RIS-assisted MIMO semi-blind joint channel estimation method

Abstract

The invention provides an RIS-assisted semi-blind joint channel estimation and symbol detection method for estimating a four-dimensional data signal. The method comprises the following steps: firstly, constructing an RIS-assisted MIMO communication system model, and converting a signal received by a base station end into a tensor form; the estimation process is divided into the following two stages: in a direct link stage, a KRF algorithm is utilized to convert a channel estimation problem into an approximate value approximation problem, and a high-quality approximation result is obtained by means of a left and right dominant singular value comparison process; and entering an RIS auxiliary link stage, further converting the problem into an iterative optimization problem of a cost function according to PARAFAC nested decomposition and a TALS algorithm principle, and obtaining a more accurate threshold result along with the increase of the number of iterations. According to the method and the device, the problem that the channel state information is difficult to quickly and accurately acquire due to the fact that the dimension of the received signal is too high in an RIS-assisted communication scene is successfully solved. Compared with an ALS algorithm based on pilot assistance, a TALS semi-blind receiver and a KAKF semi-blind receiver which are widely applied in the same communication scene, the method disclosed by the invention shows remarkable advantages in the aspects of improving the estimation precision, reducing the bit error rate and reducing the pilot overhead.

Description

Technical Field

[0001] The present invention belongs to the field of wireless communication technology, and particularly relates to a RIS-assisted MIMO semi-blind joint channel estimation system. Background Art

[0002] Reconfigurable intelligent surface (RIS) is considered as one of the promising technologies for future sixth-generation (6G) wireless communication due to its characteristics such as low cost, low power consumption, simple deployment, and the ability to configure the wireless propagation environment to weaken interference in the channel estimation process. RIS consists of a planar array of multiple passive (or semi-passive) controllable reflecting elements. According to its internal structure and composition, it mainly realizes the control of the signal propagation environment by intelligently controlling the reflection, refraction, and scattering characteristics of electromagnetic waves. Each reflecting element can independently adjust its phase and amplitude according to the dynamic needs of users. Under the condition that the channel state information (CSI) is known, RIS can adjust and change the reflection coefficients of different reflecting elements in real time according to the CSI, so as to effectively increase the useful signal and reduce the interference signal at the receiver. Although RIS can bring certain gains to the wireless communication system, completely passive RIS does not have the ability to independently process signals. How to quickly and accurately obtain CSI in a multiple-input multiple-output (MIMO) system assisted by RIS is the key issue in research.

[0003] In recent years, some studies have begun to use tensor-based methods to solve the CE (channel estimation) problem in RIS-assisted communication systems, and some classic algorithms have emerged. For example, L. Wei et al. proposed a tensor-based alternating least squares (ALS) CE algorithm. It uses the principle of tensor decomposition to process high-dimensional data. However, the increase in the number of estimated coefficients leads to a sharp increase in the demand for pilots, and the excessive pilot overhead significantly reduces its spectral efficiency. To balance the pilot overhead and CE accuracy, recent research has made achievements in semi-blind estimation. G. T. De Araujo et al. proposed a tensor-based two-stage semi-blind KAKF receiver, which jointly estimates the channel and the symbol matrix and removes the pilot training stage. However, there is still an error between the estimations of the two stages. Subsequently, this team proposed a tensor-based trilinear alternating least squares (TALS) CE algorithm. On the basis of removing the pilot training stage, through iteration, the error caused by the stage is removed. However, there are still errors that cannot be eliminated and cannot be ignored in the three-step estimation process.

[0004] The inventors found that the high-dimensional data estimated by the above solutions are all three-dimensional data, and there are new challenges in data processing and channel estimation for dimensions exceeding three. Since the scenarios constructed by signal information exceeding three dimensions are more complex, the number of parameters involved in the system is larger and the composition is complex. If general methods are used to process and estimate high-dimensional information data, certain requirements for the computing power of the algorithm are needed, which is time-consuming and laborious; at the same time, if a pilot-assisted method is used for channel estimation, a huge pilot overhead will be generated, resulting in a reduction in spectral efficiency. Therefore, how to solve the CE problem of high-dimensional signals and data above three dimensions in the system and accurately and quickly obtain CSI has become an urgent issue to be solved currently. Summary of the Invention

[0005] An embodiment of the present invention provides a RIS-assisted semi-blind joint channel estimation and symbol detection method to supplement the lack of existing technologies in dealing with possible four-dimensional signal information in RIS assistance.

