A near-field large-scale reconfigurable intelligent surface topology optimization and beam design method

Abstract

Reconfigurable intelligent surface (RIS) assisted wireless communication in near-field propagation suffers from array gain loss, which limits the improvement of system capacity. To further improve the system capacity of large-scale RIS (XL-RIS) assisted wireless communication in near-field and compensate for the severe path loss, an optimized deployment of XL-RIS assisted multi-user wireless communication system is proposed to further expand the near-field range and improve the system capacity. The joint optimization of XL-RIS topology and beam design is performed to minimize the user's expected beam gain error. To solve this non-convex problem, the adaptive tabu search algorithm and Majorization-Minimization (MM) algorithm are used to alternately optimize the transformed sub-problems. Especially, without increasing the hardware cost, the selection of element positions obtains additional spatial degrees of freedom, effectively improves the user and rate, and makes the beam pointing closer to the user's expected direction. The simulation results verify the effectiveness of the proposed method, which provides a new solution for the application of XL-RIS in near-field communication systems.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless communication, in particular to a near-field large-scale reconfigurable intelligent surface topology optimization and beam design method. BACKGROUND

[0002] Extremely large-scale array / surface (XL-array / surface) has become a promising technology to achieve ultra-high spectral efficiency and ultra-space resolution in the future sixth generation (6G) communication system. As wireless communication migrates to higher frequency bands, such as millimeter wave (mm Wave) and even terahertz (THz), the near-field range expands to tens of meters or even hundreds of meters, making the near-field characteristics more prominent. However, these technologies are faced with the problems of serious free space path loss and obstruction sensitivity, and it is still difficult to completely solve them by using massive multiple input multiple output (m-MIMO) technology.

[0003] Reconfigurable intelligent surface (RIS) can actively and intelligently control spatial electromagnetic waves through programming, thereby forming an electromagnetic environment with controllable amplitude, phase, polarization and frequency. RIS has the characteristics of low power consumption and low cost. In the RIS-assisted wireless communication system, deploying a large number of reflection units can compensate for the serious "multiplicative fading" effect in the cascaded channel, so extremely large-scale RIS (XL-RIS) is considered to be one of the future development directions of RIS. At the same time, with the increase in the number of reflection units, the range of the near-field region will also expand, and scatterers and users are more likely to be located in the near-field region.

[0004] In the existing disclosed technical content, the research mainly focuses on codebook design and beam training. The system can select the beam pointing to the desired spatial angle through beam training from the pre-designed codebook. The existing beam training scheme is affected by the constant modulus constraint, which causes the effective direction of the beam to deviate from the user's desired area, and there is still a large space for improvement between the user's achievable transmission rate and the actual performance. And the current work usually considers the classic XL-RIS, XL-RIS elements are regularly arranged on the grid with constant spacing, but the current work usually considers the classic XL-RIS, XL-RIS elements are regularly arranged on the grid with constant spacing, at this time the performance is highly dependent on the number of XL-RIS elements, and to improve the array gain of XL-RIS can only increase the number of elements. Although RIS is low in cost, expanding the number of elements will increase the computational complexity and cause certain resource waste. Therefore, in the multi-user XL-RIS assisted wireless communication, it is necessary to deeply study the new design scheme to overcome these challenges.

[0005] Therefore, in the multi-user XL-RIS assisted wireless communication, it is necessary to deeply study the new design method to overcome these challenges. SUMMARY

[0006] The purpose of the present application is to provide a method for large-scale reconfigurable intelligent surface topology optimization and beam design based on near-field propagation, further reducing the array gain loss of XL-RIS, making the designed beam closer to the user's desired direction, expanding the range of near-field distance and improving the capacity of the communication system.

