Intelligent metasurface assisted robust transmission method for terahertz communication system

Abstract

The application belongs to the field of terahertz wireless communication transmission, and relates to a kind of intelligent metasurface assisted terahertz communication system robust transmission method, including constructing RIS assisted terahertz communication system and initialization;According to the minimum rate constraint of user, base station transmission power constraint, channel uncertainty constraint and RIS discrete phase shift constraint, the total energy efficiency of RIS assisted terahertz communication system is used as the optimization target to construct a robust beam forming model;Robust beam forming model is converted into deterministic problem model using S-Procedure;Solving based on block coordinate descent alternating iteration criterion;Compared with the method under perfect channel state information, the application scheme has higher robustness, improves the energy efficiency and received signal quality of terahertz communication network, reduces the user transmission interruption probability, and is suitable for RIS that cannot provide continuous phase shift regulation capability, and is more suitable for practical application.

Description

Technical Field

[0001] This invention belongs to the field of terahertz wireless communication transmission and relates to a robust transmission method for a smart metasurface-assisted terahertz communication system. Background Technology

[0002] Terahertz technology has been proposed as a potential 6G technology, promising to address the issues of spectrum scarcity and capacity limitations in current wireless systems. However, terahertz communication quality is susceptible to propagation loss and molecular absorption, which significantly limits its transmission distance and communication capacity. Furthermore, obstacles in the propagation environment can directly block communication. RIS (Reflective Radiation Array) is a candidate 6G technology, considered a new communication paradigm capable of reconfiguring the wireless channel environment. This technology addresses obstacle-avoiding communication problems with low power consumption and by providing a direct path. As described in the RIS technology white paper, RIS is a uniform array plane composed of many low-power, passive, reconfigurable reflective elements. By independently adjusting the reflective elements, the reflection path of the received signal can be changed, effectively improving the capacity and coverage performance of wireless networks.

[0003] In practical communication systems, the communication link between transceivers suffers from quantization errors, feedback delays, and estimation errors, making it extremely difficult for base stations to obtain accurate channel state information. Furthermore, due to the limited processing capabilities of low-power, passively transmitted RIS (Resource-Assisted Terahertz) systems, and the impact of quantization errors and channel delays in terahertz communication systems, designing transmission algorithms requires not only improving system throughput but also considering energy consumption. Therefore, energy efficiency optimization in RIS-assisted terahertz communication network scenarios with channel uncertainties is highly challenging. Summary of the Invention

[0004] The purpose of this invention is to provide a robust transmission method for a smart metasurface-assisted terahertz communication system. Considering user minimum rate constraints, base station transmit power constraints, channel uncertainty constraints, and RIS discrete phase shift constraints, the method aims to maximize system energy efficiency. A network model and a system model are established for the RIS-assisted terahertz communication network. The S-Procedure is used to transform the uncertain non-convex optimization problem into a deterministic non-convex optimization problem; the Tinkelbach method is used to transform the fractional programming problem into an equivalent problem in subtraction form; and the non-convex problem is transformed into an equivalent convex problem for solution based on variable substitution, continuous convex approximation theory, block coordinate descent method, and semidefinite programming method.

[0005] The specific solution provided by this invention includes the following steps:

[0006] S1. Construct and initialize a RIS-assisted terahertz communication system, the system comprising K users, a RIS with N reflective elements, and a terahertz base station with M antennas;

[0007] S2. Based on the user minimum rate constraint, base station transmit power constraint, channel uncertainty constraint and RIS discrete phase shift constraint, a robust beamforming model is constructed with the optimization objective of maximizing the total energy efficiency of the RIS-assisted terahertz communication system.

[0008] S3. Use the S-Procedure to convert the robust beamforming model into a deterministic problem model;

[0009] S4. Calculate the overall energy efficiency of the system;

[0010] S5. With the phase shift matrix of the fixed RIS fixed, calculate the first... Beamforming matrix for each user;

[0011] S6. Fix the beamforming matrix of the k-th user and calculate the phase shift matrix of RIS;

[0012] S7. Determine whether the current system overall energy efficiency has converged. If yes, output the system optimal energy efficiency, base station beamforming matrix, and RIS continuous phase shift matrix, and then execute step S9; otherwise, execute step S8.