[0006] According to an embodiment of the present invention, a RIS-assisted semi-blind joint channel estimation and symbol detection method is provided, including:

[0007] Build a RIS-assisted MIMO communication system model; where the time is divided into three units: time unit, time sub-block, and time slot according to preset conditions;

[0008] Based on the built RIS-assisted MIMO communication system model, construct the noise-free received signal of the base station as a parallel factor tensor model; use the nested PARAFAC decomposition method to decompose the four-dimensional signal into two three-way signals; PARAFAC decomposition, as a classical three-dimensional tensor decomposition method, usually cuts along three directions: horizontal, vertical, and transverse, to obtain three different product forms of a matrix Khatri-Rao product matrix and a single matrix; based on the construction of the decomposition result, an abstract and complex three-dimensional problem can be transformed into three specific problems for decomposition and estimation.

[0009] Subsequently, according to the decomposition characteristics, we use the trilinear alternating least squares algorithm; set three parameters to enter the alternating optimization process, fix the other two variables, optimize the remaining variable, and alternately optimize and iterate the three cost functions to obtain the estimated value of the final channel matrix, realizing semi-blind joint channel estimation and symbol detection of four-dimensional signal information.

[0010] At the same time, the present invention embodiment does not ignore the direct link during the channel estimation process, and uses the obtained symbol matrix data as the initial value of the RIS-assisted link, and uses the KRF algorithm to estimate the direct link; singular value decomposition is the core of the algorithm, and approximate estimation is achieved by redistributing the energy of the original space, usually transforming the channel estimation problem into sub-problems of approximating N rank-one matrices.

[0011] Furthermore, the mathematical model for the base station to receive signals is specifically expressed as follows:

[0012]

[0013] In the channel estimation stage of the direct link, the RIS switch is in the "off" state, and the received signals are collected as:

[0014]

[0015] Among them, the KRF (Khatri-Rao) algorithm is used to estimate two parameters of the channel matrix and the transmission symbol matrix:

[0016] First, define as:

[0017]

[0018] Furthermore, we collect the signals received in the first time sub-blocks to obtain:

[0019]

[0020] Furthermore, through derivation and transformation, we get:

[0021]

[0022] Furthermore, the Khatri-Rao decomposition algorithm is used to estimate the and values:

[0023]

[0024] Through the above formula, using the principle of the KRF algorithm, the channel estimation problem is converted into an approximation problem of a rank-1 matrix:

[0025]

[0026] Among them, the symbol represents the Khatri-Rao operation. The KRF algorithm mainly uses the principle of singular value decomposition (SVD) to set a threshold range for approximation estimation.

[0027] In the second stage, i.e., the RIS-assisted link stage, the RIS switch is in the "on" state. At this time, data transportation and signal estimation are carried out simultaneously on the direct link and the RIS-assisted link. In the second stage, the small influence brought by the direct link is ignored in the example of the present invention, and the received signals are obtained:

[0028]

[0029] Among them, this invention patent utilizes an innovative algorithm to estimate the cascaded estimation of two channel matrices and a symbol matrix, and obtains their respective estimated values through a decomposition process.

[0030] Obtain a four-dimensional vector through nested PARAFAC decomposition , represented in scalar form as:

[0031]

[0032] Furthermore, the four-dimensional vector is decomposed into two three-dimensional vectors:

[0033]

[0034]

[0035] Furthermore, according to the contraction form of, obtain their respective decomposition representations:

[0036]

[0037]

[0038] Among them, .

[0039] Furthermore, according to different decomposition forms of the three-dimensional tensor, some of its modes are obtained as follows:

[0040]

[0041]

[0042]

[0043] Furthermore, through the deformation and conversion of the above expressions, obtain the channel estimation expressions for the matrices , and :

[0044]

[0045]

[0046]

[0047] Among them, is the representation symbol of the pseudo-inverse, represents the Kronecker operation.