[0007] The above technical purpose of the present application is realized by the following technical scheme:

[0008] A near-field large-scale reconfigurable intelligent surface topology optimization and beam design method is applied to a wireless communication system, characterized in that the wireless communication system comprises the following steps:

[0009] Step 1: Establish an optimized deployment XL-RIS assisted multi-user wireless communication system, give the signal model of XL-RIS assisted wireless communication, and give the channel model in the near-field environment, the receiving user is located in the near-field area and is affected by the spherical wave beam reflected by the large-scale reconfigurable intelligent surface deployed optimally;

[0010] Step2: Establish a target function to represent the minimization of user expected beam gain error problem, considering the actual performance of k XL-RIS beam pointing to the user, the optimization deployment of XL-RIS topology constraint and the restriction of beam weight, decompose the problem into irregular reconfigurable intelligent surface topology structure, beam forming vector sub-problems; Through Majorization-Minimization algorithm, adaptive tabu search algorithm, convert the non-convex sub-problem into convex problem;

[0011] Step3: Design a joint optimization scheme to jointly optimize the irregular large-scale reconfigurable intelligent surface topology structure and beam forming vector, minimize the user expected beam gain error, so as to ensure that the beam is more accurately aligned to the direction expected by the user, and thus significantly improve the transmission rate in multi-user scenario.

[0012] Further preferably, Step1, specifically expressed as:

[0013] Step1.1: An optimized deployment XL-RIS assisted multi-user wireless communication system, including a base station, an optimized deployment large-scale reconfigurable intelligent surface, K communication users, and a signal reflected by the large-scale reconfigurable intelligent surface spherical wave beam affecting the user. Due to the propagation characteristics of the large-scale reconfigurable intelligent surface, all users are located in the near-field region.

[0014] Step1.2: Define for the user set, denotes the XL-RIS topology matrix after optimization deployment, wherein Define z n ∈{1,0} to determine whether the XL-RIS element is deployed on the nth grid point, i.e. z n =1 indicates that the XL-RIS element selects the nth grid point, and z n =0 indicates that the nth grid point is not selected, and the kth user receives the signal

[0015]

[0016] In the formula: is the precoding transmission signal at the base station; denote the channel from XL-RIS to the kth user and the channel from the base station to XL-RIS, respectively; denotes the beam forming matrix of XL-RIS; denotes the additive white Gaussian noise received by the kth user, and a large number of reflective elements constitute XL-RIS, which can adjust the phase of the incident signal by designing Θ;

[0017]

[0018] where θ denotes the XL-RIS beamforming vector; denotes the reflection coefficient of the n-th XL-RIS unit optimized for deployment, the received signal model is

[0019]

[0020] where the reflection channel of the k-th user is defined as

[0021] Step 1.3: Based on the system model, the transmission and rate of the k-th user in the XL-RIS-aided communication system after optimized deployment are optimized as:

[0022]

[0023] Step 1.4: Since the deployment positions of the base station and the XL-RIS are fixed, it is assumed that the precoding at the base station has been designed, and the base station is regarded as a single-antenna transmitter in this system model. Referring to the far-field propagation model, the effective reflection channel under near-field propagation can be written as

[0024]

[0025] where: is the effective gain; r re,k is the spatial coordinate vector from the XL-RIS to the k-th user; r in is the spatial coordinate vector from the base station to the XL-RIS; the system directly determines the XL-RIS beamforming vector through a beam training process, and the channel remains unchanged at each small time scale. The first process is independently performed by all users to obtain the expected beam direction of each user with the expected beam gain. According to these beam parameters, the beamforming vector of the XL-RIS optimized for deployment is designed in the design process, and finally data transmission is performed using the beamforming vector. In order to obtain the expected XL-RIS beamforming, the beam training process is modeled as

[0026]

[0027] where: c i is the XL-RIS beamforming code word selected from the code book C.

