Terahertz multi-intelligent reflecting surface communication beam optimization method based on two-layer delay lines

Abstract

This invention proposes a terahertz multi-intelligent reflector communication beam optimization method based on a two-layer time delay line. The steps are as follows: First, a distributed RISS-assisted broadband terahertz MIMO communication system is constructed, wherein the BS comprises a uniform linear array of N antennas and N... RF One radio frequency chain provides services to K single-antenna users. Secondly, a simulated beamforming matrix is ​​generated at the BS end using two layers of time delay lines. Then, the achievable rate of the users is calculated based on the simulated beamforming matrix, and maximizing the achievable rate of the users is used as the objective function. Finally, the objective function is optimized and solved according to the designed joint beamforming model and alternating iterative optimization algorithm to obtain the reflection coefficients of digital beamforming and RIS. This invention, based on a hierarchical structure, combines time delay lines capable of achieving large delays with those capable of small delays to minimize the number of large-delay time delay lines, effectively mitigating beam dispersion problems, improving practical feasibility, and reducing hardware complexity.

Description

Technical Field

[0001] This invention relates to the field of terahertz communication technology, and in particular to a terahertz multi-intelligent reflector communication beam optimization method based on two-layer time delay lines. Background Technology

[0002] To achieve data rates of terahertz per second (Tb / s), future 6G networks are expected to utilize terahertz frequencies (0.1-10 THz) with ultra-wide bandwidth. However, terahertz signals suffer from severe path loss and poor diffraction, making it difficult to achieve ideal coverage. This problem can be addressed through massive MIMO and reconfigurable smart reflector (RIS) technologies. Specifically, on the one hand, due to the short wavelength of terahertz signals, the physical aperture of the antenna can be very small, allowing massive MIMO antenna arrays to be used for terahertz communication, supporting the generation of very narrow beams and very high array gain, which can significantly compensate for severe path loss. On the other hand, using RIS as a passive repeater between the transmitter and receiver establishes a new non-line-of-sight link, solving the problem of poor diffraction capability of terahertz signals.

[0003] Massive MIMO means that fully digital beamforming is difficult to implement due to its enormous power consumption, necessitating the use of hybrid beamforming, including digital / analog beamforming. The most important function of analog beamforming is generating directional beams to compensate for the severe path loss of terahertz signals. To achieve this, phase shift compensation is required for different antenna elements. The most widely used phase shifting element is the analog phase shifter. Typically, the required phase shift of the phase shifter is determined by the carrier frequency. This is not a problem for narrowband systems with a small frequency range. However, in broadband systems with a large frequency range, because phase shifters can only achieve frequency-independent phase shifts, the generated beam may disperse in surrounding directions, leading to beam dispersion problems.

[0004] In RIS-assisted broadband terahertz MIMO communication systems, beam dispersion also exists at the RIS end. Specifically, since the RIS is usually equipped with frequency-independent phase-shifting circuits, RIS-assisted communication can only perform frequency-independent precoding. The deployment method of the RIS has a significant impact on beam dispersion. On the one hand, the more RIS units there are, the better the user can receive signals transmitted by the base station; on the other hand, the more RIS units there are, the greater the beam dispersion effect. Summary of the Invention

[0005] To address the shortcomings of the aforementioned background technology, this invention proposes a terahertz multi-intelligent reflector communication beam optimization method based on two-layer time delay lines, which solves the beam dispersion problem while better balancing performance and hardware cost.

[0006] The technical solution of this invention is implemented as follows:

[0007] A terahertz multi-intelligent reflector communication beam optimization method based on two-layer time delay lines, the steps of which are as follows:

[0008] Step 1: Construct a distributed RISS-assisted broadband terahertz MIMO communication system, where BS consists of a uniform linear array of N antennas and N... RF A radio frequency chain serves K single-antenna users, and N≥N RF ;

[0009] Step 2: Use two layers of time delay lines at the BS end to generate an analog beamforming matrix;

[0010] Step 3: Calculate the user's reachable rate based on the simulated beamforming matrix, and use maximizing the user's reachable rate as the objective function;

[0011] Step 4: Design a joint beamforming model. Based on the joint beamforming model and the alternating iterative optimization algorithm, optimize the objective function to obtain the reflection coefficients of digital beamforming and RIS.