[0013] S8. Determine if the current iteration count is greater than the maximum iteration count. If yes, output the system's optimal energy efficiency, base station beamforming matrix, and RIS continuous phase shift matrix, and then execute step S9. If not, increment the iteration count by 1 and return to step S4.

[0014] S9. Calculate the discretized phase shift matrix of RIS.

[0015] Furthermore, the robust beamforming model constructed in step S2 is expressed as follows:

[0016]

[0017] in, Let represent the transmission rate of the k-th user during communication, considering the channel uncertainty problem. This represents the signal-to-interference-plus-noise ratio (SIR) of the k-th user during communication, considering channel uncertainty. This represents the total throughput of a RIS-assisted terahertz communication system considering channel uncertainty. C1 represents the total power consumption of the RIS-assisted terahertz communication system; C1 is the maximum transmit power constraint of the terahertz base station, where... This represents the beamforming vector from the terahertz base station to the k-th user. This represents the total power consumption of a terahertz base station for beamforming K users. C1 represents the maximum transmission power threshold of the terahertz base station; C2 is the discrete phase shift constraint of RIS, where Representing the first RIS The amplitude and phase reflection coefficients of each array element Indicates that RIS contains A discrete phase shift set of n elements; There is an estimation error in the channel. At that time, the user's minimum transmission rate constraint, C4 represents the minimum transmission rate threshold; C5 represents the channel uncertainty set constraint. This indicates the channel uncertainty problem.

[0018] Furthermore, step S3 uses the S-Procedure to convert the robust beamforming model into a deterministic problem model, which is expressed as:

[0019]

[0020] Among these, Tinkelbach transformed the fractional objective function into an equivalent subtraction of two terms, namely, , Let represent the transmission rate of the k-th user in the deterministic problem model during the communication process. An auxiliary variable representing the system's energy efficiency. For the set of auxiliary variables, This indicates the total power consumption of the RIS-assisted terahertz communication system. This is the conjugate transpose of the matrix; The maximum transmit power constraint for terahertz base stations, where Indicates the first Beamforming matrix for each user Let be the beamforming vector from the terahertz base station to the k-th user. This represents the total power consumption of the base station for beamforming K users. Tr() represents the maximum transmission power threshold of the terahertz base station; C2 represents the trace of the matrix; C2 is the discrete phase shift constraint of RIS. C1 represents the minimum transmission rate constraint for users based on the continuous convex approximation and Taylor series expansion; C2 represents the channel uncertainty set constraint. For robust constraints, , , , and Both represent auxiliary variables. and express and The previous iteration, Represents the identity matrix. Indicates intermediate variables. ,in , and These are the antenna gains for the transmitting and receiving ends, respectively. This represents the path loss compensation factor. For Kronecker product, Represents RIS Phase shift of each array element, , and They represent the first, second, and third RIS, respectively. The amplitude and phase reflection coefficients of each array element Indicates from the base station to the... Cascaded channel estimates for individual users Indicates the first The upper bound of the cascaded channel estimation error for each user, vec() denotes matrix vectorization. Indicates intermediate variables. , C6 represents the background noise power at the receiver; C6 is the rank-one constraint of the beamforming matrix. C is the rank of the matrix; C7 is the rank-one constraint of the RIS phase shift matrix, where This represents the RIS phase shift matrix.

[0021] Furthermore, the formula for calculating the overall system energy efficiency in step S4 is as follows:

[0022]

[0023] in, This represents the total system energy efficiency at the l-th iteration. This represents the lower limit of the system's energy efficiency. This represents the upper limit of the system's energy efficiency; the Tinkelbach function is used. Perform a judgment update when hour, ,when hour, .

[0024] Furthermore, in step S5, the phase shift matrix of RIS is fixed to obtain information about the first... Beamforming matrix for individual users and auxiliary variable set The sub-optimization problem, the solution process includes:

[0025] S51. Regarding the first Beamforming matrix for individual users and auxiliary variable set The sub-optimization problem is represented as:

[0026]

[0027] in, Let represent the transmission rate of the k-th user in the deterministic problem model during the communication process. An auxiliary variable representing the system's energy efficiency. This indicates the total power consumption of the RIS-assisted terahertz communication system. The maximum transmit power constraint for terahertz base stations. For the minimum transmission rate constraint of users based on the continuous convex approximation and Taylor series expansion, C6 is a robust constraint, and C6 is a rank-one constraint for the beamforming matrix.