[0028] Further preferably, Step 2 is specifically expressed as:

[0029] Step 2.1: Considering the actual performance of the k-th XL-RIS beam pointing to the user, which is restricted by the optimized deployment of the XL-RIS topology and the beam weight, the p1 objective function represents the minimization of the user expected beam gain error:

[0030] ​

[0031] C3:1 T z=N

[0032] In the formula: XL-RIS beamforming vector θ, topology Z is the variable to be optimized; The code word θ obtained in the beam training process k The K-column steering vector matrix composed of is the expected beam gain vector, C1 represents the unit module constraint between beams; C2 and C3 represent the optimization deployment XL-RIS topology constraint, wherein the N diagonal elements of the topology matrix Z are assigned as 1, and the remaining N S -N diagonal elements are assigned as 0, and the objective function P1 is rewritten as

[0033]

[0034] C3:1 T z=N

[0035] Step2.2: search for the optimal XL-RIS topology using the adaptive tabu search method, specifically, for a given topology, P2 can be simplified as:

[0036]

[0037] C2: z=z0

[0038] Step2.3: after obtaining the fixed suboptimal topology Z, the Minorize-Maximization (MM) algorithm is used to solve the problem, and θ and are optimized, and P3 is decomposed into a series of approximate subproblems, which are solved iteratively, and the optimal solution of the original objective function is approached, and for the optimization problem min x f(x), the MM algorithm constructs a strict surrogate function {q(x|x t )}(t=1, 2, …), and each surrogate function is used as the objective function of the subproblem, and q(x|x t ) is the objective function of the t+1 iteration, wherein x t+1 =argmin x q(x|x t ), taking the t+1 iteration as an example, q(x|x t ) satisfies the following four properties:

[0039]

[0040] The first feature in Equation (1) ensures that each surrogate function is an upper bound of the original objective function, the second and third features in Equations (2) and (3) ensure that at the unique intersection point, the first-order gradient between the original objective function and the surrogate function is equal, and the fourth feature in Equation (4) ensures that the upper bound is strictly decreasing from the tth iteration to the (t+1)th iteration, so that f(x t+1 ) = q(x t+1 |x t ) ≤ q(x t |x t ) = f(x t ) is established, and then by solving the corresponding sequence of sub-problems {q(x|x t )} (t = 1, 2, …), the optimal value of min q(x|x t ) is monotonically decreasing with respect to t, and finally converges to the optimal value of the original problem;

[0041] Step 2.4: Lemma 1: The surrogate function q(θ|θ t ) that satisfies the above four features can be defined as Equation (11) for any given θ t and any available θ, where λ max represents the maximum eigenvalue of ΞΞ H .

[0042]

[0043] Step 2.5: The XL-RIS beamforming θ is solved by iteratively optimizing a series of minimization problems {q(θ|θ t )} (t = 1, 2, …), for the (t+1)th iteration, M is substituted into λ max I N×N , and the optimization is represented as

[0044]

[0045] Step 2.6: To enrich the degrees of freedom of the optimization objective, a beam phase gain optimization variable is introduced to solve the phase factor of the expected beam gain as an optimizable variable, and the optimization problem P3 is rewritten as

[0046]

[0047] C2: z = z0

[0048]

[0049] For a given θ t and z0, the phase factor can be optimized as

[0050]

[0051] Further preferably, Step3 is specifically expressed as:

[0052] Step3.1: randomly generate the topology matrix Z1, generate the neighborhood of Z1, obtain the sum rate of the above-mentioned neighbors through beamforming, select the optimal neighbor and obtain the topology matrix Z2 through the adaptive tabu search algorithm, perform the next iteration according to Z2, and obtain the suboptimal solution of P1 when the maximum iteration number is reached or the termination condition is met, which is expressed as: Z opt and θ opt ;

[0053] Step3.2: the feasible selection position of the XL-RIS element has an additional spatial degree of freedom, which not only expands the near-field area under the condition of a limited number of elements, but also improves the user carrying capacity and the overall capacity of the system in the area.

