RIS enabling CF network downlink communication method under complex non-ideal CSI and power limited conditions

Abstract

The invention relates to an RIS enabling CF network downlink communication technology under complex non-ideal CSI and power limited conditions, and belongs to the technical field of communication. Comprising the following steps: constructing a complex non-ideal CSI and power-limited CF network model based on an RIS auxiliary scene, and constructing a transmission model of a receiving end user according to the RIS auxiliary complex non-ideal CSI and power-limited CF network model; obtaining a signal to interference plus noise ratio of the receiving end user according to the transmission model of the receiving end user; designing a model for maximizing the worst user communication rate according to the transmission model of the receiving end user and the signal to interference plus noise ratio of the receiving end user; and solving the optimization problem of the model for maximizing the worst user communication rate by using an alternating algorithm to obtain a final information transmission scheme. Compared with the existing scheme of only optimizing the beam forming and only optimizing the beam forming and the phase shift and amplitude of the RIS electromagnetic unit, the technology can more effectively improve the system communication rate in a more complex environment.

Description

Technical Field

[0001] The present invention belongs to the field of communication technology, and in particular relates to a RIS-enabled CF network downlink communication method under complex non-ideal CSI and power-limited conditions. Background Art

[0002] With the rapid development of wireless communication technology and the advent of the era of the Internet of Everything, the number of communication users and their requirements for communication speed are increasing rapidly. The problems of signal interference and frequent handover in the traditional cellular architecture are gradually becoming prominent, which seriously affects the communication quality and user experience. Therefore, a user-centric Cell-Free (CF) large-scale Multiple Input Multiple Output (MIMO) architecture is proposed to effectively solve the spectrum efficiency and energy efficiency bottlenecks faced by wireless communication. In the CF network, a large number of distributed small access points (APs) are deployed around the user, and the central processing unit allocates resources to each AP and coordinates and controls the APs to provide data communication services to the user. This architecture reduces the average path loss between APs and users, effectively reduces interference between users, and thus obtains a better user communication experience. However, due to the uncertain factors of the changing surrounding environment, the communication between APs and users is likely to be blocked by objects such as trees and buildings, resulting in poor user experience. In recent years, a new type of wireless device called Reconfigurable Intelligent Surfaces (RIS) has been proposed by the industry and is considered to be one of the potential key technologies for the future 6G. RIS is composed of a large number of reconfigurable reflective electromagnetic units, which can intelligently adjust the phase, amplitude, polarization and frequency of the incident signal and then reflect it to the user. The emergence of RIS has realized the control of the uncontrollable wireless propagation environment and built a new paradigm of intelligent programmable wireless environment. RIS does not have signal processing capabilities, but it has the characteristics of low cost, low complexity and strong flexibility. It is widely used in CF network architecture as a bridge for communication between AP and users, thereby completely solving the problem of signal obstruction. Then, the electromagnetic unit is used to adjust the amplitude and phase of the signal, which can more effectively improve the overall transmission performance of the network.

[0003] Many existing studies have proved that RIS can significantly expand the coverage of signals and ensure the reliability of communication of CF network users. Then, through the algorithm solution, the electromagnetic beams of AP and RIS are jointly designed to achieve a significant improvement in network transmission performance. However, the existing technology has two main shortcomings. First, the existing technology considers that the power resources of RIS are sufficient to ensure the normal operation of all its electromagnetic units. Although RIS is a low-power device, the power consumption of the circuit on each electromagnetic unit cannot be ignored, especially the active RIS newly proposed in the industry, the power amplifier devices on its circuit have higher power requirements. Therefore, when the power supply power of RIS in the system is limited, it is often unable to support all electromagnetic units to work normally. How to select limited electromagnetic units to improve the transmission performance of CF network in this case has high application value, but there is still a lack of effective technical support. Secondly, in practical applications, due to the factor of quantization error, the system usually has estimation errors in the process of estimating channel state information (CSI), and the existing technology only considers the cascade channel estimation error of AP-RIS-user, but ignores the estimation errors of the more complex AP-RIS and RIS-user channels. In some application scenarios, the CSI of these two segments need to be estimated separately, such as MIMO scenarios with multi-antenna APs and users, and scenarios assisted by active RIS. Especially in active RIS-assisted scenarios, in order to suppress the impact of RIS-side amplified noise on user reception performance, it is necessary to obtain the CSI of the RIS-user segment. In addition, new errors may also be generated in the estimation process, thereby affecting the accurate pointing of the signal beam and causing a reduction in user reception performance. Therefore, the new problem of improving user data transmission performance under the complex conditions of considering the cascade of RIS and the segmented CSI estimation errors is of greater positive significance for promoting the application of RIS technology in CF networks to maturity and practicality. At present, no existing technology has been found to be able to solve this problem. Summary of the invention

[0004] The technical problems to be solved by the present invention are:

[0005] In order to avoid the shortcomings of the prior art, the present invention provides a RIS-enabled CF network downlink communication method under complex non-ideal CSI and power-constrained conditions, so as to solve the problems of single link error estimation and limited system power resources in the prior art.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] A RIS-enabled CF network downlink communication method under complex non-ideal CSI and power-limited conditions, characterized by comprising:

[0008] Based on the RIS-assisted scenario, a CF network model under complex non-ideal CSI and power limitation is constructed, and a transmission model of a receiving end user is constructed according to the RIS-assisted complex non-ideal CSI and power limitation CF network model;

[0009] Obtaining a signal to interference and noise ratio of the receiving end user according to the transmission model of the receiving end user;

[0010] Designing a model for maximizing the worst user communication rate according to the transmission model of the receiving end user and the signal to interference noise ratio of the receiving end user;

[0011] The optimization problem of maximizing the worst user communication rate model is solved by using an alternating algorithm to obtain a final information transmission solution.