[0028] S52. Use the semidefinite relaxation method to relax the constraints and determine whether the relaxed conditions satisfy the rank-one constraint. If they do, use the CVX toolbox to solve the problem. If they do not, use Gaussian randomization to obtain an approximate solution.

[0029] Furthermore, in step S6, the beamforming matrix of the k-th user is fixed to obtain the phase shift matrix with respect to RIS. and auxiliary variable set The sub-optimization problem, the solution process includes:

[0030] S61. Regarding the phase shift matrix of RIS and auxiliary variable set The sub-optimization problem is represented as:

[0031]

[0032] in, Let represent the transmission rate of the k-th user in the deterministic problem model during the communication process. An auxiliary variable representing the system's energy efficiency. C1 represents the total power consumption of the RIS-assisted terahertz communication system, and C2 represents the discrete phase shift constraint of the RIS. For the minimum transmission rate constraint of users based on the continuous convex approximation and Taylor series expansion, For robust constraints, C7 is a rank-one constraint on the RIS phase shift matrix;

[0033] S62. Constraint C2 is a non-convex optimization constraint; continuous phase constraint should be considered. The sub-optimization problem of S61 can be transformed as follows:

[0034]

[0035] in, Representing the first RIS The phase reflection coefficient of each array element, Represents a set of continuous phases;

[0036] S63. Use the CVX toolbox to solve the sub-optimization problem transformed from S62.

[0037] The beneficial effects of this invention are:

[0038] This invention addresses the issue of high communication interruptions in terahertz wireless communication systems due to obstruction. Considering user minimum rate constraints, base station transmit power constraints, channel uncertainty constraints, and RIS discrete phase shift constraints, it introduces a smart metasurface to construct a robust transmission model that maximizes energy efficiency based on bounded channel uncertainty. The S-Procedure is used to transform the uncertain non-convex optimization problem into a deterministic non-convex optimization problem; the Tinkelbach method is used to transform the fractional programming problem into a subtractive form; and the variable substitution method, continuous convex approximation theory, block coordinate descent method, and semidefinite programming method are used to transform the non-convex optimization problem into an equivalent convex optimization problem. This invention effectively improves the overall energy efficiency of terahertz communication systems under channel uncertainty conditions while reducing user communication interruptions.

[0039] Compared with methods under perfect channel state information, the present invention has higher robustness, improves the energy efficiency and received signal quality of terahertz communication networks, reduces the probability of user transmission interruption, and is applicable to RIS that cannot provide continuous phase shift control capability, making it more suitable for practical applications. Attached Figure Description

[0040] Figure 1 This is a system model diagram of the present invention;

[0041] Figure 2 This is a flowchart of the method of the present invention;

[0042] Figure 3 This is the energy efficiency convergence diagram of the method of the present invention;

[0043] Figure 4 This is a robustness diagram of the method of the present invention. Detailed Implementation

[0044] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0045] like Figure 2 As shown, the present invention proposes a robust transmission method for a smart metasurface-assisted terahertz communication system, comprising the following steps:

[0046] S1. Construct and initialize a RIS-assisted terahertz communication system, the system comprising K users, a RIS with N reflective elements, and a terahertz base station with M antennas.

[0047] Specifically, this invention considers a RIS-assisted downlink multi-user MISO terahertz communication system. The RIS is deployed between the terahertz base station and the user to improve reception performance. By adjusting the phase shift of the RIS, the channel gain of the user is changed, thereby enhancing the useful signal received by the user. Based on the above analysis, the RIS-assisted terahertz communication system constructed in this embodiment is as follows: Figure 1 As shown, in this communication system, a terahertz cellular base station with M antennas serves K single-antenna users through a RIS (Radio Reflector System). A RIS with N reflector elements reflects signals from the terahertz cellular base station, thereby enhancing the received signals of the K users and solving the problem of Loss of Signal (LoS) being blocked by obstacles. The user sets are defined respectively. RIS array set .