[0048] The above one or more technical solutions have the following beneficial effects:

[0049] (1) The present invention provides a RIS-assisted semi-blind joint channel estimation method. By establishing a RIS-assisted MIMO communication system model, the signal received by the base station is modeled as a parallel factor tensor model. Based on the tensor decomposition principle of PARAFAC, the four-dimensional tensor is decomposed into two three-dimensional vectors. Further, in the auxiliary link stage, the channel estimation problem is transformed into an iterative optimization problem of a cost function, and the cost function is solved by trilinear least squares (TALS), realizing the RIS-assisted semi-blind joint channel estimation and symbol detection of four-dimensional data signals. The proposed scheme solves the problem that in the RIS-assisted communication scenario, due to time nesting or other factors, the dimension of the received signal is too high, making it difficult to accurately and quickly obtain the channel state information. Compared with the pilot-assisted ALS algorithm, TALS semi-blind receiver, and KAKF semi-blind receiver in the same assisted communication scenario, this method can effectively improve the estimation accuracy and further reduce the bit error rate and pilot overhead.

[0050] (2) The proposed scheme of the present invention considers the direct link stage when RIS is in the off state. Based on the KRF algorithm, the signal estimation problem is transformed into an approximation convergence problem, and the transmission symbol matrix result obtained in the first-stage direct link is used as the initial value and brought into the second-stage RIS-assisted link. At this time, due to the substitution of the initial value, the estimation accuracy is greatly improved and the bit error rate is reduced.

[0051] (3) The proposed scheme of the present invention is based on the principles of the ALS algorithm and tensor decomposition to solve the channel estimation problem of the RIS-assisted system for four-dimensional data signals. The simulation results show that under the same signal-to-noise ratio, the proposed semi-blind joint channel estimation algorithm is superior to the pilot-assisted ALS scheme, TALS semi-blind receiver scheme, and KAKF semi-blind receiver scheme in the same scenario, effectively improving the estimation accuracy and reducing the pilot overhead at the same time.

[0052] The advantages of the additional aspects of the present invention will be given in the following description section, some of which will become more obvious and intuitive from the following description, or can be understood through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The accompanying drawings forming a part of this specification are used to assist in the further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention.

[0054] Figure 1 It is a block diagram of the RIS-assisted MIMO communication system model described in the embodiments of the present invention;

[0055] Figure 2Flowchart and technical route of the RIS-assisted semi-blind joint channel estimation and symbol detection method described in the embodiments of the present invention

[0056] Figure 3 It is a legend of channel protocol estimation in the embodiments of the present invention, which intuitively expresses the time allocation during the whole process of channel estimation;

[0057] Figure 4 It is a schematic diagram of time division in any stage in the embodiments of the present invention;

[0058] Figure 5 It is a comparison diagram of simulation curves of the cascaded channel normalized mean square error (NMSE) versus signal-to-noise ratio (SNR) of the single-RIS-assisted semi-blind joint channel estimation and symbol detection method described in the embodiments of the present invention, the pilot-assisted ALS algorithm, the KAKF semi-blind receiver, and the TALS algorithm in the same communication scenario;

[0059] Figure 6 It is a comparison diagram of simulation curves of the symbol error rate (SER) versus signal-to-noise ratio (SNR) of the single-RIS-assisted semi-blind joint channel estimation and symbol detection method described in the embodiments of the present invention, the pilot-assisted ALS algorithm, and the KRKF semi-blind estimation algorithm in the same communication scenario. Detailed implementation manners

[0060] It should be noted that the following detailed descriptions are all exemplary and are intended to provide further descriptions of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0061] It should be noted that the terms used herein are only for describing specific implementation manners and are not intended to limit the exemplary embodiments according to the present invention.

[0062] Without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0063] As Figure 2 shown, the embodiments of the present invention provide a RIS-assisted semi-blind joint channel estimation method, which specifically includes the following processing procedures:

[0064] Step 1: Construct a RIS-assisted MIMO communication system model; wherein, the preset conditions of the time unit are divided into three levels;

[0065] In specific implementation, the communication system adopted is a RIS-assisted end-to-end MIMO system. Considering the uplink input path, the initial parameters are set as (in specific implementation, the following parameters can be set according to actual requirements): the number of base station transmitting antennas is 5, the number of time units is 6, the number of reflecting elements per RIS is 4, the number of user antennas is 4, the transmission time blocks within a transmission frame is 8, is 4, each transmission time block contains 8 time slots, the transmitted data symbols are selected from a 16-PSK alphabet, and the coding matrix and the RIS phase matrix are taken from and the first rows and rows of the discrete Fourier transform matrix, and the system block diagram is as Figure 1 shown.