[0054] Compared with the prior art, the present application has the following beneficial effects:

[0055] The present application compares with the conventional deployment of XL-RIS beam design under the existing far-field and near-field propagation channel, not only significantly improves the system capacity, but also ensures that the beam is more accurately aligned in the direction expected by the user, thereby significantly improving the transmission rate in a multi-user scenario. Under the condition of not increasing the number of elements, the potential of optimizing the deployment of XL-RIS structure to improve the performance of wireless communication system is provided as a new strategy direction for the design of near-field communication system. BRIEF DESCRIPTION OF DRAWINGS

[0056] Figure 1 It is a schematic diagram of the structure of the wireless communication system assisted by the optimized deployment of XL-RIS in an embodiment;

[0057] Figure 2 It is a schematic diagram of the optimization scheme flow of the transmission scheme of the optimized deployment of XL-RIS in an embodiment;

[0058] Figure 3 It is the topology structure of the optimized deployment of XL-RIS in the present application, and too many elements are selected to illustrate the local situation of the topology structure, wherein the blue square represents the selected position of the XL-RIS element;

[0059] Figure 4 The relationship between the sum rate and the number of grid points of the optimized deployment of XL-RIS;

[0060] Figure 5 It is the performance comparison of the sum rate and the number of users. DETAILED DESCRIPTION

[0061] The application will be described in further detail below with reference to the drawings.

[0062] Embodiment, a near-field large-scale reconfigurable intelligent surface topology optimization and beam design method, comprising the following steps:

[0063] Step1: Establish an optimized deployment XL-RIS assisted multi-user wireless communication system, give the signal model of XL-RIS assisted wireless communication, and give the channel model in the near-field environment. The receiving user is located in the near-field region and is affected by the spherical wave beam reflected by the large-scale reconfigurable intelligent surface of the optimized deployment.

[0064] The XL-RIS assisted multi-user wireless communication system model optimization deployment method, characterized by Step1, is specifically described as follows:

[0065] Step1.1: An optimized deployment XL-RIS assisted multi-user wireless communication system contains a base station, an optimized deployment large-scale reconfigurable intelligent surface, K communication users, and a signal reflected by a large-scale reconfigurable intelligent surface. The spherical wave beam affects the user, and due to the propagation characteristics of the large-scale reconfigurable intelligent surface, all users are located in the near-field region.

[0066] Step1.2: Define for the user set, XL-RIS topology matrix after optimization, wherein Define z n ∈{1,0} to determine whether the XL-RIS element is deployed on the nth grid point, i.e. z n =1 indicates that the XL-RIS element selects the nth grid point, and z n =0 indicates that the nth grid point is not selected. The received signal of the kth user is represented as

[0067]

[0068] In the formula: is the precoding transmission signal at the base station; respectively represent the channel from the XL-RIS to the kth user and the channel from the base station to the XL-RIS; represents the beamforming matrix of the XL-RIS; represents the additive white Gaussian noise received by the kth user. The XL-RIS composed of a large number of reflecting elements can adjust the phase of the incident signal by designing Θ;

[0069]

[0070] In the formula: θ represents the XL-RIS beamforming vector; denotes the reflection coefficient of the nth XL-RIS unit for optimal deployment. According to formula (2), the received signal model is

[0071]

[0072] where the reflection channel of the kth user is defined as

[0073] Step 1.3: Based on the system model, the transmission and rate of user k in the XL-RIS-aided communication system after optimal deployment are optimized as follows:

[0074]

[0075] Step 1.4: Since the deployment locations of the base station and the XL-RIS are fixed, it is assumed that the precoding at the base station has been designed, and the base station is regarded as a single-antenna transmitter in this system model. Referring to the far-field propagation model, the effective reflection channel under near-field propagation can be written as

[0076]

[0077] where is the effective gain; r re,k is the spatial coordinate vector from the XL-RIS to the kth user; r in is the spatial coordinate vector from the base station to the XL-RIS; the system directly determines the XL-RIS beamforming vector through a beam training process, and the channel remains unchanged at each small time scale. The first process is independently performed by all users to obtain the desired beam direction of each user with the desired beam gain. According to these beam parameters, the beamforming vector of the XL-RIS for optimal deployment is designed in the design process, and finally data transmission is performed using the beamforming vector. To obtain the desired XL-RIS beamforming, the beam training process is modeled as

[0078]

[0079] where c i is the XL-RIS beamforming code word selected from the code book C.