[0048] Specifically, the parameters of the RIS-assisted terahertz communication system are initialized. These parameters include: the number of terahertz base station antennas M, the number of RIS reflectors N, the number of system users K, and the maximum transmission power threshold of the terahertz base station. Circuit power consumption of terahertz base stations The circuit power consumption of each reflective element in the RIS Minimum transmission rate threshold for the k-th user The previous iteration value of the auxiliary variable and Composite antenna gain and path loss compensation factor Upper bound of the cascaded channel error for the k-th user Total system energy efficiency Lower limit of system energy efficiency Upper limit of system energy efficiency Maximum number of iterations Convergence accuracy and number of iterations .

[0049] S2. Based on the user minimum rate constraint, base station transmit power constraint, channel uncertainty constraint and RIS discrete phase shift constraint, a robust beamforming model is constructed with the optimization objective of maximizing the total energy efficiency of the RIS-assisted terahertz communication system.

[0050] In one embodiment, if channel uncertainty is disregarded, and only user minimum rate constraints, base station transmit power constraints, and RIS discrete phase shift constraints are considered under perfect channel state information, and the objective is to maximize the overall energy efficiency of the RIS-assisted terahertz communication system, then this optimization problem P can be expressed as:

[0051]

[0052] in, This represents the total throughput of the RIS-assisted terahertz communication system. This represents the transmission rate of the k-th user during the communication process. C1 represents the total power consumption of the RIS-assisted terahertz communication system, and C1 is the maximum transmit power constraint of the terahertz base station. This represents the beamforming vector from the terahertz base station to the k-th user. This represents the total power consumption of a terahertz base station for beamforming K users. C1 represents the maximum transmission power threshold of the terahertz base station; C2 is the discrete phase shift constraint of RIS, with the number of discrete phase shift bits being... L represents the number of adjustable phase shifts in a single element of the RIS array. Representing the first RIS The phase reflection coefficient of each array element, Indicates that RIS contains A discrete phase shift set of n elements; Minimum transmission rate constraint for users, This represents the minimum transmission rate threshold.

[0053] Specifically, The signal-to-interference-plus-noise ratio (SIR) of the k-th user during communication is represented by the following formula:

[0054] (1)

[0055] Indicates intermediate variables. , and These are the antenna gains for the transmitting and receiving ends, respectively. This represents the path loss compensation factor. This indicates the background noise power at the receiver. Let RIS be the channel vector from user k. , Indicates RIS to the 1st Channels for individual users Indicates the RIS containing terahertz composite path gain up to the th The channel for each user, where f is the center frequency. This represents the distance from user k to the center point of the RIS. For composite path gain, The distance from the terahertz base station to the RIS center point. Represents the phase shift matrix of RIS. , and These represent the first and second generations of passive RISC. The amplitude and phase reflection coefficients of each array element Normally take , , , , This represents the smallest adjustable phase shift. The channel matrix from the terahertz base station to the RIS, which includes terahertz composite path gain. , This represents the channel matrix from the terahertz base station to the RIS.

[0056] Specifically, the total power consumption of the RIS-assisted terahertz communication system The calculation formula is as follows:

[0057] (2)

[0058] The Euclidean criterion for vectors. Indicates base station to Total power consumption of beamforming per user , and These represent the circuit power consumption at the terahertz base station, a single RIS array element, and the user equipment, respectively.

[0059] In practical communication systems, the communication link between transceivers suffers from quantization errors, feedback delays, and estimation errors, making it extremely difficult for base stations to obtain accurate channel state information. To overcome the impact of channel uncertainty, this uncertainty is considered in the optimization problem P described above. According to robust optimization theory, the channel uncertainty problem can be described as follows:

[0060] (3)

[0061] in, This represents the channel from the base station to the k-th user via RIS reflection. It is an estimated CSI. It is an estimation error. This is the upper bound of the estimation error. Therefore, the signal-to-interference-plus-noise ratio (SIR) with channel uncertainty is... It can be represented as:

[0062] (4)

[0063] in, Represents RIS The transpose of each array element phase shift.

[0064] Therefore, the robust energy efficiency optimization problem P1, i.e., the robust beamforming model based on the channel uncertainty problem, can be expressed as:

[0065]

[0066] in, Let represent the transmission rate of the k-th user during communication, considering the channel uncertainty problem. This represents the signal-to-interference-plus-noise ratio (SIR) of the k-th user during communication, considering channel uncertainty. This represents the total throughput of a RIS-assisted terahertz communication system considering channel uncertainty. There is an estimation error in the channel. At that time, the user's minimum transmission rate constraint, C4 represents the minimum transmission rate threshold; C5 represents the channel uncertainty set constraint. This indicates the channel uncertainty problem.