[0066] Step 2: Based on the constructed RIS-assisted MIMO communication system model, preset a transmission protocol, assuming that the model is a quasi-static flat fading channel;

[0067] Among them, as Figure 3 shown, the whole process of channel estimation transmission protocol is presented. A transmission frame consists of time slots, is denoted as the data transmission phase, is denoted as the semi-blind estimation phase, and the preset coherence time is long enough to fully ensure that the overall transmission time is not interfered. During the phase, it is divided into and according to the direct link and the RIS-assisted link.

[0068] Among them, during the direct link phase, is divided into time units, each time unit contains time sub-blocks, and each time sub-block contains time slots; during the RIS-assisted link phase, is divided into time units, each time unit contains time sub-blocks, and each time sub-block contains time slots.

[0069] Among them, as Figure 4 shown, the time division diagram within the phase is presented. Obtain . Considering the assistance of the RIS, the base station receives the independent data streams encoded by the mobile user; considering the direct link estimation result as the initialization of the RIS-assisted link, that is, turning off the RIS during the time and turning on the RIS during the time. Considering the base station at the th unit and the The discrete-time signal vector received in the th time slot of a sub-block can be expressed as (ignoring the minor direct-link impact during the RIS on phase):

[0070]

[0071] Step 3: In the first time units, the RIS is in the "off" state. Using the Khatri-Rao (KRF) algorithm, the estimation problem of the channel matrix between the user terminal (UE) - base station (BS) and the transmission symbol matrix ( ) is converted into an approximation problem.

[0072] In specific implementation, Step 3 specifically includes the following processing procedures:

[0073] Step 301: The discrete-time signal vector received by the base station in the th unit and the th time sub-block and the th time slot can be expressed as:

[0074]

[0075] where the UT-BS direct-link channel , the mobile user terminal uses a space-time coding scheme to encode "diagonally", , and the additive white Gaussian noise vector .

[0076] Assume that the coding vector remains unchanged in the th time sub-block and only changes between sub-blocks, and is regarded as a known quantity during the channel estimation process, that is:

[0077]

[0078] Step 302: According to Step 2 and Step 301, define and transform to obtain:

[0079]

[0080] Step 303: According to the transformation, mathematical derivation, and principle, collect the signals of the first time sub-blocks to obtain:

[0081] where , and through transformation, we can get:

[0082]

[0083] is the noise version of the virtual MIMO channel (with Khatri-Rao structure) that models RIS-aided MIMO transmission.

[0084] Step 304: Use the Khatri-Rao least squares approximation problem to and estimate the values:

[0085]

[0086] Therefore, the channel estimation problem is transformed into an approximation problem of a rank-1 matrix:

[0087]

[0088] where , is expressed as the first column of the and respectively represent the first column of the matrix and and The estimates of can be obtained from the left and right dominant singular vectors of . The symbol represents the Khatri-Rao operation.

[0089] Step 4: In the subsequent time units, the RIS switch is in the "on" state. Using the innovative algorithm (E-NP) of this embodiment, the estimation problems of the user equipment (UE)-RIS, RIS-base station (BS), and the transmission symbol matrix ( ) are transformed into cost function iterative optimization problems. Using the trilinear alternating least squares (TALS) algorithm, fix the cost function and implement the update iteration of the remaining parameters. Repeat this process until it reaches below the threshold range.

[0090] In a specific implementation, Step 4 specifically includes the following processing procedures:

[0091] Step 401: The discrete-time signal vector received by the base station in the th unit, the th time sub-block, and the th time slot can be expressed as:

[0092]

[0093] where the RIS-BS channel , the UT-RIS channel , the RIS phase vector , , the coded symbol vector is transmitted in the th sub-block at the th time slot. Additive white Gaussian noise .