[0012] A further technical solution of the present invention: the RIS-assisted scenario is a passive RIS-assisted scenario or an active RIS-assisted scenario.

[0013] A further technical solution of the present invention is as follows: the RIS-assisted complex non-ideal CSI and CF network model under power limitation is a passive RIS-assisted complex non-ideal CSI and CF network model under power limitation or an active RIS-assisted complex non-ideal CSI and CF network model under power limitation.

[0014] A further technical solution of the present invention is to solve the optimization problem of the model of maximizing the worst user communication rate by using an alternating algorithm to obtain the final information transmission solution; including:

[0015] The RIS-assisted scenario maximizes the worst user communication rate model and uses beamforming w and RIS electromagnetic unit coefficients to calculate the maximum worst user communication rate model. The RIS electromagnetic unit selects the matrix A as two sub-optimization models of the optimization variables to optimize the model of maximizing the worst user communication rate;

[0016] The final information transmission scheme is obtained within the framework of the alternating algorithm.

[0017] A further technical solution of the present invention is to decompose the maximization worst user communication rate model in the RIS-assisted scenario, and respectively use beamforming w and RIS electromagnetic unit coefficients to obtain the worst user communication rate model. The RIS electromagnetic unit selects the matrix A as the two sub-optimization models of the optimization variables to optimize the step of maximizing the worst user communication rate model, including:

[0018] The electromagnetic unit parameters and selection matrix parameters of RIS are set, and the beamforming optimization strategy in this scenario is obtained by combining the maximization worst user communication rate model in the RIS-assisted scenario;

[0019] According to the beamforming optimization strategy of RIS and combined with the model of maximizing the worst user communication rate, the optimization strategy of electromagnetic unit coefficients and selection matrix of RIS is obtained;

[0020] Based on the electromagnetic unit and selection matrix optimization strategy of RIS, the electromagnetic unit parameters and selection matrix parameters of RIS are iteratively updated, and the above steps are repeated until the variables of the beamforming optimization strategy of RIS and the electromagnetic unit and selection matrix optimization strategy of RIS converge, and a solution for maximizing the worst user communication rate is obtained.

[0021] A computer system, characterized in that it comprises: one or more processors, and a computer-readable storage medium for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned method.

[0022] A computer-readable storage medium, characterized in that it stores computer-executable instructions, which are used to implement the above method when executed.

[0023] A computer program product, characterized by comprising computer executable instructions, wherein the instructions are used to implement the above method when executed.

[0024] The beneficial effects of the present invention are:

[0025] The present invention provides a RIS-enabled CF network downlink communication method under complex non-ideal CSI and power-limited conditions. Under the complex non-ideal CSI conditions that take into account the estimation errors of cascaded links and segmented links, high-quality communication under the CF large-scale MIMO architecture is completed. At the same time, under power-limited conditions, the present invention proposes to optimize the selection of a limited number of RIS electromagnetic units that can work normally through the RIS selection matrix, further improving the RIS's ability to regulate the electromagnetic environment, thereby improving the communication rate under power-limited conditions. Moreover, the present invention uses active RIS to amplify the amplitude of the reflected signal to effectively suppress the "double fading" effect that is difficult to solve in the existing passive RIS-assisted scenario. Compared with the existing solutions of only optimizing beamforming and only optimizing beamforming and RIS electromagnetic unit phase shift and amplitude, this technology can more effectively improve the system communication rate in a more complex environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The drawings are only for the purpose of illustrating particular embodiments and are not to be considered limiting of the present invention. Like reference symbols denote like components throughout the drawings.

[0027] Figure 1A step diagram of a method for enabling CF network downlink communication with RIS under complex non-ideal CSI and power-limited conditions in an exemplary embodiment of the present disclosure is shown.

[0028] Figure 2 A schematic diagram of a scenario of RIS-enabled CF network downlink communication technology under complex non-ideal CSI and power-limited conditions in an exemplary embodiment of the present disclosure is shown.

[0029] Figure 3 A comparison diagram showing the convergence of the worst user communication rate of the present technology under different numbers of RIS electromagnetic units in an exemplary embodiment of the present disclosure is shown.

[0030] Figure 4 A comparison diagram of transmission performance between the present technology and only the optimized beamforming solution as AP transmit power increases in an exemplary embodiment of the present disclosure is shown.

[0031] Figure 5 A transmission performance comparison diagram of the present technology in an exemplary embodiment of the present disclosure with only optimized beamforming and only optimized beamforming and RIS electromagnetic unit phase shift and amplitude scheme as the error channel uncertainty increases is shown. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0033] This example embodiment provides a RIS-enabled CF network downlink communication method under complex non-ideal CSI and power-limited conditions. Figure 1 As shown in , the RIS enabling cellular network technology may include: steps S101 to S104.

[0034] Step S101: constructing a complex non-ideal CSI and a CF network model under power limitation based on a RIS-assisted scenario, and constructing a transmission model of a receiving end user according to the RIS-assisted complex non-ideal CSI and the CF network model under power limitation; wherein the RIS-assisted scenario is a passive RIS-assisted scenario or an active RIS-assisted scenario, and the RIS-assisted complex non-ideal CSI and the CF network model under power limitation is a passive RIS-assisted complex non-ideal CSI and the CF network model under power limitation or an active RIS-assisted complex non-ideal CSI and the CF network model under power limitation;

[0035] Step S102: obtaining a signal-to-interference-and-noise ratio of the receiving end user according to the transmission model of the receiving end user;

[0036] Step S103: designing a model for maximizing the worst user communication rate according to the transmission model of the receiving end user and the signal to interference noise ratio of the receiving end user;

[0037] Step S104: using an alternating algorithm to solve the optimization problem of the maximization worst user communication rate model to obtain a final information transmission solution.