[0067] S3. Use the S-Procedure to convert the robust beamforming model into a deterministic problem model.

[0068] Specifically, as can be seen from the robust energy efficiency optimization problem P1 above, its objective function contains mutually coupled optimization variables and channel uncertainties in the constraints, making it a very difficult non-convex optimization problem to solve. Therefore, the objective function must first be transformed. According to the Tinkelbach method, the fractional objective function can be transformed into a subtraction form:

[0069]

[0070] in, It is an auxiliary variable that represents the energy efficiency of the system.

[0071] because It's a semi-infinite constraint and needs to be transformed. The optimization variables are coupled with user rate. This makes the transformation difficult. Therefore, based on the continuous convex approximation and Taylor series expansion, C3 can be rewritten as:

[0072]

[0073] in, Represents auxiliary variables. and express and The previous iteration; using and Calculate by substituting into equation (1) respectively The numerator and denominator; after using the continuous convex approximation Equivalent to Therefore, problem P1 can be transformed into:

[0074]

[0075] However, and It is still a semi-infinite constraint with uncertainties, so the S-Procedure is used to... Transformed into an equivalent form of a finite number of linear matrix inequalities:

[0076]

[0077] in, Indicates intermediate variables. ,, Indicates the first Beamforming matrix for each user Represents the identity matrix. Indicates the first The upper bound of the cascaded channel estimation error for each user, vec() denotes matrix vectorization. For Kronecker product, Let be the rank of the matrix.

[0078] Similarly, It can be transformed into:

[0079]

[0080] in, .

[0081] At this point, all uncertain constraints have been transformed into deterministic constraints. Therefore, optimization problem P1 can be transformed into problem P2:

[0082]

[0083] in, For the set of auxiliary variables, This represents the RIS phase shift matrix.

[0084] S4. Calculate the overall energy efficiency of the system.

[0085] Specifically, the formula for calculating the total energy efficiency of the system in step S4 is as follows:

[0086]

[0087] in, This represents the total system energy efficiency at the l-th iteration. This represents the lower limit of the system's energy efficiency. The upper limit of the system's energy efficiency is represented by the Tinkelbach function. Perform a judgment update when hour, ,when hour, .

[0088] S5. With the phase shift matrix of the fixed RIS fixed, calculate the first... Beamforming matrix for each user.

[0089] Specifically, due to the optimization variables and Coupling remains a multivariate coupled nonconvex optimization problem, specifically problem P2. Therefore, this invention uses an alternating iteration criterion based on block coordinate descent, fixing one variable and alternating the iterations of the variables... and Optimization is performed to solve the problem.

[0090] Specifically, in step S5, the phase shift matrix of RIS is fixed, and the first step is calculated. Beamforming matrix for individual users , to obtain information about the first Beamforming matrix for individual users and auxiliary variable set Sub-optimization problem:

[0091]

[0092] Due to the existence of the rank-one constraint This sub-optimization problem is non-convex and difficult to solve. Therefore, a semidefinite relaxation method is considered. After relaxation, the sub-optimization problem transforms into a convex optimization problem, which can be solved using the CVX toolbox. Specifically, after relaxing the original problem, only satisfying the rank-one constraint can yield the beamforming matrix. The beamforming vector is transformed into a feasible solution through eigenvalue decomposition. Otherwise, Gaussian randomization is needed to obtain an approximate solution.

[0093] S6. Fix the beamforming matrix of the k-th user and calculate the phase shift matrix of RIS.

[0094] Specifically, in step S6, the beamforming matrix of the k-th user is fixed, and the phase shift matrix of RIS is calculated. The phase shift matrix of RIS is obtained. and auxiliary variable set Sub-optimization problem:

[0095]

[0096] Because the phase in constraint C2 is discrete, this constraint is a non-convex optimization constraint. Let's first consider the continuous phase constraint. . A semidefinite relaxation method is used. The problem can be transformed into:

[0097]

[0098] At this point, the sub-optimization problem has been transformed into a convex optimization problem, which can be solved using the CVX toolbox.