[0012] The implementation method for step two is as follows:

[0013] First, connect the subarray consisting of P antennas to the delay line of the first layer; then, K L The subarrays are synthesized into a primary array with a smaller aperture, and then connected to the delay line of the second layer; finally, K... H A large-aperture array is composed of primary arrays; wherein, the number of bits of the delay lines in the first layer is P. L The number of bits in the second-layer delay line is P. H ;

[0014] Deploy R RISs and define This serves as the index set for RISs; all RISs have the same size, and each RIS consists of a uniform planar array with N cells. RIS =M x ×M y , of which M x M represents the row number of each RIS. y Represents the number of columns for each RIS; definition For the set of element indices on the row, For the set of element indices on the column;

[0015] Assume the number of radio frequency chains is N RF Equal to the number of RIS, i.e., N RF=R, where each RIS is served by a beam generated by a radio frequency chain; reliable broadband transmission is achieved using M subcarriers orthogonal frequency division multiplexing; the frequency corresponding to the m-th subcarrier is Among them, f c B is the center frequency, and B is the bandwidth.

[0016] The equivalent channel h between BS and the k-th user on the m-th subcarrier m,k Represented as:

[0017]

[0018] in, It is the channel matrix between the BS and the r-th RIS on the m-th subcarrier. It is the channel between the k-th user and the r-th RIS on the m-th subcarrier. It is the diagonal matrix of the reflection coefficients of the r-th RIS. Indicates the reflection amplitude, φ r Indicates the phase shift of the reflection;

[0019] According to the equivalent channel h m,k Calculate the received signal y of the k-th user on the m-th subcarrier. m,k :

[0020]

[0021] Where F = F A F L F H A simulated beamforming matrix implemented for a two-layer time delay line structure; Both represent digital beamforming vectors. It is the additive white Gaussian noise of the k-th user on the m-th subcarrier. For variance; s m,k For the information of the k-th user on the m-th subcarrier, s m,j This refers to the information of the j-th user on the m-th subcarrier;

[0022]

[0023] in, This indicates that P phase shifters pass through the k-th phase shifter. l The delay line and the kth time line h The delay line is connected to the nth delay line. rf Beamforming vectors generated by each radio frequency chain; This represents the frequency-shifted phase shift implemented by the first-layer delay line network, namely:

[0024]

[0025] in, It is connected to the nth rf The RF chain and the kth h K with time delay L The delay vector is implemented by a delay line; The frequency-shifted phase shift implemented for the second-layer delay line network is as follows:

[0026]

[0027] in, It is connected to the nth rf K under a radio frequency chain H A time delay vector implemented by a time delay line.

[0028] The channel matrix G r,m The expression is:

[0029]

[0030] Where L1 represents the number of paths, This represents the gain of the l1-th path leading to the r-th RIS. This represents the delay of the l1-th path to the r-th RIS. This represents the array response vector at point BS. for It is the physical direction of the l1-th path from BS to the r-th RIS. This represents the array response vector at RIS. This represents the azimuth angle of the l1-th path at the r-th RIS. This represents the elevation angle of the l1-th path at the r-th RIS arrival angle;

[0031] The They are represented as follows:

[0032]

[0033]

[0034] Where c is the speed of light, d is the distance between two consecutive antennas, and d = λ c / 2,λ c This represents the corresponding center frequency f. c The wavelength; n represents the nth antenna element;

[0035] The channel between the k-th user and the r-th RIS on the m-th subcarrier for:

[0036]

[0037] Where L2 represents the number of paths, This represents the gain of the l2-th path from the r-th RIS to the k-th user. This represents the delay of the l2-th path from the r-th RIS to the k-th user; Let be the emitter array response vector at RIS, expressed as:

[0038]

[0039] in, This represents the azimuth angle of the l2-th path from the r-th RIS to the k-th user. This represents the elevation angle of the l2th path from the r-th RIS to the k-th user.

[0040] The achievable speed for the user is:

[0041]

[0042] in, Let SINR be the SINR of the k-th user on the m-th subcarrier.

[0043] The objective function is obtained as follows:

[0044] For the delay lines of the first layer and the delay lines of the second layer, a uniform quantization method is used to generate them respectively. and A discrete value; therefore, a discrete time delay value and They are represented as follows:

[0045]

[0046]

[0047] Where D represents the delay step size;

[0048] For a phase shifter, a b-bit quantization phase shift is set, and the beamforming vector generated by the phase shifter... for:

[0049]

[0050] The reflection coefficient of RIS Discretize to obtain discrete values:

[0051]

[0052] Where Q represents that each RIS contains 2 QThere are discrete phase shifts; therefore, the objective function to maximize the achievable rate is:

[0053]

[0054] Where Θ=diag(Φ1,…,Φ R ), P max This represents the maximum transmission power of the base station.