[0038] Through the above-mentioned RIS-enabled CF network downlink communication method under complex non-ideal CSI and power-constrained conditions, high-quality communication under the CF large-scale MIMO architecture is completed under the complex non-ideal CSI conditions that simultaneously consider the cascade link and segment link estimation errors. At the same time, the embodiment of the present disclosure proposes to optimize the selection of a limited number of RIS electromagnetic units that can work normally through the RIS selection matrix under power-constrained conditions, further improving the RIS's ability to regulate the electromagnetic environment, thereby improving the communication rate under power-constrained conditions. Moreover, the embodiment of the present disclosure uses active RIS to amplify the amplitude of the reflected signal to effectively suppress the "double fading" effect that is difficult to solve in the existing passive RIS-assisted scenario. Compared with the existing solutions of only optimizing beamforming and only optimizing beamforming and RIS electromagnetic unit phase shift and amplitude, this technology can more effectively improve the system communication rate in a more complex environment.

[0039] Next, we will refer to Figures 1 to 5 The various steps of the above-mentioned RIS-enabled CF network downlink communication method under complex non-ideal CSI and power-limited conditions in this example implementation are described in more detail.

[0040] In step S101, based on the passive RIS assisted scenario and the active RIS assisted scenario, respectively, a complex non-ideal CSI and a CF network model under power limitation under the assistance of passive RIS or a complex non-ideal CSI and a CF network model under power limitation under the assistance of active RIS is constructed, wherein the active RIS (i.e., the active RIS assisted scenario) can adjust the amplitude of the signal compared with the passive RIS (i.e., the passive RIS assisted scenario) to effectively overcome the "double fading" effect in the passive RIS assisted scenario; for the passive RIS and active RIS scenarios, a transmission model of the receiving end user is constructed according to the communication network model.

[0041] In the passive RIS-assisted scenario, a CF network model with complex non-ideal CSI and power limitation under passive RIS assistance is constructed.

[0042] Specifically, a CF massive MIMO system model is constructed with B access points APs, 1 passive RIS, and T single-antenna users. Each AP is equipped with M antennas, and N electromagnetic units are set on the RIS. However, due to the limited power resources of the system, it is impossible to make all electromagnetic units of the RIS work at the same time. The power currently provided can only meet the normal operation of a maximum of V electromagnetic units on the RIS, where V <N。

[0043] In the active RIS-assisted scenario, a CF network model with complex non-ideal CSI and power limitation under active RIS assistance is constructed.

[0044] Specifically, a CF massive MIMO system model is constructed with B access points APs, 1 active RIS, and T single-antenna users. Each AP is equipped with M antennas, and N electromagnetic units are set on the RIS. Similar to the passive RIS-assisted scenario, due to the limited power resources of the system, only V electromagnetic units can work normally, where V <N。

[0045] In step S102, a transmission model of a wireless signal at a receiving end user is determined according to a communication network model.

[0046] Specifically, in the communication network model, the power resources of the system are limited, which is insufficient to meet the basic circuit consumption of all electromagnetic units on the RIS, so that all electromagnetic units cannot work normally. Therefore, the present application introduces a RIS electromagnetic unit selection matrix to select electromagnetic units that can work normally under limited power, thereby constructing a wireless signal transmission model in the power-limited system. In the actual scenario where the direct link is blocked by an obstacle, preferably, the transmission model expression of the wireless signal at the receiving end user is as follows:

[0047]

[0048] Among them, x b is the signal transmitted by the bth AP, is the channel state matrix from RIS to the t-th user, is the channel state matrix from the bth AP to the RIS, n t The power at the receiving end user is The background Gaussian noise, is the reflection coefficient matrix of RIS, where β n and θ n are respectively the modulation of the signal amplitude by the nth reflection unit on the RIS and the phase shift of the reflection unit. For a passive RIS, A is the RIS electromagnetic unit selection matrix, A satisfies trace(A)=V, where Z[i,i] represents the i-th row and i-th column element of the matrix Z, and trace(Z) represents the trace of the matrix Z.

[0049] In order to obtain the signal-to-interference-noise ratio of the receiving end user according to the transmission model, the transmission model of the receiving end user can be rewritten as:

[0050]

[0051] in, For the beamforming of the b-th AP to the t-th user, satisfy s t is the signal information received by the tth user, satisfying satisfy Where diag(z) represents a diagonal matrix whose diagonal elements are composed of z, and the remaining symbols are defined as follows:

[0052] According to the user's transmission model, the signal-to-interference-noise ratio of the user at the receiving end can be expressed as:

[0053]

[0054] In the scenario assisted by active RIS, the transmission model of the wireless signal at the receiving end user is determined according to the communication network model. In the actual scenario where the direct link is blocked by an obstacle, preferably, the transmission model expression of the wireless signal at the receiving end user is as follows:

[0055]

[0056] Among them, n I The power generated by the active RIS is thermal noise.