[0099] S7. Determine whether the current system total energy efficiency has converged. If yes, proceed to step S9; otherwise, proceed to step S8.

[0100] S8. Determine if the maximum number of iterations has been reached. If yes, proceed to step S9; otherwise, increment the iteration count by 1 and return to step S4.

[0101] S9. Calculate the discretized phase shift matrix of RIS.

[0102] Specifically, the continuous phase is quantized into L discrete phases using the quantization mapping method, and the corresponding discrete phases are calculated:

[0103]

[0104] in, Represents the RIS discrete phase shift matrix. This represents the solution to the RIS continuous phase shift matrix.

[0105] In one embodiment, it is assumed that the system has a terahertz base station located at (0,0), a RIS located at (3,3), and the user located within a circle with a radius of 1.5 meters centered at (5,0). The angle of arrival and the angle of departure follow... Assuming uniform distribution within the antenna range, and an antenna distance of half a wavelength, the noise power at the transceiver is equal. Define an upper bound for normalized uncertainty. Other parameters are shown in Table 1.

[0106] Table 1 Simulation Parameter Table

[0107]

[0108] In this embodiment, Figure 3 An energy efficiency convergence graph of the iterative method in this example is provided. Figure 4 Robustness diagrams for the iterative method in this example are provided. Figure 3The results show that the method of the present invention can achieve convergence quickly, thus proving that the method of the present invention can effectively guarantee the communication quality of users and has real-time performance. Figure 4 This shows the upper bound of channel uncertainty ( As the number of methods increases, the probability of user interruption also increases, but at the same time... Under these conditions, the actual interference power received by the user using the method of the present invention is minimal and less than 5%, thus proving that the method of the present invention has strong robustness. Figure 3 and Figure 4 Experimental results show that the method of the present invention ensures both real-time performance and service quality for users, and has strong robustness.

[0109] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "setting," "connection," "fixing," "rotation," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Unless otherwise explicitly limited, those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0110] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for robust transmission of intelligent metasurface assisted terahertz communication system, characterized in that, The method comprises the following steps: S1. Constructing and initializing a RIS-aided terahertz communication system, the system comprising K users, one RIS with N reflecting elements, and one terahertz base station with M antennas; S2. Constructing a robust beamforming model with the optimization objective of maximizing the total energy efficiency of the RIS-aided terahertz communication system according to the user minimum rate constraint, the base station transmit power constraint, the channel uncertainty constraint, and the RIS discrete phase shift constraint; The robust beamforming model constructed in step S2 is represented as: in, Let represent the transmission rate of the k-th user during communication, considering the channel uncertainty problem. This represents the signal-to-interference-plus-noise ratio (SIR) of the k-th user during communication, considering channel uncertainty. This represents the total throughput of a RIS-assisted terahertz communication system considering channel uncertainty. C1 represents the total power consumption of the RIS-assisted terahertz communication system; C1 is the maximum transmit power constraint of the terahertz base station, where... This represents the beamforming vector from the terahertz base station to the k-th user. This represents the total power consumption of a terahertz base station for beamforming K users. C1 represents the maximum transmission power threshold of the terahertz base station; C2 is the discrete phase shift constraint of RIS, where Representing the first RIS The phase reflection coefficient of each array element, Indicates that RIS contains A discrete phase shift set of n elements; There is an estimation error in the channel. At that time, the user's minimum transmission rate constraint, C4 represents the minimum transmission rate threshold; C5 represents the channel uncertainty set constraint. This indicates the channel uncertainty problem; Represents the phase shift matrix of RIS; S3. Converting the robust beamforming model into a deterministic problem model by using an S-Procedure; S4. Calculating the total energy efficiency of the system; S5. Fix the phase shift matrix of RIS, calculate the beamforming matrix of the i-th user; user; S6. Fixing the beamforming matrix of the kth user, and calculating the phase shift matrix of the RIS; S7. Judging whether the current total energy efficiency of the system converges, if yes, outputting the optimal energy efficiency of the system, the base station beamforming matrix, and the continuous phase shift matrix of the RIS, and then performing step S9; if not, performing step S8; S8. Judging whether the current iteration number is greater than the maximum iteration number, if yes, outputting the optimal energy efficiency of the system, the base station beamforming matrix, and the continuous phase shift matrix of the RIS, and then performing step S9; if not, increasing the iteration number by 1, and returning to step S4; S9. Calculating the discretization phase shift matrix of the RIS.

