuration matrix is shown in Figure 3 , the dimension reconstruction cascaded channel model is shown in formula (1), the corresponding relationship between the pilot signal of links 2 and 3 and the intelligent reconfigurable surface configuration matrix is shown in Figure 4 , and the dimension reconstruction cascaded channel model is:

[0078] Link 2: Where H config is the same as in formula (1), and J Tconfig =[Ψ0J T ,Ψ1J T,…, Ψ N′-1 J T ], (N a is the number of antennas of the secondary transceiver);

[0079] Link 3: where J config = ([(JΨ0) T ,(JΨ1) T ,…,(JΨ N′-1 ) T ] T And G config is the same as in formula (1), it can be proved that J config and J Tconfig are transposes of each other.

[0080] Finally, the partial channel state information of links 1 and 2 is estimated at the base station: (Link 1) the phase-corrected dimensionally reconstructed H channel direction matrix: the dimensionally reconstructed G channel direction matrix: the dimensionally reconstructed H and G channel coupling gain matrix: (Link 2) the phase-corrected dimensionally reconstructed H channel direction matrix: the dimensionally reconstructed J T channel direction matrix: the dimensionally reconstructed H and J T channel coupling gain matrix In the secondary transceiver, the information related to link 3 is obtained: the phase-corrected dimensionally reconstructed J channel direction matrix: the dimensionally reconstructed G channel direction matrix the dimensionally reconstructed J and G channel coupling gain matrix: It is worth noting that the dimensionally reconstructed channel state information matrices are all N s ×N s matrices.

[0081]

Transmission scheme from the secondary transceiver to the base station

[0082] Since the base station does not obtain the information related to link 3 at this time, it cannot decouple the gain matrix, and the transmission scheme is designed to transmit the coupling gain matrix of link 3 to the base station. Considering the dimension restriction on the to-be-transmitted signal matrix in the channel model of link 2, it is necessary to convert into a series of N a ×N a submatrices E0, E1, …, E N′-1 , and the splitting diagram of the matrix and the correspondence diagram of the submatrix and the intelligent reconfigurable surface configuration during transmission are shown in Figure 5The received signal can be represented as a matrix and decomposed into a dimension-reconstructed channel model as follows:

[0083]

[0084] The estimation of the sub-matrix is calculated as (3) and the gain matrix of link 3 is recovered according to Figure 5

[0085]

[0086]

Channel state information decoupling and final estimation at the base station

[0087] Considering that the base station has obtained the directional matrix of each of the three segmented channels H, G, and J, the decoupling is mainly directed to the coupled gain matrix: According to the expansion of the three, the following equation group is obtained:

[0088]

[0089] The gain matrix of each channel and the estimated reconstructed channel are obtained by solving:

[0090]

[0091] The matrix W in the above is a random disturbance term, and its form is relatively complex and is not further researched in the present scheme.

[0092] Taking the first N r rows of , the first N s rows of , and the first N a rows of , the final estimated channel matrix

[0093] Embodiment:

[0094] After introducing the auxiliary transceiver, a three-point cooperative intelligent reconfigurable surface communication model is constructed, the preliminary channel training method of a single link is applied to each of the three links, and the base station and the auxiliary transceiver obtain part of the channel state information of link 1, link 2, and link 3, respectively. The sending of the pilot signal in the channel estimation stage and the matching intelligent reconfigurable surface configuration matrix during transmission are shown in Figure 2 and Figure 3 , and the base station and the auxiliary transceiver recover the gain matrix of link 3 according to Figure 3 ​The diagram in the figure organizes this series of received signals into the form of the dimension-reconstructed concatenated channel model shown in formula (1), and then estimates the partial channel state information of the dimension-reconstructed channel according to the following algorithm (taking link 1 as an example):

[0095] [Estimation Algorithm for Dimensional Reconstruction Cascaded Channels]

[0096] Input: Initial configuration matrices Φ0 and Φ1 of the intelligent reconfigurable surface; pilot signal A series of auxiliary configuration matrices [Ψ0,Ψ1,…,Ψ] constructed by the bisection reconstruction method L (L≥M, L≥N).