[0057] In order to obtain the signal-to-interference-noise ratio of the receiving end user according to the transmission model, the transmission model of the receiving end user can be rewritten as:

[0058]

[0059] According to the user's transmission model, the signal-to-interference-noise ratio of the user at the receiving end can be expressed as:

[0060]

[0061] In this communication network model, due to factors such as quantization error, the system has estimation errors in estimating CSI, and the CSI of the cascaded channel cannot be perfectly captured. At the same time, in order to reduce the impact of thermal noise generated by active RIS on user receiving performance and accurately meet the maximum power limit that active RIS can carry, the errors of the two channels, RIS-user and AP-RIS, need to be estimated separately. Therefore, the present application proposes a technology for improving user communication rate under complex non-ideal CSI containing multiple channel estimation errors.

[0062] In step S103, a model for maximizing the worst user communication rate is designed according to the transmission model of the receiving end user and the signal to interference noise ratio of the receiving end user;

[0063] In the passive RIS-assisted scenario, according to the signal-to-interference-noise ratio of the receiving end user, the communication rate of the receiving end user can be expressed as:

[0064] R1 t =ln(1+S1 t )

[0065] Due to the influence of quantization error, the system cannot perfectly capture the CSI of the cascaded channel. Preferably, the AP-RIS-user channel model can be expressed as:

[0066]

[0067] in, represents the estimated channel, ΔH t Represents the error channel. This application studies the bounded CSI error, that is, the boundary value of the channel error can be characterized. Specifically, its error channel satisfies the following expression:

[0068]

[0069] in, Represents the radius of the uncertainty region of the channel error of the AP-RIS-user channel.

[0070] According to the communication rate of the receiving user and the channel model under bounded CSI error, the optimization problem of maximizing the worst user communication rate in the passive RIS-assisted scenario is shown as P0:

[0071]

[0072] C3:A∈{0,1},trace(A)=V.

[0073] Where C0 is the bounded CSI error constraint in the passive RIS-assisted scenario; C1 represents the maximum power constraint of each AP, where P AP is the maximum transmit power of each AP, is a diagonal matrix used to select the beamforming of the bth AP, which satisfies C2 is the constant modulus constraint of the passive RIS electromagnetic unit; C3 indicates that the selection matrix A is a 0-1 matrix, and the number of electromagnetic units selected for normal operation is V.

[0074] In the active RIS-assisted scenario, according to the signal-to-interference-noise ratio of the receiving end user, the communication rate of the receiving end user can be expressed as:

[0075] R2 t =ln(1+S2 t )

[0076] The amplification circuit of active RIS generates thermal noise, which affects the transmission quality of the receiving end user. At the same time, its power carrying capacity is limited, which causes the system to only estimate the quantization error of the cascade channel and cannot accurately establish the optimization problem model. Therefore, in the scenario assisted by active RIS, it is necessary not only to estimate the error of the AP-RIS-user channel, but also to estimate the errors of the RIS-user and AP-RIS channels separately. The channel models can be expressed as follows:

[0077]

[0078] in, and denote the estimated channels of the AP-RIS link and the RIS-user link, ΔH AI and They represent the error channels of the AP-RIS link and the RIS-user link respectively. According to the bounded CSI error, specifically, their error channels satisfy the following expression:

[0079]

[0080] in, and They represent the radius of the uncertainty region of the channel error of the AP-RIS link and the RIS-user link respectively.

[0081] According to the communication rate of the receiving end user and the channel model under bounded CSI error, the optimization problem of maximizing the worst user communication rate in the active RIS assisted scenario is shown as P0':

[0082]

[0083] K4:A∈{0,1},trace(A)=V.

[0084] Where K0 represents the bounded CSI error constraint in the active RIS assisted scenario; K1 represents the maximum power constraint of each AP; K2 is the maximum power constraint of active RIS, where P I represents the maximum power that the active RIS can carry; K3 represents the constraint of each electromagnetic unit of the active RIS on the signal modulation amplitude capability, where: Indicates the maximum magnification of each electromagnetic unit; K4 indicates that the selection matrix A is a 0-1 matrix, and the number of electromagnetic units selected for normal operation is V.

[0085] However, the optimization problems P0 and P0' for maximizing the worst user communication rate constructed in the passive RIS-assisted scenario and the active RIS-assisted scenario are both complex non-convex optimization problems with highly coupled optimization variables. The present application combines the alternating algorithm with semidefinite relaxation (Semidefinite Relaxation, SDR), Dinkelbach method, successive convex approximation (Successive Convex Approximation, SCA), "S-process" and Schur complement methods to transform the non-convex optimization problem into a convex problem, thereby effectively solving the constructed complex non-convex optimization problem.

[0086] In step S104, since the maximization of the worst user communication rate models in the two scenarios in step S103 are both multi-variable coupled optimization problems, an alternating optimization method is used to solve the optimization problems in the two scenarios respectively.

[0087] According to the main idea of ​​the alternating optimization method, the model of maximizing the worst user communication rate in the passive RIS-assisted scenario is decomposed, and the beamforming w and RIS electromagnetic unit coefficients are used respectively. The RIS electromagnetic unit selects two sub-optimization models with matrix A as the optimization variable, and obtains the maximum communication rate solution of the original model under the framework of the alternating algorithm.

[0088] The electromagnetic unit parameters and selection matrix parameters of the passive RIS are set, and the beamforming optimization strategy in this scenario is obtained by combining the maximization worst user communication rate model in the passive RIS-assisted scenario.