2. The method of claim 1, wherein, Step S3 converts the robust beamforming model into a deterministic problem model by using an S-Procedure, and the deterministic problem model is represented as: in, Let represent the transmission rate of the k-th user in the deterministic problem model during the communication process. An auxiliary variable representing the system's energy efficiency. For the set of auxiliary variables, This indicates the total power consumption of the RIS-assisted terahertz communication system. This is the conjugate transpose of the matrix; The maximum transmit power constraint for terahertz base stations, where Indicates the first Beamforming matrix for each user Let be the beamforming vector from the terahertz base station to the k-th user. This represents the total power consumption of the base station for beamforming K users. Tr() represents the maximum transmission power threshold of the terahertz base station; C2 represents the trace of the matrix; C2 is the discrete phase shift constraint of RIS. C1 represents the minimum transmission rate constraint for users based on the continuous convex approximation and Taylor series expansion; C2 represents the channel uncertainty set constraint. For robust constraints, , , , and Both represent auxiliary variables. and express and The previous iteration, Represents the identity matrix. Indicates intermediate variables. ,in , and These are the antenna gains for the transmitting and receiving ends, respectively. This represents the path loss compensation factor. For Kronecker product, Represents RIS Phase shift of each array element, , and They represent the first, second, and third RIS, respectively. The amplitude and phase reflection coefficients of each array element Indicates from the base station to the... Cascaded channel estimates for individual users Indicates the first The upper bound of the cascaded channel estimation error for each user, vec() denotes matrix vectorization. Indicates intermediate variables. , denotes the background noise power at the receiver; C6 is a beamforming matrix rank-one constraint, is the rank of the matrix; C7 is a RIS phase shift matrix rank-one constraint, where denotes the RIS phase shift matrix.

3. The method of claim 1, wherein, The formula for calculating the total energy efficiency of the system in step S4 is: wherein, represents the total system energy efficiency at the lth iteration, represents the lower bound of the system energy efficiency, represents the upper bound of the system energy efficiency; the Tinker-Habach function is used the judgment update is performed when , when , .

4. The method of claim 1, wherein, Step S5 fixes the phase shift matrix of RIS, and obtains information about the first... Beamforming matrix for individual users and auxiliary variable set The sub-optimization problem, the solution process includes: S51. Regarding the first Beamforming matrix for individual users and auxiliary variable set The sub-optimization problem is represented as: wherein, denotes the transmission rate of the kth user in the communication process in the deterministic problem model, denotes an auxiliary variable of system energy efficiency, denotes the total power consumption of the RIS-assisted terahertz communication system, is the maximum transmit power constraint of the terahertz base station, is the minimum transmission rate constraint of the user based on the successive convex approximation and Taylor series expansion, is a robust constraint, and C6 is a beamforming matrix rank one constraint; S52. Relaxing the constraint condition by using a semi-positive relaxation method, and judging whether the relaxed condition satisfies the rank-one constraint, if yes, solving by using a CVX toolbox, if not, obtaining an approximate solution by using a Gaussian randomization.

5. The method of claim 1, wherein, Step S6 fixes the beamforming matrix of the kth user, obtaining the phase shift matrix of the RIS and the auxiliary variable set The solving process includes: S61. The phase shift matrix with respect to the RIS and the set of auxiliary variables suboptimization problem, denoted as: wherein, denotes the transmission rate of the kth user in the communication process in the deterministic problem model, denotes an auxiliary variable of system energy efficiency, denotes the total power consumption of the RIS-assisted terahertz communication system, C2 is the discrete phase shift constraint of the RIS, is the minimum transmission rate constraint of the user based on continuous convex approximation and Taylor series expansion, is a robust constraint, and C7 is a rank-one constraint of the RIS phase shift matrix. S62. Constraint C2 is a non-convex optimization constraint, consider continuous phase constraint The sub-optimization problem of S61 is converted as follows: wherein, denotes the amplitude and phase reflection coefficients of the element of the RIS, denotes a set of consecutive phase shifts; S63. Solving the converted sub-optimization problem in S62 by using a CVX toolbox.

Images