[0080] Step 2: The objective function is established to represent the problem of minimizing the user expected beam gain error, considering the actual performance of the k XL-RIS beams pointing to the user, the constraints of the optimized deployment XL-RIS topology and the beam weight, and the problem is decomposed into the topology of the irregular reconfigurable intelligent surface and the beamforming vector sub-problems; through the Majorization-Minimization (MM) algorithm and the adaptive tabu search algorithm, the non-convex sub-problems are converted into convex problems. ​

[0081] The objective function represents a minimization of user desired beam gain error problem, characterized by Step2, specifically expressed as:

[0082] Step2.1: Considering the actual performance of k XL-RIS beams pointing to users, constrained by the optimized deployment of XL-RIS topology and beam weight, the p1 objective function represents the minimization of user desired beam gain error:

[0083]

[0084] In the formula: XL-RIS beamforming vector θ, topology structure Z are to be optimized variables; is the code word θ obtained in the beam training process k consisting of K columns of steering vector matrix; is the expected beam gain vector. C1 represents the unit module constraint between beams; C2 and C3 represent the optimized deployment of XL-RIS topology constraints, where the N diagonal elements of the topology matrix Z are assigned as 1, and the remaining N S -N diagonal elements are assigned as 0. The objective function P1 is rewritten as

[0085]

[0086] Step2.2: Use the adaptive tabu search method to search for the optimal XL-RIS topology, specifically, for a given topology, P2 can be simplified as:

[0087]

[0088] Step2.3: After obtaining the fixed suboptimal topology Z, the Minorize-Maximization (MM) algorithm is used to solve the problem, and θ and are optimized. P3 is decomposed into a series of approximate subproblems, which are solved iteratively to approach the optimal solution of the original objective function. For the optimization problem min x f(x), the MM algorithm constructs a strict surrogate function {q(x|x t )}(t=1,2,…), and each surrogate function is used as the objective function of the subproblem. Let q(x|x t ) be the objective function of the (t+1)th iteration, where x t+1 =argmin x q(x|x t ). Taking the (t+1)th iteration as an example, q(x|x t ) satisfies the following four properties:

[0089]

[0090] The first feature in equation (10a) ensures that each surrogate function is an upper bound of the original objective function, the second and third features in equations (10b) and (10c) ensure that at the unique intersection point, the first order gradient between the original objective function and the surrogate function is equal, and the fourth feature in equation (10d) ensures that the upper bound is strictly decreasing from the tth iteration to the t+1th iteration.

[0091] f(x t+1 ) = q(x t+1 |x t ) ≤ q(x t |x t ) = f(x t ) holds, then by solving the corresponding sequence of sub-problems {q(x|x t )}, (t = 1, 2,...), the optimal value of min q(x|x t ) is monotonically decreasing with respect to t, and finally converges to the optimal value of the original problem.

[0092] The optimization problem is reformulated as

[0093]

[0094] Lemma 1: A surrogate function q(θ|θ t ) that satisfies the above four features simultaneously can be defined as equation (12) for any given θ t and any available θ, where λ max represents the maximum eigenvalue of ΞΞ H .

[0095]

[0096] The XL-RIS beamformer θ is solved by iteratively optimizing a sequence of minimization problems {q(θ|θ t )}, (t = 1, 2,...). For the t+1th iteration, M is substituted into λ max I N×N , and the optimization is represented as

[0097]

[0098] To enrich the degrees of freedom of the optimization objective, a beam phase gain optimization variable is introduced to solve the phase factor of the expected beam gain as an optimizable variable, the optimization problem P3 is rewritten as

[0099]

[0100] For a given θ t and z0, the phase factor can be optimized as

[0101]

[0102] Step 3: Design a joint optimization scheme to jointly optimize the irregular large-scale reconfigurable smart surface topology and beamforming vectors to minimize the user's desired beam gain error, thereby ensuring that the beam is more accurately aligned in the user's desired direction, thereby significantly improving the transmission rate in multi-user scenarios.