[0089] Specifically, according to the alternating optimization idea, the optimization problem variable of maximizing the worst user communication rate in the passive RIS-assisted scenario is and A (l) Treated as a constant, and set the initial passive RIS electromagnetic unit parameters and the selection matrix A (l) , l is the number of iterations, solving the variable

[0090] Preferably, in a given and A (l) Under the conditions of The subproblem can be expressed as P1:

[0091]

[0092] Next, due to the monotonicity of the logarithmic function, P1 can be equivalent to the optimization problem of maximizing the worst user signal to interference and noise ratio. In order to further solve this optimization problem, P1 can be transformed into P2:

[0093]

[0094] Among them, C4 represents the lower limit of the signal-to-interference-to-noise ratio of all users in the system. U , t U is an introduced auxiliary variable.

[0095] However, according to S1 t The specific expression of C4 is a complex fractional non-convex constraint condition. Therefore, C4 needs to be converted into an integer convex constraint condition. Specifically, C4 can be converted into:

[0096]

[0097] Among them, β t is the introduced auxiliary variable.

[0098] At this point, C5 is still a non-convex constraint. Therefore, using the idea of ​​the SCA method, C5 can be converted into a convex constraint. According to the lemma “|x| 2 ≥2Re{x (n) x H}-x (n) x (n),H ”, where Re{z} means taking the real part of z, x (n) Represents the optimal value in the previous iteration, C5 can be converted to:

[0099]

[0100] Will Substitute into C7 and use Among them, vec(Z) represents the straightening operation of the matrix Z. After a series of derivations, C7 can be transformed into C8:

[0101]

[0102] in,

[0103]

[0104] In order to further solve the optimization problem, this application uses the lemma "S-process" to transform the estimated error in the constraint into a solvable form. The specific content of the "S-process" is as follows:

[0105] Definition: f k (x) = xA k x H +2Re{b k x H}+C k ,k=1,...,K, where If f k≠1 (x)≥0, there exists μ k ≥0,k=2,...,K, if and only if the following conditions are met:

[0106]

[0107] f1(x)≥0 always holds.

[0108] At this time, C0 can be equivalently expressed as:

[0109]

[0110] According to Lemma (1), by introducing the auxiliary variable μ H =[μ H,1 ,μ H,2 ,...,μ H,T ],μ H,t ≥0, C8 and C9 can be transformed into the following linear inequalities:

[0111]

[0112] Furthermore, the constraint C6 is transformed into the following form using “Schur complement”:

[0113]

[0114] Among them, K t =[θ H H t W -t ],W -t =[w1,w2,...,w t-1 ,w t+1 ,...,w T ].

[0115] In order to transform the bounded CSI error in C11 into a solvable form, the following important lemma is used:

[0116] For the matrix W = W H and If and only if the following inequality is satisfied:

[0117]

[0118] in, Linear Inequality Heng established.

[0119] Will Substituting into C11, we get C12:

[0120]

[0121] in, At this time, according to Lemma (2), the auxiliary variable τ is introduced H =[τ H,1 ,τ H,2 ,...,τ H,T ],τ H,t ≥0, C9 and C12 can be transformed into the following linear inequalities:

[0122]

[0123] Furthermore, the optimization problem P2 under bounded CSI error estimation is transformed into the following optimization problem P3, and the specific expression is as follows:

[0124]

[0125] stC1,C10,C13,

[0126]

[0127] P3 can use binary search t U The maximum feasible solution of , that is, it is regarded as a feasible point search problem to solve, and finally P3 can be transformed into P4:

[0128]

[0129] stC1,C10,C13,

[0130]

[0131] By solving the above formula, we can get

[0132] According to the beamforming optimization strategy of passive RIS and combined with the model of maximizing the worst user communication rate, the optimization strategy of electromagnetic unit coefficients and selection matrix of passive RIS is obtained.

[0133] Specifically, in a given Under the condition, this subproblem can be expressed as P5:

[0134]

[0135] C3:A∈{0,1},trace(A)=V.

[0136] and optimization The processing method is similar. In order to further solve the optimization problem, P5 is transformed into P6:

[0137]

[0138] C3:A∈{0,1},trace(A)=V.

[0139]

[0140] In this subproblem, since the electromagnetic unit coefficients of RIS are included The two optimization variables, C4 and the selection matrix A, need to be decoupled to be optimized simultaneously. At the same time, C4 is a complex fractional non-convex constraint. Therefore, this application combines the Dinkelbach method and the SDR method to solve this complex problem.

[0141] Specific, definition To make the RIS electromagnetic coefficient Decoupled from A, C2 can be transformed into the following form:

[0142]

[0143] For A∈{0,1} in C3, it can be converted into an equivalent form:

[0144]

[0145] Among them, C15 is a non-convex constraint. Using the first-order Taylor expansion and penalty function method, C15 can be transformed into a convex constraint C17:

[0146]

[0147] Among them, S n A is the introduced auxiliary variable. (n) is the optimal selection matrix in the previous iteration.

[0148] For the complex fractional non-convex constraint C4, according to the core idea of ​​Dinkelbach method, the auxiliary parameter ν is introduced, and C4 can be transformed into the following form:

[0149]

[0150] Among them, t′ U is an introduced auxiliary variable.

[0151] Will Substitute into C18 and use After a series of derivations, C18 can be converted to C19:

[0152]

[0153] in,

[0154]

[0155] h P =vec H (H t ).

[0156] According to Lemma (1), by introducing the auxiliary variable C9 and C19 can be transformed into the following linear inequalities:

[0157]

[0158] At this point, P6 can be transformed into a convex optimization problem P7, which is expressed as follows:

[0159]

[0160] stC14,C16~C17,C20,

[0161]

[0162] in, is the penalty factor. By solving P7, optimizing the electromagnetic unit coefficients of the passive RIS and selecting the matrix optimization model, and performing Gaussian random decomposition on the optimization result Ψ, we can obtain and A (l+1) , that is, the electromagnetic unit coefficients and selection matrix optimization strategy of passive RIS.