[0055] The method for optimizing the objective function is as follows:

[0056] S41. Design a joint beamforming model:

[0057] The kth digit of the first layer is calculated using the following formula. l The delay line and the kth layer of the second layer h Optimal delay for each delay line:

[0058]

[0059]

[0060] Among them, T d θ0 represents the time delay between two consecutive antennas, and θ0 represents the direction of the target.

[0061] Based on the approximation principle, the solution is... Mapped to sets respectively and The elements in the array are represented as follows:

[0062]

[0063]

[0064] Where τ1 is The element in, τ2 is The elements in the above mapping can be used to obtain a suboptimal solution for two delay lines;

[0065] Due to the array response vector It is of equal amplitude and phase, and can be used as F A The column, in which We obtain it from the following formula:

[0066]

[0067] in, At the center frequency f c The spatial direction of the l1-th path from BS to the r-th RIS; a 1→P This represents the array response vector; similarly, the solved phase shift... Projected onto a set of discrete values It can be written as:

[0068]

[0069] The suboptimal solution for the phase shifter can be obtained from the above equation;

[0070] S42, Fixed Θ optimization d m,k :

[0071] Given F A ,F L ,F H In this case, the objective function P1 can be reformulated as an optimization problem P2:

[0072]

[0073] The equivalent channel vector for the k-th user on the m-th subcarrier can be written as: Based on the extension of the Sherman-Morrison-Woodbury formula, we obtain:

[0074]

[0075] Therefore, optimization problem P2 can be transformed into P3:

[0076]

[0077] P3 and They are respectively:

[0078]

[0079]

[0080] in,

[0081]

[0082]

[0083] P3 is solved by a numerical convex programming solver until convergence, thus obtaining the local optimal solution of the digital beamforming vector.

[0084] S43, fixed d m,k Optimization Θ:

[0085] because Transform P3 into an optimization problem P4:

[0086]

[0087] Assumption And define a vector containing the reflection coefficients of all units, i.e. Set Q = 1 bit; transform optimization problem P4 into P5:

[0088]

[0089] Design a CUA algorithm for the reflection matrix, and optimize it during the optimization process. Where i represents the current iteration number, and i-1 represents the previous iteration number;

[0090] P5 is solved using the CUA algorithm for the reflection matrix until R... sum Convergence occurs, and the reflection coefficient matrix Θ is output.

[0091] The method for solving P5 using the CUA algorithm based on the reflection matrix is ​​as follows:

[0092] S43.1, Input: Channel matrix f r,m,k G r,m Analog precoding F; Digital precoding d m,k Number of iterations I o ;

[0093] S43.2 Initialization:

[0094] S43.3, Order but

[0095] S43.4 Transform the vector φ1 into the reflection matrix Θ1;

[0096] S43.5, Order but

[0097] S43.6 Transform the vector φ2 into the reflection matrix Θ2;

[0098] S43.7. Use the formula q=argmax q＝1,2 {R sum (Θ q Calculate the value of q and assign it to Θ = Θ q ;

[0099] S43.8. Repeat steps S43.3 to S43.7 until the maximum number of iterations is reached, and output the reflection coefficient matrix Θ.

[0100] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0101] 1) This invention is based on a hierarchical structure, which combines delay lines that can achieve large delays with delay lines that can achieve small delays, in order to minimize the number of delay lines with large delays, thereby improving practical feasibility and reducing hardware complexity.

[0102] 2) A distributed small-scale RIS was deployed to alleviate the beam splitting effect at the RIS end, which can effectively alleviate the beam dispersion problem and achieve near-optimal achievable rate performance. Attached Figure Description

[0103] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0104] Figure 1 This is a block diagram of the distributed RISS-assisted broadband terahertz MIMO communication system of the present invention.

[0105] Figure 2 This represents the relationship between the achievable rate and the maximum transmit power under different time delay step sizes D.

[0106] Figure 3 represents the achievable rate for different number of iterations.

[0107] Figure 4 This represents the achievable rate for different iterations under discrete phase shifter conditions. Detailed Implementation

[0108] 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.

