A UAV-assisted NOMA two-way relay communication network and rate maximization method

Abstract

The present invention claims a non-orthogonal multiple access (NOMA) two-way relay communication system and rate maximization method assisted by an unmanned aerial vehicle (UAV), belonging to the field of NOMA network resource allocation, maximizing the sum rate of the UAV-assisted NOMA two-way relay network under the constraints of NOMA decoding order, UAV transmission power, and user transmission power. Its innovation lies in that, unlike the traditional NOMA two-way relay network, the present invention considers the impact of hardware damage on system performance. The present invention adopts variable replacement, alternating optimization, and successive convex approximation (SCA) to design a resource allocation scheme. The method provided by the present invention improves the system sum rate compared to other resource allocation schemes that do not consider hardware damage and optimize the position of the UAV, and has better practicality and feasibility.

Description

Technical Field

[0001] The present invention relates to the technical field of NOMA network power control, and in particular to a UAV-assisted NOMA bidirectional relay communication network and a rate maximization method under hardware damage conditions. Background Art

[0002] As a new type of communication method, UAV-assisted wireless communication has attracted widespread attention from academia and industry. UAVs can be deployed as aerial communication platforms, such as flying base stations or mobile relay stations. Compared with ground communication facilities, UAV-assisted communication has more outstanding advantages. For example, in most cases, UAVs communicate directly or relay with ground users through long-distance strong line-of-sight links, usually with higher channel gain; UAVs can dynamically adjust their positions according to real-time needs to better adapt to the communication environment, while having lower investment costs. These outstanding advantages and characteristics make UAV-assisted communication a promising solution for achieving air-ground integration. NOMA technology has the advantages of large-scale connection, low latency, and high spectrum efficiency, and is regarded as a key technology for the next generation of wireless communication technology. Compared with traditional orthogonal multiple access technology, the power domain NOMA system allows multiple users to share the same time-frequency resources through power domain multiplexing, breaking the limitation of the number of access users on the orthogonal division of communication resources. In order to further improve the spectrum efficiency, the researchers studied the bidirectional relay forwarding protocol based on the physical layer network coding technology. After adopting physical layer network coding, the communication system with relay nodes can complete the complete mutual transmission of uplink and downlink in two or three time slots, which greatly improves the spectrum efficiency compared with the four time slots of the one-way relay system.

[0003] At present, through the study of existing resource allocation methods in UAV-assisted two-way relay networks, it is found that the existing methods have two main problems, which makes the methods not very practical. First, the scenarios considered in the study are mostly based on the ideal hardware conditions of the RF transceiver at the node. For example, Yipeng Liang et al. published an article entitled "Joint Trajectory and Resource Optimization for UAV-Aided Two-Way Relay Networks" in "IEEE Transactions on Vehicular Technology, 2022, 71 (1): 639-652.", which studied the system and rate of the UAV-assisted two-way relay system, but did not consider the impact of hardware damage on the system at the transceiver end. This type of simple scenario design method cannot be applied to actual networks. Secondly, the method is too complex, resulting in poor network scalability. For example, Lei Li et al. proposed a system and rate optimization method based on a penalty method in the article "UAV Positioning and Power Control for Two-Way Wireless Relaying" published in "IEEE Transactions on Wireless Communications, 2020, 19 (2), 1008-1024."

[0004] Therefore, considering the hardware damage conditions of the transceiver and the layout of the UAV, it is of great practical application value and significance to study the method of maximizing the sum rate of the UAV-assisted NOMA two-way relay communication network. Since the sum rate optimization problem of the UAV-assisted NOMA two-way relay communication network is a hybrid binary non-convex optimization problem, how to design a low-complexity resource allocation method for this problem is a technical difficulty. Patent CN114760695A "A method for maximizing the sum rate of a non-orthogonal multiple access system assisted by a UAV under hardware damage conditions" maximizes the sum rate of the system by jointly optimizing the location deployment of the UAV and the power allocation of the user. Different from patent CN114760695A, which only considers the downlink communication application scenario where the UAV is used as a base station to send information to a single-cell ground user. The present invention considers the application scenario where the UAV is used as a two-way relay for information exchange between users in two different cells. In this application scenario, it is necessary to consider the uplink capacity limit from the user to the UAV and the downlink capacity limit from the UAV to the user at the same time. At the same time, the objective function of the system sum rate maximization problem is a non-differentiable function. How to design a resource allocation solution for such a non-differentiable non-convex optimization problem is a major challenge. To this end, the patent of the present invention introduces auxiliary variables into the objective function, adds the objective function to the constraints, first converts the objective function into a continuous differentiable function, and then designs a resource optimization plan through variable replacement, continuous convex approximation, and alternating optimization methods.