[0163] Based on the electromagnetic unit and selection matrix optimization strategy of the passive RIS, the electromagnetic unit parameters and selection matrix parameters of the passive RIS are iteratively updated, and the above steps are repeated until the variables of the passive RIS beamforming optimization strategy and the passive RIS electromagnetic unit and selection matrix optimization strategy converge, and a solution for maximizing the worst user communication rate is obtained.

[0164] Specifically, update the iteration number l→l+1, based on the optimization solution of the electromagnetic unit and selection matrix optimization model of the passive RIS, update the electromagnetic unit parameters and selection matrix of the passive RIS, and repeat the above steps until the parameters meet the condition Where ε is the convergence threshold, and the final maximized worst user communication rate solution in the passive RIS-assisted scenario can be obtained.

[0165] According to the main idea of ​​the alternating optimization method, the model of maximizing the worst user communication rate in the active RIS-assisted scenario is decomposed, and the beamforming w and RIS electromagnetic unit coefficients are used respectively. The RIS electromagnetic unit selects two sub-optimization models with matrix A as the optimization variable, and obtains the maximum communication rate solution of the original model under the framework of the alternating algorithm.

[0166] The electromagnetic unit parameters and selection matrix parameters of active RIS are set, and the beamforming optimization strategy in this scenario is obtained by combining the maximization worst user communication rate model in the active RIS assisted scenario.

[0167] Specifically, according to the alternating optimization idea, the optimization problem variables for maximizing the worst user communication rate in the active RIS-assisted scenario are and A (l) Treated as a constant, and set the initial active RIS electromagnetic unit parameters and the selection matrix A (l) , solving for the variable

[0168] Preferably, in a given and A (l) Under the conditions of The subproblem can be expressed as P1':

[0169]

[0170] Next, due to the monotonicity of the logarithmic function, P1' can be equivalent to the optimization problem of maximizing the worst user signal to interference and noise ratio. In order to further solve this optimization problem, P1' can be transformed into P2':

[0171]

[0172] Among them, K5 represents the lower limit of the signal-to-interference-to-noise ratio of all users in the system. U .

[0173] However, according to S2 t The specific expression of K5 is a complex fractional non-convex constraint condition. Therefore, K5 needs to be converted into an integer convex constraint condition. Specifically, K5 can be converted into:

[0174]

[0175] At this point, K6 is still a non-convex constraint. The method of converting it into an integral convex constraint is the same as the method of converting C5 into an integral convex constraint in the passive RIS scenario. After a series of derivations, K6 can be converted to K8:

[0176]

[0177] in,

[0178]

[0179] In order to further solve the optimization problem in the active RIS-assisted scenario, this application uses Lemma (1) to transform the estimated error in the constraint into a solvable form. Specifically, The equivalent expression is:

[0180]

[0181] According to Lemma (1), by introducing the auxiliary variable μ AH =[μ AH,1 ,μ AH,2 ,...,μ AH,T ],μ AH,t ≥0, K8 and K9 can be transformed into the following linear inequalities:

[0182]

[0183] Furthermore, in the active RIS-assisted scenario, the constraint K7 is transformed into the following form using the “Schur complement”:

[0184]

[0185] in,

[0186] This application uses Lemma (2) to transform the bounded CSI errors of the cascaded channel and the independent channel RIS-user in K11 into a solvable form. Specifically, Substituting into K11, we can get K12:

[0187]

[0188] in, B and D are selection matrices, respectively satisfying B = [I T-1 ,0 (T-1)×NR ],D=[0 NR×(T-1) ,I NR ], ΔH IUt satisfy

[0189] According to Lemma (2), we introduce the auxiliary variable τ AH =[τ AH,1 ,τ AH,2 ,...,τ AH,T ],τ AH,t ≥0, and τ IU =[τ IU,1 ,τ IU,2 ,...,τ IU,T ],τ IU,t ≥0, K0 and K12 can be transformed into the following linear inequality:

[0190]

[0191] In the scenario assisted by active RIS, the active RIS has a maximum power carrying constraint K2. Using Lemma (2), the bounded CSI error of the independent channel AP-RIS can be transformed into a solvable form.

[0192] Specifically, use "Schur complement" to transform constraint K2 into the following form:

[0193]

[0194] in, θ = diag(θ).

[0195] Specifically, Substituting into K14, we can get K15:

[0196]

[0197] in, H AI satisfy

[0198] Further, according to Lemma (2), we introduce the auxiliary variable τ AI ≥0, K0 And K15 can be transformed into the following linear inequality:

[0199]

[0200] The optimization problem P2' under bounded CSI error estimation is transformed into the following optimization problem P3', and the specific expression is as follows:

[0201]

[0202] stK10,K13,K16,

[0203]

[0204] τ AI ≥0,

[0205]

[0206] P3' can use binary search t U The maximum feasible solution of , that is, it is regarded as a problem of searching for feasible points, and finally P3' can be transformed into P4':

[0207]

[0208] stK10,K13,K16,

[0209]

[0210] τ AI ≥0,

[0211]

[0212] By solving the above formula, we can get

[0213] According to the beamforming optimization strategy of active RIS and combined with the model of maximizing the worst user communication rate, the optimization strategy of electromagnetic unit coefficients and selection matrix of active RIS is obtained.