[0109] This invention provides a terahertz multi-intelligent reflector communication beam optimization method based on a two-layer delay line architecture. Based on a hierarchical structure, it combines delay lines capable of achieving large delays with those capable of achieving small delays to minimize the number of large-delay delay lines. In the two-layer delay line structure, antenna elements and phase shifters are connected to the first-layer delay line to form a sub-array with a smaller aperture. Then, the sub-array with the smaller aperture is connected to the second-layer delay line to synthesize an antenna array with a larger aperture. Specifically, the delay lines in different layers only need to compensate for the delay caused by the aperture of the sub-array in that layer; that is, the first-layer and second-layer delay lines provide phase control for the small-aperture array and the large-aperture array, respectively, and the remaining phase shift is compensated by the phase shifter. In this two-layer delay line architecture, the number of required large-delay delay lines can be significantly reduced, improving practical feasibility and reducing hardware complexity.

[0110] The specific steps are as follows:

[0111] Step 1: Build a distributed RISS-assisted wideband terahertz MIMO communication system, such as... Figure 1 As shown, the direct path between the BS and the user is blocked; the BS consists of a uniform linear array of N antennas and N... RF A radio frequency chain serves K single-antenna users, and N≥N RF .

[0112] Step 2: At the BS end, a two-layer time delay line is used to generate an analog beamforming matrix; specifically, firstly, a subarray consisting of P antennas is connected to the first-layer time delay line; then, K... L The subarrays are synthesized into a primary array with a smaller aperture, and then connected to the delay line of the second layer; finally, K... H A large-aperture array is composed of primary arrays; wherein, the number of bits of the delay lines in the first layer is P. L The number of bits in the second-layer delay line is P. H .

[0113] In addition, deploy R RISs and define This serves as the index set for RISs; assuming all RISs have the same size, each RIS consists of a uniform planar array with N cells. RIS =M x ×M y , of which M x M represents the row number of each RIS. y Represents the number of columns for each RIS; definition For the set of element indices on the row, This is the set of element indices for the column.

[0114] Assume the number of radio frequency chains is N RF Equal to the number of RIS, i.e., NRF =R, where each RIS is served by a beam generated by a radio frequency chain; reliable broadband transmission is achieved using M subcarriers orthogonal frequency division multiplexing (OFDM); the frequency corresponding to the m-th subcarrier is Among them, f c B is the center frequency, and B is the bandwidth.

[0115] The equivalent channel h between BS and the k-th user on the m-th subcarrier m,k Represented as:

[0116]

[0117] in, It is the channel matrix between the BS and the r-th RIS on the m-th subcarrier. It is the channel between the k-th user and the r-th RIS on the m-th subcarrier. It is the diagonal matrix of the reflection coefficients of the r-th RIS. Indicates the reflection amplitude, φ r This represents the reflection phase shift. Assume the amplitude of the reflection coefficient... The Saleh-Valenzuela terahertz channel model was adopted, with channel matrix G. r,m The expression is:

[0118]

[0119] Where L1 represents the number of paths, This represents the gain of the l1-th path leading to the r-th RIS. This represents the delay of the l1-th path to the r-th RIS. This represents the array response vector at point BS. for of……, It is the physical direction of the l1-th path from BS to the r-th RIS. This represents the array response vector at RIS. This represents the azimuth angle of the l1-th path at the r-th RIS. This represents the elevation angle of the l1-th path at the r-th RIS arrival angle.

[0120] The They are represented as follows:

[0121]

[0122]

[0123] Where c is the speed of light, d is the distance between two consecutive antennas, and d = λc / 2,λ c This represents the corresponding center frequency f. c The wavelength; n represents the nth antenna element.

[0124] The channel between the k-th user and the r-th RIS on the m-th subcarrier for:

[0125]

[0126] Where L2 represents the number of paths, This represents the gain of the l2-th path from the r-th RIS to the k-th user. This represents the delay of the l2-th path from the r-th RIS to the k-th user; Let be the emitter array response vector at RIS, expressed as:

[0127]

[0128] in, This represents the azimuth angle of the l2-th path from the r-th RIS to the k-th user. This represents the elevation angle of the l2th path from the r-th RIS to the k-th user.