[0005] The present invention aims to solve the above problems of the prior art. A UAV-assisted NOMA bidirectional relay communication network and a rate maximization method are proposed. The technical solution of the present invention is as follows:

[0006] A UAV-assisted NOMA bidirectional relay communication network and rate maximization method, comprising the following steps:

[0007] Step 1) Establish a UAV-assisted NOMA two-way relay communication network and rate maximization model under hardware damage conditions;

[0008] Step 2), initialize the drone position, NOMA decoding order, user transmit power, drone transmit power and minimum rate decision threshold;

[0009] Step 3) Using the given user transmission power and drone power, the drone position and NOMA decoding order are solved and updated;

[0010] Step 4), using the given drone position and NOMA decoding order, solve the user transmission power and drone power allocation coefficient and update them;

[0011] Step 5), judge the convergence of the sum rate update, calculate the updated sum rate value, if the absolute value of the difference between the updated sum rate and the previous sum rate is not greater than the sum rate decision threshold, then judge that the sum rate converges, give the maximum sum rate value, output the drone position, NOMA decoding order coefficient, user transmission power and drone power allocation coefficient, and the method ends; if the absolute value of the difference between the updated sum rate and the previous sum rate is greater than the sum rate decision threshold, then save the newly calculated sum rate value as the sum rate value at this time, and go to step 2) to update the system parameters until the sum rate meets the convergence conditions and gives the maximum sum rate.

[0012] Furthermore, the step 1) establishes a sum rate maximization model of a drone-assisted NOMA bidirectional relay network as follows:

[0013]

[0014]

[0015] Where R represents the sum rate of the system, S k To represent the rate from user k to user 2K-k+1, p k is the transmission power of user k, p represents p k The vector composed of UAV is the maximum transmission power of the UAV, β k is the power allocation coefficient allocated by the UAV to user k, and β represents β k The three-dimensional coordinates of the drone can be expressed as [X U ,Y U ,H] T , H is the height of the drone, X U and Y U is the horizontal coordinate position of the UAV, a k,i 、a i,k 、a i,h 、a k,h 、b k,i 、b i,k 、b i,h 、b k,h represents the control coefficient of the SIC decoding order in the uplink and downlink, and is a binary variable, A represents a k,i The matrix of B represents b k,i The matrix of Represents the drone in [X U ,Y U ,H] T The channel gain at and They are the hardware damage coefficients of the user sending end and the drone receiving end, and they obey and and Represent the hardware damage coefficients of the transmitter and receiver respectively, n is additive white Gaussian noise, and obeys β k is the power allocation coefficient allocated to user k, r min is the minimum rate constraint, k and 2K represent the kth user and the number of users respectively, i and j represent the ith user and the jth user respectively, δ 2 is the system noise power; the P1 objective function R represents the sum rate, which is the sum rate of all users. The maximum rate of each user is equivalent to the maximum value of the two-stage rate from user to UAV and from UAV to target user. Constraint C1a is the user transmission power constraint, C1b is the UAV power allocation coefficient constraint, C1c-C1h are the SIC decoding order coefficient restriction constraints, and the alternating optimization method is used to convert the P1 optimization problem into three sub-problems for solution.

[0016] Furthermore, in step 2), the drone position is initialized to q 0 =(X,Y,H), the UAV transmission power allocation coefficient is β 0 , Uplink decoding order coefficient A 0 , Downlink decoding order coefficient B 0 , user transmit power p 0 , the rate decision threshold is ε, where X, Y, and H represent the horizontal coordinate, vertical coordinate, and flight altitude of the UAV, respectively.