[0214] Specifically, in a given Under the condition, this subproblem can be expressed as P5':

[0215]

[0216] K4:A∈{0,1},trace(A)=V.

[0217] and optimization The processing method is similar. In order to further solve the optimization problem, P5' is transformed into P6':

[0218]

[0219] K4:A∈{0,1},trace(A)=V,

[0220]

[0221] In this subproblem, since the electromagnetic unit coefficients of RIS are included The two optimization variables, y and the selection matrix A, need to be decoupled to be able to optimize them simultaneously. At the same time, K5 is a complex fractional non-convex constraint. Therefore, this application combines the Dinkelbach method and the SDR method to solve this complex problem.

[0222] Specific, definition To make the RIS electromagnetic coefficient Decoupled from A, K3 can be transformed into the following form:

[0223]

[0224] For A∈{0,1} in K4, it can be converted into an equivalent form:

[0225]

[0226] Among them, K18 is a non-convex constraint. Using the first-order Taylor expansion and penalty function method, K18 can be transformed into a convex constraint K20:

[0227]

[0228] Among them, S n A is the introduced auxiliary variable. (n) is the optimal selection matrix in the previous iteration.

[0229] Through the above process, the RIS electromagnetic coefficient After decoupling from A, further, Substitute into K2 and use K2 can be transformed into K21:

[0230]

[0231] in, X⊙Y represents the Hadamard product of matrix X and matrix Y. In order to transform the bounded CSI estimation error in K21 into a solvable form, we use Lemma (1). Specifically, Can be equivalent to Introducing auxiliary variables According to Lemma (1), in K0 and K21 can be expressed as the following linear inequality:

[0232]

[0233] For the complex fractional non-convex constraint K5, according to the core idea of ​​Dinkelbach method, the auxiliary parameter ν is introduced, and K5 can be transformed into the following form:

[0234]

[0235] Among them, t′ U is an introduced auxiliary variable.

[0236] Will and Substitute into K23 and use And after deduction, K23 can be converted to K24:

[0237]

[0238] Where Δh Q =[vec H (ΔH t ),vec H (ΔH IUt )],

[0239]

[0240] According to Lemma (1), by introducing the auxiliary variable K0 And K24 can be transformed into the following linear inequality:

[0241]

[0242] At this point, P6' can be transformed into a convex optimization problem P7', which is expressed as follows:

[0243]

[0244] stK17,K19~K20,K22,K25,

[0245]

[0246] in, is the penalty factor. By solving P7', optimizing the electromagnetic unit coefficients of the active RIS and selecting the matrix optimization model, and performing Gaussian random decomposition on the optimization result Ψ, we can obtain and A (l+1) , that is, the electromagnetic unit coefficients and selection matrix optimization strategy of active RIS.

[0247] Based on the electromagnetic unit and selection matrix optimization strategy of active RIS, the electromagnetic unit parameters and selection matrix parameters of active RIS are iteratively updated, and the above steps are repeated until the variables of the beamforming optimization strategy of active RIS and the electromagnetic unit and selection matrix optimization strategy of active RIS converge, and a scheme for maximizing the worst user communication rate is obtained.

[0248] Specifically, update the iteration number l→l+1, based on the optimization solution of the electromagnetic unit and selection matrix optimization model of the active RIS, update the electromagnetic unit parameters and selection matrix parameters of the active RIS, and repeat the above steps until the parameters meet the condition The final maximized worst user communication rate solution in the active RIS-assisted scenario can be obtained.

[0249] The effects of the present application are described in detail below in conjunction with simulation.

[0250] This application simulates the RIS-enabled CF network downlink communication method under complex non-ideal CSI and power-constrained conditions to verify the superiority of the solution of this application. The specific steps are as follows: The basic parameters set are the coordinates of APs, RIS and users, where the number of APs B=4, their coordinates are (-60,0,4), (-30,0,4), (30,0,4), (60,0,4), the coordinates of RIS are (0,30,8), the number of users is T=4, they are randomly generated on the horizontal line Y=25m, within the range of X∈[-30,30], and the height is 0 meters. Each AP has 2 antennas and each user has 1 receiving antenna. When the reference distance is 1 meter, the large-scale fading coefficient is -30dB, and the background noise power is This application uses the Rice fading model for channel modeling, with a Rice factor of 5 and a path loss exponent of 2.2 associated with RIS. Thermal noise of active RIS The power budget P of each RIS amplifier circuit I =10dBm. When solving the optimization problem, the threshold for ending the algorithm is ε = 10 -3 .

[0251] like Figure 3 As shown, a convergence comparison diagram of the worst user communication rate under different numbers of RIS electromagnetic units in active RIS and passive RIS assisted scenarios is shown under the parameters of the exemplary embodiment. It can be seen from the figure that in the scenarios of active RIS and passive RIS assisted, the scheme proposed in this application can effectively improve the communication rate of the system. As the number of RIS units increases, the convergence curve will converge to a higher communication rate. This is because the more electromagnetic units N, the higher the degree of freedom of signal transmission, the stronger the control of RIS on the wireless electromagnetic channel, and the more significantly the user's signal reception performance can be enhanced.

[0252] like Figure 4 As shown, it is shown that under the parameters of the exemplary embodiment, as the AP transmit power P AP The increase of , the communication performance improvement of the worst user in the active RIS and passive RIS assisted scenarios, and at the same time, Figure 4This technology is also compared with the existing solution that only optimizes beamforming. As can be seen from the figure, the wireless communication solution proposed in this application can effectively improve the communication rate of the system. As can be seen from the figure, active RIS can more effectively improve the user communication rate than passive RIS. This is because the CF system is a system with a wide range and long transmission distance. In this system, the "double fading" effect of passive RIS is obvious, which increases the attenuation of the reflection link of passive RIS and cannot effectively assist the communication in this system. Active RIS can effectively overcome the "double fading" effect of long-distance transmission by adjusting the amplitude of the signal, so that RIS can better assist the communication in the CF network.