[0129] According to the equivalent channel h m,k Calculate the received signal y of the k-th user on the m-th subcarrier. m,k :

[0130]

[0131] Where F = F A F L F H A simulated beamforming matrix implemented for a two-layer time delay line structure; Both represent digital beamforming vectors. It is the additive white Gaussian noise of the k-th user on the m-th subcarrier. For variance; s m,k For the information of the k-th user on the m-th subcarrier, satisfying E[|s m,k | 2 ] = 1; s m,j This refers to the information of the j-th user on the m-th subcarrier;

[0132]

[0133] in, This indicates that P phase shifters pass through the k-th phase shifter. l The delay line and the kth time lineh The delay line is connected to the nth delay line. rf Beamforming vectors generated by each radio frequency chain; This represents the frequency-shifted phase shift implemented by the first-layer delay line network, namely:

[0134]

[0135] in, It is connected to the nth rf The RF chain and the kth h K with time delay L The delay vector is implemented by a delay line; The frequency-shifted phase shift implemented for the second-layer delay line network is as follows:

[0136]

[0137] in, It is connected to the nth rf K under a radio frequency chain H A time delay vector implemented by a time delay line.

[0138] Step 3: Calculate the user's reachable rate based on the simulated beamforming matrix, and use maximizing the user's reachable rate as the objective function;

[0139] The SINR of the k-th user on the m-th subcarrier can be calculated as:

[0140]

[0141] The achievable speed for the user is:

[0142]

[0143] Considering practical hardware limitations, delay lines and phase shifters can only achieve discrete delays and phase shifts. For the delay lines of the first and second layers, uniform quantization is used to generate... and A discrete value; therefore, a discrete time delay value and They are represented as follows:

[0144]

[0145]

[0146] Where D represents the delay step size.

[0147] For a phase shifter, a b-bit quantization phase shift is set, and the beamforming vector generated by the phase shifter... for:

[0148]

[0149] The reflection coefficient of RIS Discretize to obtain discrete values:

[0150]

[0151] Where Q represents that each RIS contains 2 Q There are discrete phase shifts; therefore, the objective function to maximize the achievable rate is:

[0152]

[0153] Where Θ=diag(Φ1,…,Φ R ), P max This represents the maximum transmit power of the base station. The objective of P1 is to jointly optimize the reflection coefficient matrix Θ and the frequency-independent beamforming matrix F. A Frequency-dependent phase shift F L and F H and digital beamforming vector d m,k To maximize the achievable rate. The first constraint is the total transmit power constraint, due to F A F H F H and d m,k The coupling between the components is such that the power constraint is non-convex. Furthermore, the remaining constraints restrict the optimization parameters to discrete values. Therefore, problem P1 is an NP-hard problem, and there is currently no standard method to efficiently find its global optimum. This invention proposes an efficient algorithm to solve this problem.

[0154] Step 4: Design a joint beamforming model. Based on the joint beamforming model and the alternating iterative optimization algorithm, optimize the objective function to obtain the reflection coefficients of digital beamforming and RIS.

[0155] This invention proposes a joint beamforming framework to solve the optimization problem. To solve for P1, firstly, an analog beamforming framework including a phase shifter and two-layer time delay lines is designed based on different RISs physical directions; then, an alternating iterative optimization algorithm is proposed to solve for the reflection coefficients of digital beamforming and RISs. Specific algorithm details are as follows:

[0156] S41. Design a joint beamforming model:

[0157] The simulated beamforming matrix is ​​F = F A F L F HIt is implemented using a phase shifter and two layers of delay lines. First, a frequency-dependent phase shift implemented by a delay line network is designed; then, for the non-convex constraint problem of discrete delay, an approximate projection method is adopted. The k-th layer of the first layer is calculated according to the following formula. l The delay line and the kth layer of the second layer h Optimal delay for each delay line:

[0158]

[0159]

[0160] Among them, T d θ represents the time delay between two consecutive antennas, and θ0 represents the direction of the target.

[0161] Based on the approximation principle, the solution is... Mapped to sets respectively and The elements in the array are represented as follows:

[0162]

[0163]

[0164] Where τ1 is The element in, τ2 is The elements in the above mapping can be used to obtain the suboptimal solution for the two-layer delay lines.

[0165] Based on this, phase shifters can be used to compensate for the residual phase shift of the preceding delay line network and generate a beam aligned with the RISS physical direction. This is due to the array response vector... It is of equal amplitude and phase, and can be used as F A The column, in which We obtain it from the following formula:

[0166]

[0167] in, At the center frequency f c The spatial direction of the l1-th path from BS to the r-th RIS; a 1→P This represents the array response vector; similarly, the solved phase shift... Projected onto a set of discrete values It can be written as:

[0168]

[0169] The suboptimal solution for the phase shifter can be obtained from the above formula.