[0103] The joint optimization scheme jointly optimizes the irregular large-scale reconfigurable smart surface topology and the beamforming vector to minimize the user's desired beam gain error. It is characterized by Step 3, which is specifically expressed as follows:

[0104] Step 3.1: Randomly generate a topology matrix Z1, generate the neighborhood of Z1, obtain the sum rate of the neighbors through beamforming, select the optimal neighbor and obtain the topology matrix Z2 through the adaptive tabu search algorithm, and perform the next iteration based on Z2. When the maximum number of iterations is reached or the termination condition is met, the suboptimal solution of P1 is obtained, which is expressed as: Z opt and θ opt .

[0105] Step 3.2: Utilizing the additional spatial freedom of the feasible selection positions of XL-RIS components, not only does it expand the near-field area under the condition of a limited number of components, but it also improves the user carrying capacity in this area and the overall capacity of the system.

[0106] like Figure 1 The figure shows an optimized deployment of XL-RIS assisted multi-user wireless communication system. The receiving user is located in the near field area and is affected by the spherical beam reflected from the XL-RIS. The N XL-RIS elements of the optimized large-scale reconfigurable smart surface are irregularly distributed on the Ns grid points of the amplified surface (Ns>N). Based on this, the signal model of XL-RIS assisted wireless communication is established, and the channel model in the near field environment is given. For the non-convex optimization problem established in Step 2, as shown in Figure 2 As shown, the objective function is established to represent the problem of minimizing the user's expected beam gain error. Only the actual performance of k XL-RIS beams pointing to the user is considered. Subject to the constraints of the optimized XL-RIS topology and beam weights, the problem is decomposed into two sub-problems: irregular reconfigurable smart surface topology structure and beamforming vector. Through the Majorization-Minimization (MM) algorithm and the adaptive tabu search algorithm, the non-convex sub-problem is converted into a convex problem for alternating optimization and solution, thereby ensuring that the beam is more accurately aligned with the user's desired direction, thereby significantly improving the transmission rate in multi-user scenarios. The specific steps are as follows:

[0107] First, the topology matrix Z1 is randomly generated, the neighborhood of Z1 is generated based on the cross-entropy algorithm of neighbor extraction, and then the sum rate of the above neighbor is obtained through beamforming, the optimal neighbor is selected and the topology matrix Z2 is obtained through the ATS algorithm, and the next iteration is performed according to Z2. Finally, when the maximum iteration number is reached or the termination condition is met, the suboptimal solution of P1 is obtained, which is represented as: Z opt and θ opt .

[0108] For the fixed beamforming vector, an optimized XL-RIS topology ATS algorithm is proposed, which iteratively searches for possible topologies through adaptive moving criteria, thereby obtaining a suboptimal XL-RIS topology, as shown in Figure 3 .

[0109] After obtaining the fixed suboptimal topology Z, the Minorize-Maximization (MM) algorithm is used to solve the problem P2, and θ and are optimized. P2 is decomposed into a series of approximate subproblems, which are iteratively solved in turn to approximate the optimal solution of the original objective function.

[0110] In this embodiment, a three-dimensional scene in Cartesian coordinates is considered, and it is assumed that the optimized deployment of XL-RIS is in the yOz plane with the center at the coordinate origin (0, 0, 0), and the XL-RIS element spacing is d R = λ / 2. In this scenario, 512 XL-RIS elements are deployed, and the grid is deployed in a manner of multiplying the number of elements. The near-field channel and the far-field channel are generated using codebooks, and the remaining parameters are set as follows: the carrier frequency is 10 GHz, when the grid size is 64*16, the XL-RIS array aperture reaches 0.9674 m, and the near-field coverage distance is 62.39 m; when the grid size is 128*8, the XL-RIS array aperture reaches 1.9237 m, and the near-field propagation coverage distance expands to 246.72 m. Based on these parameter settings, MATLAB is used for simulation.

[0111] As shown in Figure 4 , by increasing the surface size to change the irregularity ratio of the optimized deployment of XL-RIS, the user sum rate performance can be effectively improved. With the increase of the grid points of the optimized deployment of XL-RIS, the user sum rate performance of the optimized deployment of XL-RIS can be effectively improved.

