Maximum throughput optimization method for uplink NOMA system assisted by smart reflector

Abstract

The present invention discloses a maximum throughput optimization method in a smart reflector-assisted uplink NOMA system, which is applicable to a network system including K users, a base station, and a smart reflector, and all devices are configured with a single antenna. The optimization method in the present invention is that K users send their information to the base station through a direct link with the base station and a reflection link of the smart reflector, and optimize the phase shift matrix of the smart reflector under the limitation of the user's transmission power to obtain the maximum throughput of all users, thereby improving the performance of the system. The present invention improves the throughput of the system and has great application value.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to a maximum throughput optimization method in an intelligent reflector-assisted uplink NOMA system. Background Art

[0002] Smart reflective surface (IRS) is a sub-wavelength artificial two-dimensional material, usually composed of metal, dielectric and adjustable elements, which can be equivalently characterized as a radio link control circuit. Smart reflective surface has a large number of low-power reflective elements, which are the core of smart reflective surface. The main function of smart reflective surface is to control the reflection phase and amplitude of the incident signal through software programming according to the communication link information fed back by signal propagation. It is essentially the adjustment of capacitance, resistance and inductance, so as to intelligently adjust the wireless environment, provide new spatial freedom for the system, achieve the purpose of enhancing the performance of wireless link, and pave the way for realizing intelligent and programmable wireless environment. According to the channel parameters, the reflected signal is constructed and added with the signal of other paths, which can enhance the expected signal power at the receiving end, improve the communication quality, enhance the capacity and expand the coverage. The basic idea of ​​non-orthogonal multiple access (NOMA) is to use non-orthogonal transmission at the transmitting end, actively introduce interference information, and achieve correct demodulation at the receiving end through serial interference cancellation (SIC) receiver, which solves the problem of spectrum resource shortage to a certain extent. Compared with traditional orthogonal multiple access technology, the significant feature of NOMA is that it allows multiple users to transmit different superimposed signals on the same spectrum and time resources through superposition coding and power multiplexing, and the signals are extracted at the receiving end using continuous interference cancellation. In addition, NOMA can ensure user fairness by allocating more transmission power to remote users and less power to near users.

[0003] Although there are many existing studies on the combination of smart reflective surfaces and NOMA, there is still a certain gap in the research on the benefits of smart reflective surfaces in uplink NOMA. Summary of the invention

[0004] In summary, in order to solve the existing technical problems, the technical solution adopted by the present invention is:

[0005] The maximum throughput optimization method in the uplink NOMA system assisted by a smart reflector is characterized by being applicable to a network system including K users, one base station, and one smart reflector, and all devices are configured with a single antenna. The method includes the following steps:

[0006] Step A: The intelligent reflection surface of the system has N reflection elements, there are K users, and the channel parameters from the kth user to the intelligent reflection surface, the intelligent reflection surface to the base station, and the kth user to the base station are respectively and Use Θ=diag(exp(θ1),exp(θ2),…,exp(θ N )) represents the phase matrix of IRS, where θ n is the phase of the nth reflective element, satisfying

[0007] The signal received by the base station is Where P k ≤P max is the transmission power of the kth user, which is less than the maximum transmission power P max , n is additive white Gaussian noise, x k is the unit power information of the kth user. After serial interference elimination, the signal-to-interference-noise ratio of the kth user can be obtained as According to Shannon's formula, the throughput of the kth user can be obtained as

[0008]

[0009] where σ 2 represents the noise power, then the sum throughput of K users can be expressed as

[0010]

[0011] Step B: By optimizing the phase shift matrix Θ to maximize the total throughput, the problem under consideration can be mathematically expressed as

[0012]

[0013]

[0014]

[0015] Where P = [P1, P2, ..., P K ] is the transmit power vector;

[0016] Step C: Solve the optimization problem in B and obtain the optimized sum throughput:

[0017] C1, due to the non-convex constraints in the objective function, it is difficult to directly obtain the optimal solution. The present invention adopts two methods to solve the phase shift matrix, namely Semidefinite Programming (SDP) and MajorizeMinimization (MM) to solve the phase matrix. First, because the logarithmic function is a monotonically increasing function, the objective function is equivalent to solving To simplify the optimization, we assume that all users use the maximum power P max , let d k =diag(g k )g0,θ=diag(θ) then the objective function will be transformed into

[0018]

[0019]

[0020] C2, use the Semidefinite Programming algorithm to solve the objective function (4), let f k =[d k ,h k ] H and Then formula (4) can be converted into

[0021]

[0022]

[0023] Expand the above formula make

[0024]

[0025]

[0026] in when When , we can get an upper bound value of (6) as λ(N+1) 2 , where λ is the maximum eigenvalue of the matrix F, a is the normalized eigenvector corresponding to the maximum eigenvalue, and there is an equation

[0027]

[0028] Therefore (6) can be transformed into

[0029]

[0030]

[0031]

[0032]

[0033] in is a square matrix with only one 1 at position (i, i) and all other 0s. Then, the optimized phase matrix can be obtained by solving problem (7) using the CVX toolbox.

[0034] C3 is solved using the Majorize Minimization method. First, Expand to

[0035]

[0036] in Then order Then the objective function (4) can be transformed into

[0037]

[0038]

[0039] However, it is still impossible to directly solve (10), so we use convex function to approximate the iterative solution, let

[0040]

[0041]

[0042] For any given θ obtained at the i-th iteration (i) and a feasible solution θ can approximate f(θ) as

[0043]

[0044] in

[0045] make The objective function can be transformed into

[0046]

[0047]

[0048] Remove the constant term ||θ|| 2 Then solve under the constraints That is, when θ and When they are in phase, the solution is obtained, that is,

[0049]

[0050] In step D, the optimized phase is brought into the objective function to obtain the maximum throughput.

[0051] Beneficial effects:

[0052] 1. The present invention enables K users to send their information to the base station through a direct link with the base station and a reflection link of an intelligent reflection surface, and optimizes the phase shift matrix of the intelligent reflection surface under the limitation of user transmission power to obtain the maximum throughput of all users and improve the performance of the system.

[0053] 2. The present invention improves the throughput of the system and has great application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 is a system model implemented by the present invention;

[0055] Figure 2 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0056] The present invention will be further described below in conjunction with the accompanying drawings.

[0057] like Figure 1 As shown, the present invention is a network system for K users, one base station, and one intelligent reflective surface, and all devices are configured with a single antenna. The intelligent reflective surface (IRS) of the system has N reflective elements, and the users have K:

[0058] The channel parameters from the kth user to the smart reflector, from the smart reflector to the base station, and from the kth user to the base station are and Use Θ=diag(exp(θ1),exp(θ2),…,exp(θ N )) represents the phase matrix of IRS, where θ n is the phase of the nth reflective element, satisfying

[0059] The signal received by the base station is Where P k ≤P max is the transmission power of the kth user, which is less than the maximum transmission power P max , n is additive white Gaussian noise, x k is the unit power information of the kth user. After serial interference elimination, the signal-to-interference-noise ratio of the kth user can be obtained as According to Shannon's formula, the throughput of the kth user can be obtained as

[0060]

[0061] where σ 2 represents the noise power, then the sum throughput of K users can be expressed as

[0062]

[0063] By optimizing the phase shift matrix Θ to maximize the total throughput, the problem under consideration can be mathematically expressed as

[0064]

[0065]

[0066]

[0067] Where P = [P1, P2, ..., P K ] is the transmit power vector.

[0068] Due to the non-convex constraints in the objective function, it is difficult to directly obtain the optimal solution. In order to effectively solve this problem, two methods for solving the phase shift matrix are selected, namely Semidefinite Programming (SDP) and Majorize Minimization (MM). First, because the logarithmic function is a monotonically increasing function, the objective function is equivalent to solving To simplify the optimization, we assume that all users use the maximum power P max Let d k =diag(g k )g0,θ=diag(θ) then the objective function will be transformed into

[0069]

[0070]

[0071] The SDP algorithm is used to solve the objective function (4), and f k =[d k ,h k ] H and Then formula (4) can be converted into

[0072]

[0073]

[0074] Expand the above formula make

[0075]

[0076]

[0077] in when When , we can get an upper bound value of (6) as λ(N+1) 2 , where λ is the maximum eigenvalue of the matrix F, and a is the normalized eigenvector corresponding to the maximum eigenvalue.