[0170] S42, Fixed Θ optimization dm,k :

[0171] Given F A ,F L ,F H In this case, the objective function P1 can be reformulated as an optimization problem P2:

[0172]

[0173] However, due to the nonconvexity of the objective function, problem P2 remains difficult to solve. The reflection coefficient matrix Θ and the digital beamforming vector d... m,k Joint optimization is a challenge. Therefore, an alternating iterative optimization algorithm is proposed to find a feasible solution to this problem. Specifically, for a given reflection coefficient matrix Θ, this invention proposes an iterative algorithm based on the minimum mean square error (MMSE) technique to obtain the digital beamforming vector d. m,k Then, the Coordinated Update Algorithm (CUA) is used to obtain the reflection coefficient matrix of the RIS. Finally, the above process is repeated until convergence to obtain the final digital beamforming vector and the reflection coefficient matrix of the RIS.

[0174] The equivalent channel vector for the k-th user on the m-th subcarrier can be written as: Based on the extension of the Sherman-Morrison-Woodbury formula, we obtain:

[0175]

[0176] Therefore, optimization problem P2 can be transformed into P3:

[0177]

[0178] P3 and They are respectively:

[0179]

[0180]

[0181] in,

[0182]

[0183]

[0184] P3 is solved using a numerical convex programming solver until convergence, yielding a local optimum solution for the digital beamforming vector; specifically, due to the obtained This is the optimal solution for P3 in the i-th iteration. Iterative updates to these variables either increase or maintain the value of the objective function. Therefore, the algorithm will converge to at least a local optimum.

[0185] S43, fixed d m,k Optimization Θ:

[0186] After obtaining the digital beamforming vector d m,k After the simulated beamforming matrix F, consider the reflection matrix of RIS.

[0187] because Transform P3 into an optimization problem P4:

[0188]

[0189] To simplify the expression, assume And define a vector containing the reflection coefficients of all units, i.e. Without loss of generality, let Q = 1 bit; transform the optimization problem P4 into P5:

[0190]

[0191] The details of the CUA algorithm for designing the reflection matrix are shown in Algorithm 1 below. During the optimization process... Where i represents the current iteration number, and i-1 represents the previous iteration number; considering the actual set The number of discrete phase shift values ​​is usually finite, making one-dimensional search highly efficient. The CUA algorithm for the reflection matrix is ​​used to solve for P5 until R... sum Convergence occurs, and the reflection coefficient matrix Θ is output.

[0192]

[0193] Experimental Simulation: The performance of the proposed two-layer delay line scheme in a broadband terahertz distributed RISs-assisted communication system was simulated and evaluated. In this system, the base station is located at (50m, 0m, 3m), and K=4 users are randomly distributed within a circle with a radius of 1m centered at (0, 85m, 0). In addition, R=4 distributed small RISs-assisted communication networks were deployed at locations of (0, 80m, 6m), (0, 80m, 8m), (0, 85m, 6m), and (0, 85m, 8m), respectively. Since terahertz communication mainly relies on line-of-sight links, L1=L2=1 was set. Other experimental parameters are set as shown in Table 1.

[0194] Table 1: System Parameter Settings

[0195]

[0196] To evaluate the impact of the delay step size D of the proposed two-layer delay line scheme, Figure 2 The achievable rate and maximum transmit power P were plotted for different time delay steps. max The relationship is as follows. To avoid other influences, the number of bits in the phase shifter is set to infinity, and RIS-assisted communication is temporarily ignored. In the single-layer delay line-based scheme and the double-layer delay line-based scheme, U=32 and K are set respectively. H =8,K L =4. The time delay step D is set to ideal continuous and 0.15T respectively. c 0.25T c At that time, the step size D was extended from 0.25T. c It becomes 0.15T c When the delay quantization error is reduced, it can be easily observed that the achievable rate is higher and closer to the upper bound. The two-layer delay line scheme, while significantly reducing the number of high-order delay lines, achieves performance close to that of the single-layer delay line scheme. Experimental results show that this scheme can effectively balance system performance and hardware cost. Furthermore, under different schemes, the achievable rate increases with P... max The performance of hybrid precoding based on phase shifters increases with the increase of phase shifters, while the performance of hybrid precoding based on delay lines is very limited compared to delay-line based schemes.