[0112] As shown in Figure 5 , the system sum rate of the optimized deployment of XL-RIS beam design scheme in the near-field environment is better than that of the conventional method. Whether in the near-field or the far-field environment, the proposed optimized deployment scheme still has good performance and can effectively improve the performance of the entire user system.

[0113] The embodiments are only used to explain the present application, and are not used to limit the present application, and any modification without creative contribution made by the person skilled in the art according to the embodiments after reading the specification is protected by the patent law as long as it is within the scope of the claims of the present application.

Claims

1. A near-field large-scale reconfigurable smart surface topology optimization and beam design method, applied to wireless communication systems, characterized by: The wireless communication system comprises the following steps: Step 1: Establish an optimized XL-RIS-assisted multi-user wireless communication system, provide a signal model for XL-RIS-assisted wireless communication, and provide a channel model in a near-field environment. The receiving user is located in the near-field area and is affected by the spherical beam reflected from the optimized large-scale reconfigurable smart surface. Step 2: Establish an objective function to represent the problem of minimizing the user's expected beam gain error. Considering the actual performance of k XL-RIS beams pointing to the user, subject to the constraints of the optimized XL-RIS topology and beam weights, the problem is decomposed into the irregular reconfigurable smart surface topology and beamforming vector quantum problems. Using the Majorization-Minimization algorithm and the adaptive tabu search algorithm, the non-convex subproblem is converted into a convex problem. Step 3: Design a joint optimization scheme to jointly optimize the irregular large-scale reconfigurable smart surface topology and beamforming vectors to minimize the user's desired beam gain error, thereby ensuring that the beam is more accurately aligned in the user's desired direction, thereby significantly improving the transmission rate in multi-user scenarios.

2. A near-field large-scale reconfigurable smart surface topology optimization and beam design method according to claim 1, characterized in that: Step 1 is specifically stated as follows: Step 1.1: An optimally deployed XL-RIS-assisted multi-user wireless communication system consists of a base station, an optimally deployed large-scale reconfigurable smart surface, and K communication users. The spherical beam reflected by the large-scale reconfigurable smart surface affects the users. Due to the propagation characteristics of the large-scale reconfigurable smart surface, all users are located in the near-field area. Step 1.2: Definition For the user set, represents the XL-RIS topology matrix after optimized deployment, where Define z n ∈{1,0} to determine whether the XL-RIS element is deployed on the nth grid point, i.e., z n =1 means that the XL-RIS element selects the nth grid point, z n =0 means that the nth grid point is not selected, and the received signal of the kth user is expressed as: Where: is a precoded transmit signal at the base station; denote the channel from XL-RIS to the kth user and the channel from the base station to XL-RIS respectively; represents the beamforming matrix of XL-RIS; represents the additive white Gaussian noise received by the kth user. The XL-RIS composed of a large number of reflective elements can adjust the phase of the incident signal by designing Θ; Where: θ represents the XL-RIS beamforming vector; represents the reflection coefficient of the nth XL-RIS unit in the optimal deployment, and the received signal model is Where: The reflection channel of the kth user is defined as Step 1.3: Based on the system model, the transmission sum rate of user k in the optimized XL-RIS assisted communication system is: Step 1.4: Since the deployment locations of the base station and XL-RIS are fixed, it is assumed that the precoding at the base station has been designed. In this system model, the base station is regarded as a single-antenna transmitter. Referring to the far-field propagation model, the effective reflection channel under near-field propagation It can be written as; Where: is the effective gain; r re,k is the spatial coordinate vector from XL-RIS to the kth user; r in is the spatial coordinate vector from the base station to XL-RIS; the system directly determines the XL-RIS beamforming vector through the beam training process. At each small time scale, the channel remains unchanged. The first process is performed independently by all users to obtain the desired beam direction of each user with the desired beam gain. Based on these beam parameters, the beamforming vector of XL-RIS is designed and deployed in the design process. Finally, the beamforming vector is used for data transmission. To obtain the desired XL-RIS beamforming, the beam training process is modeled as Where: c i is the XL-RIS beamforming codeword selected from codebook C.