[0253] like Figure 5 As shown, it is shown that under the parameters of the exemplary embodiment, as the uncertainty of the channel error increases, the communication performance improvement of the worst user in the active RIS and passive RIS assisted scenarios is improved, and at the same time, Figure 5 The present technique is also compared with existing schemes that optimize only beamforming and only beamforming and RIS electromagnetic unit phase shift. Figure 5 It can be seen that as the uncertainty of the channel error increases, the communication rate decreases accordingly. This is because the increase in uncertainty will lead to an increase in the channel error. The technology proposed in this application has less attenuation and better robustness than the existing solution. At the same time, the wireless communication solution proposed in this application can effectively improve the communication rate of the system compared to the existing solution of only optimizing beamforming and only optimizing beamforming and RIS electromagnetic unit phase shift.

[0254] Through the above-mentioned RIS-enabled CF network downlink communication method under complex non-ideal CSI and power-constrained conditions, on the one hand, high-quality communication under the CF large-scale MIMO architecture is achieved under the complex non-ideal CSI situation that simultaneously considers the cascade link and segment link estimation errors. At the same time, the embodiment of the present disclosure proposes to optimize the selection of a limited number of RIS electromagnetic units that can work normally through the RIS selection matrix under power-constrained conditions, so as to further improve the RIS's ability to regulate the electromagnetic environment. On the other hand, the embodiment of the present disclosure utilizes the active RIS to amplify the amplitude of the reflected signal, so as to effectively suppress the "double fading" effect that is difficult to solve in the existing passive RIS-assisted scenario. Compared with the existing solutions of only optimizing beamforming and only optimizing beamforming and RIS electromagnetic unit phase shift and amplitude, this technology can more effectively improve the system communication rate in a more complex environment.

[0255] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should be included in the protection scope of the present invention.

Claims

1. A RIS-enabled CF network downlink communication method under complex non-ideal CSI and power-limited conditions, characterized in that: include: Based on the RIS-assisted scenario, a CF network model under complex non-ideal CSI and power limitation is constructed, and a transmission model of a receiving end user is constructed according to the RIS-assisted complex non-ideal CSI and power limitation CF network model; Obtaining a signal to interference and noise ratio of the receiving end user according to a transmission model of the receiving end user; Designing a model for maximizing the worst user communication rate according to the transmission model of the receiving end user and the signal to interference noise ratio of the receiving end user; The alternating algorithm is used to solve the optimization problem of the maximization worst user communication rate model to obtain the final information transmission solution.

2. According to claim 1, a RIS-enabled CF network downlink communication method under complex non-ideal CSI and power-limited conditions, characterized in that: The RIS assisted scenario is a passive RIS assisted scenario or an active RIS assisted scenario.

3. According to claim 1, a RIS-enabled CF network downlink communication method under complex non-ideal CSI and power-limited conditions, characterized in that: The RIS-assisted complex non-ideal CSI and CF network model under power limitation is a passive RIS-assisted complex non-ideal CSI and CF network model under power limitation or an active RIS-assisted complex non-ideal CSI and CF network model under power limitation.

4. According to claim 1, a RIS-enabled CF network downlink communication method under complex non-ideal CSI and power-limited conditions, characterized in that: The step of using an alternating algorithm to solve the optimization problem of maximizing the worst user communication rate model to obtain a final information transmission scheme comprises: Decomposing the maximization worst user communication rate model in the RIS-assisted scenario, and optimizing the maximization worst user communication rate model using two sub-optimization models with beamforming w, RIS electromagnetic unit coefficient φ, and RIS electromagnetic unit selection matrix A as optimization variables; The final information transmission scheme is obtained within the framework of the alternating algorithm.

5. According to claim 4, a RIS-enabled CF network downlink communication method under complex non-ideal CSI and power-limited conditions, characterized in that: The step of decomposing the maximization worst user communication rate model in the RIS-assisted scenario, and optimizing the maximization worst user communication rate model using two sub-optimization models with beamforming w, RIS electromagnetic unit coefficient φ, and RIS electromagnetic unit selection matrix A as optimization variables, includes: The electromagnetic unit parameters and selection matrix parameters of RIS are set, and the beamforming optimization strategy in this scenario is obtained by combining the maximization worst user communication rate model in the RIS-assisted scenario; According to the beamforming optimization strategy of RIS and combined with the model of maximizing the worst user communication rate, the optimization strategy of electromagnetic unit coefficients and selection matrix of RIS is obtained; Based on the electromagnetic unit and selection matrix optimization strategy of RIS, the electromagnetic unit parameters and selection matrix parameters of RIS are iteratively updated, and the above steps are repeated until the variables of the beamforming optimization strategy of RIS and the electromagnetic unit and selection matrix optimization strategy of RIS converge, and a solution for maximizing the worst user communication rate is obtained.

6. A computer system, characterized in that include: One or more processors, and a computer-readable storage medium for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement the method of claim 1.

7. A computer-readable storage medium, characterized in that Computer executable instructions are stored, and when the instructions are executed, they are used to implement the method of claim 1.

8. A computer program product, characterized in that The invention comprises computer executable instructions, which are used to implement the method of claim 1 when being executed.

Images