[0097] Initialize variables:

[0098] Step 1: Obtain the received signal matrices Y0 and Y1 according to the single-link transmission principle and formula (1);

[0099] Step 2: Estimate intermediate variables using the least squares method

[0100] Step 3: Calculate intermediate variables Eigenvalues:

[0101] Step 4: (Loop) The variable i is 1, 2, ..., N in sequence. s The process in each loop is as follows:

[0102] Calculate intermediate variables

[0103] For D i Perform singular value decomposition: And Divided into a series of column vectors:

[0104] In initializing variables

[0105] Perform phase calibration on the direction matrix:

[0106] Step 5: Calculate the matrix

[0107] Step 6: Decompose Q1 into a series of column vectors:

[0108] Step 7: (Loop) The variable i is 1, 2, ..., N in sequence. s The process in each loop is as follows:

[0109] Calculate initialization variables

[0110] Step 8: Calculate the coupling gain matrix:

[0111] Output: Obtain the dimension reconstruction channel H config and G config Direction matrix and coupling gain matrix: and

[0112] [Partial channel state information estimated for the three links]

[0113] according to Figure 3 Figure 4 The corresponding relationship is transmitted through the pilot channel, and after the above algorithm is applied to the base station and auxiliary transceiver to initially estimate the channel, the following partial channel state information is obtained:

[0114] In the base station:

[0115] (Link 1) Phase calibration followed by dimension reconstruction of the H-channel direction matrix: Dimensionally reconstructed G-channel direction matrix: Dimensionally reconstructed H and G channel coupling gain matrices:

[0116] (Link 2) H-channel direction matrix after phase calibration and dimension reconstruction Dimensional Reconstruction of J T Channel direction matrix: Dimensional reconstruction of H and J T Channel coupling gain matrix:

[0117] In the auxiliary sending and receiving side:

[0118] (Link 3): J-channel direction matrix reconstructed after phase calibration: Dimensionally reconstructed G-channel direction matrix: Dimensionally reconstructed J and G channel coupling gain matrices: In the above, W represents the disturbance term.

[0119] The signal transmission scheme from the transceiver to the base station is as follows: To determine the coupling gain matrix during decoupling estimation, the coupling gain matrix of link 3 needs to be transmitted to the base station and combined with the coupling gain matrices of links 1 and 2 already obtained by the base station to construct a ternary equation system for decoupling. The transmission scheme proceeds in the following three steps:

[0120] Link 3 Coupling Gain Matrix Dimension Adjustment Method: The coupling gain matrix dimension adjustment method is... Figure 5The matrix partitioning diagram in the figure shows N. s ×N s The link 3 coupling gain matrix is ​​split into a series of N a ×N a Submatrices E0, E1, ..., E N′-1 ;

[0121] Signal transmission scheme conforming to the two-dimensional reconstruction channel transmission law of the link: according to Figure 5 The corresponding relationship diagram in the diagram is used to transmit sub-matrix signals to the base station. After transmission is completed, the base station processes and integrates the received signals according to formula (2).

[0122] Link 3 coupling gain matrix restoration method at the base station: based on the integrated received signal Y 2E And the estimates made during the initial training The estimated formula for the sent subset of matrices is obtained.

[0123]

[0124] After obtaining the submatrix set, it can be directly based on... Figure 5 The split diagram is combined to form the coupling gain matrix of the original link 3.

[0125] Decoupling of the coupling gain matrix and estimation of H, G, J of the channel matrix: After obtaining partial channel state information of links 1 and 2 and reconstructing the coupling gain matrix of link 3, the base station can construct a ternary equation system as shown in equation (4) and obtain the decoupled gain matrix and the reconstructed channel matrix of each dimension in equation (5). Taking dimension H to reconstruct the channel. The first N r Okay, dimension reconstruction G channel The first N s Okay, dimension reconstruction J channel The first N a Linear estimation of each channel matrix

[0126] The simulation results of this invention are shown in Figure 6 and Figure 7 . Figure 6 The parameter is set to: N s =N t =N a =N r =16, Figure 7 The parameter is set to: N t =N a =N r =16, P=144. The estimation accuracy is represented by the normalized mean square error, and its formula is:

[0127] Channel H:

[0128] Channel G:

[0129] Channel J:

[0130] from Figure 6 It can be observed that the pilot with time slot 124 has the smallest estimation error, and the estimation error gradually decreases as the signal-to-noise ratio increases. From... Figure 7 It can be observed that as the number of intelligent reconfigurable surface units increases in this scheme, the estimation error shows a decreasing trend, and it can maintain high estimation accuracy in ultra-large-scale intelligent reconfigurable surface cascaded channels.

[0131] In another embodiment of the present invention, a smart reconfigurable surface cascaded channel estimation system based on dimension reconstruction is provided, which can be used to implement the above-mentioned smart reconfigurable surface cascaded channel estimation method based on dimension reconstruction. Specifically, the system includes:

[0132] The model building module is used to build a three-point collaborative intelligent reconfigurable surface communication model containing three communication links;

[0133] The auxiliary configuration module is used to introduce an auxiliary configuration matrix into the three-point collaborative intelligent reconfigurable surface communication model, simplifying the configuration process of the intelligent reconfigurable surface.