3. The near-field large-scale reconfigurable smart surface topology optimization and beam design method according to claim 1, characterized in that: Step 2 is specifically stated as follows: Step 2.1: Considering the actual performance of k XL-RIS beams pointing to the user, subject to the optimally deployed XL-RIS topology constraints and beam weights, the p1 objective function represents minimizing the user's expected beam gain error: P1: s·t C1: C2: C3:1 T z=N Where: XL-RIS beamforming vector θ and topology structure Z are variables to be optimized; is the codeword θ obtained during the beam training process k The K-column steering vector matrix composed of is the expected beam gain vector, C1 represents the unit mode constraint between beams; C2 and C3 represent the constraints of the optimized XL-RIS topology, where the N diagonal elements of the topology matrix Z are assigned the value of 1, and the remaining N S -N diagonal elements are assigned 0, and the objective function P1 is rewritten as P2: s·t C1: C2: C3:1 T z=N Step 2.2: Use the adaptive tabu search method to search for the optimal XL-RIS topology. Specifically, for a given topology, P2 can be simplified as: P3: s·t C1: C2:z=z0 Step 2.3: After obtaining a fixed suboptimal topology Z, the Minorize-Maximization (MM) algorithm is used to solve the problem. Optimize and decompose P3 into a series of approximate sub-problems, solve them iteratively in turn, and approach the optimal solution of the original objective function. For the optimization problem x∈X, min x f(x), the MM algorithm constructs a strict proxy function {q(x|x t )}(t=1,2,…), each proxy function is used as the objective function of the sub-problem, and q(x|x t ) is the target of the t+1th iteration, where x t+1 =argmin x q(x|x t ), taking the t+1th iteration as an example, q(x|x t ) satisfies the following four properties: The first feature in Equation 1 ensures that each surrogate function is the upper bound of the original objective function. The second and third features in Equations 2 and 3 ensure that the first-order gradients between the original objective function and the surrogate function are equal at the unique intersection point. Equation 4 ensures that the upper bound is strictly decreasing from the tth to the t+1th iteration, so that f(x t+1 )=q(x t+1 |x t )≤q(x t |x t )=f(x t ) holds true, and then by solving the corresponding sequence of subproblems {q(x|x t )}, (t=1,2,…), min q(x|x t ) decreases monotonically with respect to t and eventually converges to the optimal value of the original problem; Step 2.4: Lemma 1: The proxy function q(θ|θ) that satisfies the above four characteristics at the same time t ) can be defined as Equation (11), for any given θ t and any available θ, where λ max Indicates ΞΞ H The biggest feature Step 2.5: XL-RIS beamforming θ is solved by iterative optimization to obtain a series of {q(θ|θ t )}, (t=1,2,…) minimization problem is solved, for t+1 iterations, substitute M into λ max I N×N , the optimization is expressed as Step 2.6: To enrich the degree of freedom of the optimization objective, the beam phase gain optimization variable is introduced to solve the phase factor of the expected beam gain. As an optimizable variable, The optimization problem P3 is rewritten as P3: s·t C1: C2:z=z0 C3: For a given θ t and z0, the phase factor Can be optimized to 4. The near-field large-scale reconfigurable smart surface topology optimization and beam design method according to claim 1, characterized in that: Step 3 is specifically stated as follows: Step 3.1: Randomly generate a topological matrix Z1, generate the neighborhood of Z1, obtain the sum rate of neighbors through beamforming, select the optimal neighbor and obtain the topological matrix Z2 through the adaptive tabu search algorithm, and perform the next iteration based on Z2. When the maximum number of iterations is reached or the termination condition is met, the suboptimal solution of P1 is obtained, which is expressed as: Z opt and θ opt ; Step 3.2: Utilizing the additional spatial freedom of the feasible selection positions of XL-RIS components, not only does it expand the near-field area under the condition of a limited number of components, but it also improves the user carrying capacity in this area and the overall capacity of the system.

Images