[0078]

[0079] Therefore, problem (6) can be transformed into

[0080]

[0081]

[0082]

[0083]

[0084] in is a square matrix with only one 1 at position (i, i) and all other 0s. Then, the optimized phase matrix can be obtained by solving problem (7) using the CVX toolbox.

[0085] To solve Θ using the MM method, first Expand to

[0086]

[0087] in Then order Then the objective function (4) can be transformed into

[0088]

[0089]

[0090] However, it is still impossible to directly solve (10), so we use convex function to approximate the solution iteratively, let

[0091]

[0092]

[0093] For any given θ obtained at the i-th iteration (i) and a feasible solution θ can approximate f(θ) as

[0094]

[0095] in

[0096] make The objective function can be transformed into

[0097]

[0098]

[0099] Remove the constant term ||θ|| 2 Then solve under the constraints That is, when θ and When they are in phase, the solution is obtained, that is,

[0100]

[0101] The optimized phases obtained by two different methods are brought into the objective function to obtain the maximum throughput.

Claims

1. A maximum throughput optimization method in an uplink NOMA system assisted by a smart reflector is applicable to a network system including K users, one base station, and one smart reflector, and all devices are configured with a single antenna, and the method is characterized in that: The steps include: Step A: The intelligent reflection surface of the system has N reflection elements, there are K users, and the channel parameters from the kth user to the intelligent reflection surface, the intelligent reflection surface to the base station, and the kth user to the base station are respectively and Use Θ=diag(exp(θ1),exp(θ2),…,exp(θ N )) represents the phase matrix of IRS, where θ n is the phase of the nth reflective element, satisfying The signal received by the base station is Where P k ≤P max is the transmission power of the kth user, which is less than the maximum transmission power P max , n is additive white Gaussian noise, x k is the unit power information of the kth user. After serial interference elimination, the signal-to-interference-noise ratio of the kth user can be obtained as According to Shannon's formula, the throughput of the kth user can be obtained as where σ 2 represents the noise power, then the sum throughput of K users can be expressed as Step B: By optimizing the phase shift matrix Θ to maximize the total throughput, the problem under consideration can be mathematically expressed as Where P = [P1, P2, ..., P K ] is the transmit power vector; Step C: Solve the optimization problem in B and obtain the optimized sum throughput: C1, due to the non-convex constraints in the objective function, it is difficult to directly obtain the optimal solution. Two methods for solving the phase shift matrix are used, namely Semidefinite Programming and Majorize Minimization. First, because the logarithmic function is a monotonically increasing function, the objective function is equivalent to solving To simplify the optimization, all users use the maximum power P max , let d k =diag(g k )g0,θ=diag(θ) then the objective function will be transformed into C2, use the Semidefinite Programming algorithm to solve the objective function (4), let f k =[d k ,h k ] H and Then formula (4) can be converted into Expand the above formula make in when When , we can get an upper bound value of (6) as λ(N+1) 2 , where λ is the maximum eigenvalue of the matrix F, a is the normalized eigenvector corresponding to the maximum eigenvalue, and there is an equation Therefore (6) can be transformed into in is a square matrix with only one 1 at position (i, i) and all other 0s. Then, the optimized phase matrix can be obtained by solving problem (7) using the CVX toolbox. C3 uses the Majorize Minimization method to solve Θ. First, Expand to in Then order Then the objective function (4) can be transformed into However, it is still impossible to directly solve (10), so we use convex function to approximate the solution iteratively, let For any given θ obtained at the i-th iteration (i) and a feasible solution θ can approximate f(θ) as in make The objective function can be transformed into Remove the constant term ||θ|| 2 Then solve under the constraints That is, when θ and When they are in phase, the solution is obtained, that is, In step D, the optimized phase is brought into the objective function to obtain the maximum throughput.

Images