[0094] Assume that the phase vector remains unchanged in the th time unit and the th sub-block, varies between units, and the coding vector remains unchanged in the th time sub-block at the th time slot, only varies between sub-blocks, and is a known quantity during the channel estimation process, i.e.:

[0095]

[0096] Step 402: According to the principle of PARAFAC tensor decomposition, through the set model, obtain the four-dimensional vector , and the scalar form is:

[0097]

[0098] Step 403: Set as a three-dimensional vector, by concatenating the th sub-block measurement matrix , and expanding it along the second dimension, we get:

[0099]

[0100] where, represents the concatenation along the th dimension. Further, let be obtained by concatenating the measurement tensor along the fourth dimension:

[0101]

[0102] Step 404: According to the principle of tensor decomposition, the four-dimensional vector can be decomposed into two three-dimensional vectors for further decomposition processing. Further, let , considering the above four-dimensional vector can be represented as two correlated and three-dimensional vectors:

[0103]

[0104]

[0105] Step 405: Further, the four-dimensional tensor of modulus expansion, , two contraction forms of and are respectively expressed in the nested PARAFAC model as:

[0106]

[0107]

[0108] wherein, .

[0109] Step 406: Further, according to the three-dimensional tensor decomposition, it is divided into three different forms: horizontal, vertical, and transverse. According to different decomposition forms, some of the modes are obtained as follows:

[0110]

[0111]

[0112]

[0113] Step 407: Further, based on the expression deformation conversion in Step 406, the channel estimation expressions for matrices , are obtained (the estimated value of the matrix in the direct link is substituted into the RIS-assisted link as the initialization result):

[0114]

[0115]

[0116] To estimate matrix , using the Kronecker product property, that is, we get:

[0117]

[0118] It can be transformed to:

[0119]

[0120] This embodiment proposes a nested PARAFAC-based TALS (NP-TALS) receiver. The normalized mean square error of the th iteration is . When the condition is satisfied Iteration stops when

[0121] Furthermore, to verify the effectiveness of the solution described in this embodiment, the following related experimental verifications were carried out:

[0122] As Figure 5 shown, the comparison diagram of the simulation curves of the RIS-assisted semi-blind joint channel estimation and symbol detection method described in this implementation example with the pilot-assisted ALS algorithm, TALS semi-blind receiver, and KAKF semi-blind receiver in the same scenario regarding the normalized mean square error (NMSE) of the cascaded channel varying with the signal-to-noise ratio (SNR). It can be seen from the figure that under the condition that the time allocation method is different and the remaining parameters are the same, compared with the comparison scheme, the scheme proposed by the present invention can not only estimate a channel matrix with a higher dimension, but also improve the estimation accuracy and reduce the consumption of pilots.

[0123] As Figure 6 shown, the comparison diagram of the simulation curves of the RIS-assisted semi-blind joint channel estimation and symbol detection method described in this embodiment with the pilot-assisted ALS algorithm and KRKF semi-blind receiver in the same scenario regarding the symbol error rate (SER) varying with the signal-to-noise ratio (SNR). It can be seen from the figure that under the condition that the time allocation method is different and the remaining parameters are the same, the scheme proposed by the present invention is still superior to the comparison scheme when the dimension of the estimated signal data is higher. The scheme proposed by the present invention can greatly reduce the symbol error rate of the signal.

[0124] In one or more embodiments, corresponding to the above-mentioned RIS-assisted semi-blind joint channel estimation method, a RIS-assisted semi-blind joint channel estimation system is provided, including:

[0125] A system model construction unit for constructing a RIS-assisted MIMO communication system model; wherein, the number of time units is divided according to preset conditions;

[0126] A symbol matrix estimation unit for constructing the noise-free received signal of the base station as a parallel factor tensor model based on the constructed RIS-assisted MIMO communication system model; in the symbol matrix estimation of the direct link, using the KRF algorithm, transforming the channel estimation problem into an approximation approximation problem to obtain the symbol matrix estimation value and the channel matrix estimation value; in the RIS-assisted link stage, substituting the direct link estimated symbol matrix as the initial value into the second stage, and obtaining the symbol matrix estimation value and the coupled channel matrix estimation value through the balanced factorization theory;

[0127] The channel matrix estimation unit reconstructs the base station received signal based on the estimated value of the symbol matrix to obtain the reconstructed received signal; constructs the noise-free received signal after reconstruction into a parallel factor tensor model, and cuts the parallel factor tensor model along the horizontal, vertical, and transverse planes through balanced factorization, and constructs a first cost function, a second cost function, and a third cost function based on the cutting results; obtains the estimated value of the channel matrix by alternately optimizing the first cost function, the second cost function, and the third cost function, so as to realize RIS-assisted semi-blind joint channel estimation.