[0134] The training module is used to perform preliminary channel training for a single link to obtain a cascaded channel model with reconstructed dimensions.

[0135] The estimation output module is used to make a preliminary estimate of the three links through the cascaded channel model of dimension reconstruction, and send the signal in the auxiliary transceiver to the base station. The base station decouples the estimated channel state information matrix and estimates the matrix of each segment of the reflection channel.

[0136] The module division in this embodiment of the invention is illustrative and represents only one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional modules in the various embodiments of the invention can be integrated into a single processor, exist as separate physical entities, or be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0137] In another embodiment of the present invention, a computer device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions from the computer storage medium to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used for the operation of an intelligent reconfigurable surface cascaded channel estimation method based on dimension reconstruction.

[0138] In another embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the intelligent reconfigurable surface cascaded channel estimation method based on dimensional reconstruction in the above embodiments.

[0139] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0140] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations 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, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0141] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0142] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 Figure 1 The steps of the function specified in one or more boxes.

[0143] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for estimating intelligent reconfigurable surface cascaded channels based on dimensional reconstruction, characterized in that, Includes the following steps: Construct a three-point collaborative intelligent reconfigurable surface communication model containing three communication links; An auxiliary configuration matrix is ​​introduced into the three-point collaborative intelligent reconfigurable surface communication model to simplify the configuration process of the intelligent reconfigurable surface; Preliminary channel training is performed on a single link to obtain a cascaded channel model with reconstructed dimensions; The three links are initially estimated using a cascaded channel model with dimension reconstruction. The signals in the auxiliary transceiver are sent to the base station. The base station decouples the estimated channel state information matrix and estimates the matrix of each segment of the reflection channel. Preliminary channel training method for a single link: The preliminary training of the model channel includes the transmission of multiple pilot signals, configuration transformation of the intelligent reconfigurable surface, and preliminary acquisition of channel state information on the base station side; The transmitted pilot signal needs to pass through multiple intelligent reconfigurable surfaces with different configurations and obtain a series of received signals at the base station; the original received signals are integrated into a dimensionally reconstructed received signal matrix, which is then decomposed to obtain a dimensionally reconstructed cascaded channel model: in It is the transmitted pilot signal. Quantity, and These are the front and rear auxiliary matrices, respectively. This is called dimensional reconstruction. Channel, This is called dimensional reconstruction. Channel; in the estimation, the channel matrix and the dimensionally reconstructed channel matrix are split into magnitude matrices, which are... The phase matrix is ​​denoted as The direction matrix is ​​denoted as That is, the channel matrix is ​​denoted as The following breakdown is made: (1) Launcher – Intelligent Reconfigurable Surface: (2) Intelligent reconfigurable surface – receiver: The direction matrix is ​​phase-calibrated using an optimization method so that phase and direction information are estimated together; the final channel state information estimation matrix obtained after preliminary training is a summary of the channel state information. and Dimensional reconstruction direction matrix , And the coupling gain matrix: product The estimated matrix.

2. The intelligent reconfigurable surface cascaded channel estimation method based on dimension reconstruction according to claim 1, characterized in that, The three communication links are: Link 1: User—Smart Reconfigurable Surface—Base Station; Link 2: Auxiliary Transceiver—Smart Reconfigurable Surface—Base Station; Link 3: User—Smart Reconfigurable Surface—Auxiliary Transceiver; The channel models for the three links are as follows: , , , where the matrix For noise, and These are the received signal and the transmitted signal, respectively. The phase configuration matrix for the intelligent reconfigurable surface is a diagonal matrix; the relevant matrix dimensions are: , , , , , among them The number of antennas for the sending user. The number of antennas at the receiving base station. To assist in the number of antennas for both the transmitter and receiver, The number of arrays of intelligent reconfigurable surfaces. The number of time slots for the pilot signal; in this model, the auxiliary transceiver mainly plays the role of receiving the pilot signal of link three and performing preliminary channel estimation. At the same time, in link two, as a transmitting device, the auxiliary transceiver will send the estimated channel state information to the base station to facilitate the base station's channel estimation of this link.

3. The intelligent reconfigurable surface cascaded channel estimation method based on dimension reconstruction according to claim 1, characterized in that, An auxiliary configuration matrix was introduced. and ,in For the pre-position auxiliary matrix, This is the post-auxiliary matrix; after introducing the auxiliary matrix, the total configuration matrix of the intelligent reconfigurable surface is: .