[0197] Figure 3 The achievable rates at different iteration numbers in a broadband terahertz distributed RISS-assisted communication system are presented to evaluate the convergence of the proposed algorithm. To verify the impact of the number of delay lines in different schemes, U=32 and U=16 are set in the traditional single-layer delay line scheme, respectively. In the two-layer delay line scheme, K is set to... H =8,K L =4, and K H =8,K L =2. The time delay step D is 0.15T. c Furthermore, the phase shifter resolution was set to infinity, and the RIS resolution was set to Q=1. After three iterations, the reachability rate converged, proving the effectiveness of the algorithm. In addition, the results show that the more delay lines, the higher the reachability. Moreover, it can be observed that the reachability rate is slightly higher under the traditional single-layer delay line scheme. Since the difference between the two is very small, the proposed two-layer delay line scheme can replace the traditional single-layer delay line method with lower hardware costs.

[0198] Considering the hardware limitations of the phase shifter, Figure 4 The given condition is that each phase shifter takes 2... b The achievable rate at different iteration numbers when there are discrete values ​​and b=1. The resolution of RIS is set to Q=1, and other parameters are set the same. Figure 3After three iterations, the reachability tends to converge, proving the effectiveness of the approximate projection method. Furthermore, it was observed that the performance of using a discrete phase shifter is lower than that of a continuous phase shifter, but achieving near-optimal performance by reducing hardware complexity is acceptable.

[0199] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A terahertz multi-intelligent reflector communication beam optimization method based on two-layer time delay lines, characterized in that, The steps are as follows: Step 1: Construct a distributed RISS-assisted wideband terahertz MIMO communication system, where the BS includes A uniform linear array of antennas and Each radio frequency chain is Provide services to a single-antenna user, and ; Step 2: Use two layers of time delay lines at the BS end to generate an analog beamforming matrix; First, it will be by A subarray consisting of several antennas is connected to the delay line of the first layer; then... The subarrays are synthesized into a primary array with a smaller aperture, and then connected to the delay line of the second layer; finally, by A large-aperture array is composed of primary arrays; wherein, the number of bits of the delay lines in the first layer is... The number of bits in the second-layer delay line is ; deploy One RISS, and defined This serves as the index set for RISs; all RISs have the same size, and each RIS consists of a uniform planar array with a number of cells. ,in, This indicates the row number of each RIS. Represents the number of columns for each RIS; definition For the set of element indices on the row, For the set of element indices on the column; Assuming the number of radio frequency chains Equal to the number of RIS, i.e. Each RIS is served by a beam generated by an RF chain; reliable broadband transmission is achieved using M subcarrier orthogonal frequency division multiplexing; the first... The frequencies corresponding to each subcarrier are , ,in, For the center frequency, It's bandwidth; BS and the The user in the first Equivalent channel on each subcarrier Represented as: ； in, It is the first BS to the first subcarrier Channel matrix between RIS It is the first On the subcarrier The user and the first Channels between RIS It is the first The diagonal matrix of reflection coefficients of each RIS , , , ; Indicates the amplitude of reflection. Indicates the phase shift of the reflection; Based on the equivalent channel Calculate the first The user in the first Received signal on each subcarrier : ； in, A simulated beamforming matrix implemented for a two-layer time delay line structure; Both represent digital beamforming vectors. It is the first The first subcarrier Additive white Gaussian noise for each user For variance; For the first The first subcarrier Information about each user For the first The first subcarrier Information about each user; ； in, , express The phase shifter passes through the first The delay line and the first The delay line is connected to the first... Beamforming vectors generated by each radio frequency chain; This represents the frequency-shifted phase shift implemented by the first-layer delay line network, namely: ； in, It is connected to the first The first radio frequency chain and the first Offline delay The delay vector is implemented by a delay line; The frequency-shifted phase shift implemented for the second-layer delay line network is as follows: ； in, It is connected to the first Under one radio frequency chain The delay vector is implemented by a delay line; Step 3: Calculate the user's reachable rate based on the simulated beamforming matrix, and use maximizing the user's reachable rate as the objective function; Step 4: Design a joint beamforming model. Based on the joint beamforming model and the alternating iterative optimization algorithm, optimize the objective function to obtain the reflection coefficients of digital beamforming and RIS.