[0128] It can be understood that the system is consistent with the method in the above embodiment, and its specific details are described in detail in the method embodiment, so they will not be repeated here.

[0129] In more embodiments, there is also provided:

[0130] An electronic device includes a memory, a processor, and computer instructions stored on the memory and running on the processor. When the computer instructions are run by the processor, the method described in the above embodiment is completed. For the sake of brevity, it will not be repeated here.

[0131] It should be understood that in this embodiment, the processor may be a central processing unit CPU, and the processor may also be other general-purpose processors, digital signal processors DSP, application-specific integrated circuits ASIC, off-the-shelf programmable gate arrays FPGA, or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0132] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A part of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0133] A computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by the processor, the method described in the above embodiment is completed.

[0134] The method in the above embodiment can be directly embodied as being executed and completed by a hardware processor, or by a combination of hardware and software modules in the processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory, and the processor reads the information in the memory and combines its hardware to complete the steps of the above method. To avoid repetition, it will not be described in detail here.

[0135] Those of ordinary skill in the art will appreciate that the units and algorithm steps of the examples described in conjunction with this embodiment can be implemented by electronic hardware or a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. A professional technician can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this disclosure.

[0136] The foregoing is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure. For those skilled in the art, various modifications and variations can be made to the present disclosure. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

1. A RIS-assisted semi-blind joint channel estimation and symbol detection method, characterized in that: include: A RIS-assisted MIMO communication system model is constructed, wherein the time unit preset conditions are divided into three levels; Based on the constructed RIS-assisted MIMO communication system model, the noise-free received signal of the base station is constructed as a parallel factor tensor model; in the symbol matrix estimation of the direct link, the KRF algorithm is used to obtain the symbol matrix estimation value and the channel matrix estimation value; in the RIS-assisted link stage, the direct link estimation symbol matrix is ​​substituted as the initial value into the second stage, and the decomposition principle of PARAFAC and the balanced factor decomposition theory are used to obtain the symbol matrix estimation value and the coupling channel matrix estimation value.

2. A RIS-assisted semi-blind joint channel estimation and symbol detection method as claimed in claim 1, specifically comprising: Based on the constructed RIS-assisted MIMO communication system model (divided into two stages: direct link and RIS-assisted link), a mathematical model of the base station receiving signal is established; based on the mathematical model, through the division of time units, the influence of transmission changes within the time sub-unit on the channel estimation process is considered to realize the construction of a parallel factor tensor model.

3. A RIS-assisted semi-blind joint channel estimation and symbol detection method as claimed in claim 2, characterized in that: The mathematical model of the base station receiving signal is specifically expressed as follows: in, is the channel from RIS to base station (BS), is the channel from the user (UT) to the RIS, is the transmission symbol matrix, is the RIS phase shift matrix, , is the encoding matrix, is the additive white Gaussian noise matrix.

4. The mathematical model expression as claimed in claim 3, in the direct link stage, based on the principle and characteristics of the KRF algorithm, the approximate advantage value is obtained by using singular value decomposition to obtain the matrix of the user terminal-base station and symbol transmission matrix The estimated expression of is: in, , and Respectively The first column of the matrix and The first row of the matrix, symbol Expressed as Khatri-Rao operation.

5. The mathematical model expression as claimed in claim 3, in the RIS auxiliary link stage, based on the characteristics of nested decomposition, decomposes the four-dimensional signal into two mutually related three-dimensional signals as follows:

6. The decomposition expression as claimed in claim 5, based on the tensor decomposition characteristics, obtains different expressions for horizontal, vertical and lateral cutting: in, , Here n indicates the cutting method, and N indicates the Nth expression form under the nth cutting method.

7. The cutting expression as claimed in claim 6, based on the algorithm characteristics of TALS, uses alternating iterative optimization to construct the first, second and third cost functions, and obtains the matrix ,matrix and matrix The expression is: in, symbol represents the pseudo-inverse operation, Represents the Kronecker operation.

8. An electronic device comprising a memory, a processor and a computer program stored and running on the memory, characterized in that: When the processor executes the program, the RIS-assisted semi-blind joint channel estimation and symbol detection method as described in any one of claims 1 to 7 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, a multi-RIS-assisted semi-blind joint channel estimation and symbol detection method as described in any one of claims 1 to 7 is implemented.

Images