4. The intelligent reconfigurable surface cascaded channel estimation method based on dimension reconstruction according to claim 2, characterized in that, Preliminary estimates of the three links Based on the preliminary training method, preliminary channel state information estimation is performed on the three links, and the cascaded channel model with dimension reconstruction is as follows: Link 2: , , , To assist in determining the number of antennas for both the transceiver and receiver; Link 3: ,in Proof obtained and They are transposes of each other; The cascaded channel model for link 1 dimension reconstruction is shown in formula (1). Finally, partial channel state information for links 1 and 2 was estimated at the base station: the dimension reconstruction after phase calibration of link 1. Channel direction matrix: Dimensional reconstruction Channel direction matrix: Dimensional reconstruction and Channel coupling gain matrix: ; Dimensional reconstruction after link 2 phase calibration Channel direction matrix: Dimensional reconstruction Channel direction matrix: Dimensional reconstruction and Channel Coupling Gain Matrix The auxiliary transceiver obtains link 3 related information: dimension reconstruction after phase calibration. Channel direction matrix: Dimensional reconstruction Channel direction matrix Dimensional reconstruction and Channel coupling gain matrix: .

5. The intelligent reconfigurable surface cascaded channel estimation method based on dimension reconstruction according to claim 4, characterized in that, Assisting in the transmission from the transceiver to the base station: The coupling gain matrix of link 3 Transmitted to the base station; Convert to a series submatrix The received signal, after being processed, is represented as the following matrix and decomposed into a dimension reconstructed channel model: Based on the channel state information already acquired by the base station, the estimation formula (3) for the submatrix is ​​obtained, and the gain matrix of link 3 is recovered. ; 。 6. The intelligent reconfigurable surface cascaded channel estimation method based on dimension reconstruction according to claim 5, characterized in that, Decoupling and final estimation of channel state information on the base station side: The base station obtains three segmented channels , , The respective direction matrices, and the decoupling gain matrix relative to the coupling: , , Based on the expansions of the three equations, we obtain the following system of equations: The gain matrices of each channel and the estimated reconstructed channel are obtained by solving the problem. The matrix mentioned above For random disturbance terms; Pick The former OK, The former OK, The former The final estimated channel matrix is ​​obtained by traversing the rows. , , .

7. A smart reconfigurable surface cascaded channel estimation system based on dimension reconstruction, characterized in that, include: The model building module is used to build a three-point collaborative intelligent reconfigurable surface communication model containing three communication links; An auxiliary configuration module is used to introduce an auxiliary configuration matrix into the three-point collaborative intelligent reconfigurable surface communication model, simplifying the configuration process of the intelligent reconfigurable surface. The training module is used to perform preliminary channel training for a single link to obtain a cascaded channel model with reconstructed dimensions. The estimation output module is used to make a preliminary estimate of the three links through the cascaded channel model of dimension reconstruction, and send the signal in the auxiliary transceiver to the base station. The base station decouples the estimated channel state information matrix and estimates the matrix of each segment of the reflection channel. Preliminary channel training method for a single link: The preliminary training of the model channel includes the transmission of multiple pilot signals, configuration transformation of the intelligent reconfigurable surface, and preliminary acquisition of channel state information on the base station side; The transmitted pilot signal needs to pass through multiple intelligent reconfigurable surfaces with different configurations and obtain a series of received signals at the base station; the original received signals are integrated into a dimensionally reconstructed received signal matrix, which is then decomposed to obtain a dimensionally reconstructed cascaded channel model: in It is the transmitted pilot signal. Quantity, and These are the front and rear auxiliary matrices, respectively. This is called dimensional reconstruction. Channel, This is called dimensional reconstruction. Channel; in the estimation, the channel matrix and the dimensionally reconstructed channel matrix are split into magnitude matrices, which are... The phase matrix is ​​denoted as The direction matrix is ​​denoted as That is, the channel matrix is ​​denoted as The following breakdown is made: (1) Launcher – Intelligent Reconfigurable Surface: (2) Intelligent reconfigurable surface – receiver: The direction matrix is ​​phase-calibrated using an optimization method so that phase and direction information are estimated together; the final channel state information estimation matrix obtained after preliminary training is a summary of the channel state information. and Dimensional reconstruction direction matrix , And the coupling gain matrix: product The estimated matrix.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent reconfigurable surface cascaded channel estimation method based on dimension reconstruction as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the intelligent reconfigurable surface cascaded channel estimation method based on dimension reconstruction as described in any one of claims 1 to 6.

Images