2. The terahertz multi-intelligent reflector communication beam optimization method based on two-layer time delay lines according to claim 1, characterized in that, The channel matrix The expression is: ； in, Indicates the number of paths. Indicates reaching the th The first RIS at the first Gain of each path Indicates reaching the th The first RIS at the first The delay of the path, This represents the array response vector at point BS. for of……, From BS to the The first RIS The physical direction of the path, This represents the array response vector at RIS. Indicates the first The path in the first The azimuth angle of the RIS arrival angle Indicates the first The path in the first The elevation angle of the RIS arrival angle; The , They are represented as follows: ； ； in, It's the speed of light. It is the distance between two consecutive antennas. , Indicates the corresponding center frequency The wavelength; n represents the nth antenna element; The first On the subcarrier The user and the first Channel between RIS for: ； in, Indicates the number of paths. Indicates from the first RIS to the first The first user's Gain of each path Indicates from the first RIS to the first The first user's Delay of the path; Let be the emitter array response vector at RIS, expressed as: ； in, Indicates from the first RIS to the first The first user's The azimuth angle of the path departure angle. Indicates from the first RIS to the first The first user's The angle of elevation of the path away from the angle.

3. The terahertz multi-intelligent reflector communication beam optimization method based on two-layer time delay lines according to claim 1 or 2, characterized in that, The achievable speed for the user is: ； in, For the first The user in the first SINR on each subcarrier.

4. The terahertz multi-intelligent reflector communication beam optimization method based on two-layer time delay lines according to claim 3, characterized in that, The objective function is obtained as follows: For the delay lines of the first layer and the delay lines of the second layer, a uniform quantization method is used to generate them respectively. and A discrete value; therefore, a discrete time delay value and They are represented as follows: ； ； Where D represents the delay step size; For a phase shifter, a b-bit quantization phase shift is set, and the beamforming vector generated by the phase shifter... for: ； The reflection coefficient of RIS Discretize to obtain discrete values: ； in, This indicates that each RIS contains There are discrete phase shifts; therefore, the objective function to maximize the achievable rate is: ； in, , This represents the maximum transmission power of the base station.

5. The terahertz multi-intelligent reflector communication beam optimization method based on two-layer time delay lines according to claim 4, characterized in that, The method for optimizing the objective function is as follows: S41. Design a joint beamforming model: The first layer is calculated according to the following formula. The delay line and the second layer Optimal delay for each delay line: ； ； in, This represents the time delay between two consecutive antennas. Indicates the direction of the target; Based on the approximation principle, the solution is... , Mapped to sets respectively and The elements in the array are represented as follows: ； ； in, yes The elements in yes The elements in the above mapping can be used to obtain a suboptimal solution for two delay lines; Due to the array response vector It is of equal amplitude and phase, and can be used as The column, in which , We obtain it from the following formula: ； in, At the center frequency From BS to the The first RIS The spatial direction of the path; This represents the array response vector; similarly, the solved phase shift... Projected onto a set of discrete values It can be written as: ； The suboptimal solution for the phase shifter can be obtained from the above equation; S42, Fixed optimization : In the given In this case, the objective function P1 can be reformulated as an optimization problem P2: ； For the The user in the first The equivalent channel vector on each subcarrier can be written as: Based on the extension of the Sherman-Morrison-Woodbury formula, we obtain: ； Therefore, optimization problem P2 can be transformed into P3: ； P3 and They are respectively: ； ； in, ； ； P3 is solved by a numerical convex programming solver until convergence, thus obtaining the local optimal solution of the digital beamforming vector. S43, Fixed optimization : because Transform P3 into an optimization problem P4: ； Assumption And define a vector containing the reflection coefficients of all units, i.e. ;set up Bit; Transform the optimization problem P4 into P5: ； Design a CUA algorithm for the reflection matrix, and optimize it during the optimization process. , where i represents the current iteration number and i−1 represents the previous iteration number; P5 is solved using the CUA algorithm for the reflection matrix until... Convergence, output reflection coefficient matrix .

6. The terahertz multi-intelligent reflector communication beam optimization method based on two-layer time delay lines according to claim 5, characterized in that, The method for solving P5 using the CUA algorithm based on the reflection matrix is ​​as follows: S43.1, Input: Channel matrix Analog precoding Digital precoding Number of iterations ; S43.2 Initialization: ; S43.3, Order ,but ; S43.4, Transform the vector Transform into reflection matrix ; S43.5, Order ,but ; S43.6, Transform the vector Transform into reflection matrix ; S43.7, Using the formula Calculate the value of q and assign it. ; S43.

8. Repeat steps S43.3 to S43.7 until the maximum number of iterations is reached, and output the reflection coefficient matrix. .

Images