[0017] Furthermore, in step 3), using the given drone position and uplink and downlink decoding order coefficients, the user transmission power and drone power allocation coefficients are solved and updated, specifically including:

[0018] Fix q, A and B, jointly optimize p and β, and the subproblem P2 is:

[0019]

[0020]

[0021] C2d:a k,i ,b k,i ∈{0,1}

[0022] C2e:a k,i +a i,k =1

[0023] C2f:b k,i +b i,k =1

[0024]

[0025] C2h:a k,i +a i,k =1

[0026] C2i:a k,k =0,b k,k =0

[0027] C2j:a k,i a i,h ≤a k,h

[0028] C2k:b k,i b i,h ≤ b k,h

[0029] C2l:S k ≥r min

[0030] The P1 objective function uses the auxiliary variable S k Replace, and get the objective function of P2; New constraints introduced based on auxiliary variables and

[0031] Furthermore, the subproblem P2 is a non-convex optimization problem, and the objective function needs to be transformed to make the subproblem P2 equivalent to P3:

[0032]

[0033]

[0034] C3e:Z i ≤logp max

[0035]

[0036] C3h:a k,i +a i,k =1

[0037]

[0038] C3k:S k ≥r min

[0039] Among them, M, V, and Z are introduced auxiliary variables. Since constraint C2a is a non-convex constraint, we can take the logarithm of both sides of the inequality of constraint C2a and move the terms, which can be rewritten as: in ρ 0is the channel gain when the reference distance is one meter, and σ is the replaced variable; due to the existence of the form of variable product, constraint C2a is a non-convex constraint. In order to solve this problem, variable substitution is used for processing. Let Then constraint C1 can be transformed into: because There is a form of convex function subtraction, so it is transformed by first-order Taylor expansion as follows: Simplifying the constraint C2b, it can be converted to This inequality is a convex constraint and can be calculated using a convex optimization algorithm;

[0040] For constraint C2c, it can be equivalent to three constraints: a k,i ≤0, a k,i ≥1, constraint C2i, can be rewritten as Since there is a difference in convex functions in this formula, it is still a non-convex constraint; Performing a first-order Taylor expansion at can be converted to: At this time, the left side of the inequality sign of the constraint is in the form of a convex function plus a convex function, which is still a convex function. Therefore, it is expressed as a convex constraint. At this time, problem (P3) and problem (P2) are equivalent. Problem (P3) is a convex optimization problem, and its optimal solution can be solved using convex optimization tools.

[0041] Furthermore, the above-mentioned fixed p and β, jointly optimizing q, A and B, sub-problem 2 P4 is:

[0042]

[0043]

[0044] C4c:a k,i ,b k,i ∈{0,1}

[0045] C4d:a k,i +a i,k =1

[0046] Taking the logarithm of both sides of the inequality of constraint C4a and moving the terms, C4a can be rewritten as Since the constraint is still a non-convex constraint, variable substitution is used, Then C4a can be transformed into a convex constraint At the same time, the constraints and

[0047] By taking logarithms on both sides of the constraint C4b inequality, C4b can be rewritten as It is still a non-convex constraint, which can be transformed by variable substitution. get

[0048] Based on the above derivation process, the original problem is transformed into the following problem:

[0049]

[0050]

[0051] Problem (P4) is a convex optimization problem, and standard convex optimization methods can be used to find the optimal solution.

[0052] Further, in the step 5), compare |S (n+1) -S (n) | and the size of the rate decision threshold ε, where S (n+1) is the safe rate after iteration n+1 times, S (n) is the system sum rate after n iterations; if |S (n+1) -S (n) | is not greater than ε, the system and rate converge, give the maximum system and rate, output the optimized drone position, NOMA decoding order coefficient, user transmission power and drone power allocation coefficient, and end; if |S (n+1) -S (n) | is greater than ε, the newly calculated system sum rate is saved as the safe rate at this time, and go to step 3) until the system sum rate meets the conditions and the maximum system sum rate is given.

[0053] The advantages and beneficial effects of the present invention are as follows:

[0054] The present invention maximizes the sum rate of the UAV-assisted NOMA bidirectional relay network under the constraints of the UAV transmission power allocation coefficient, NOMA decoding order, and user transmission power. Its innovation lies in that, compared with the traditional ideal hardware model, it takes into account the impact of hardware damage on the communication system, and adopts different hardware damage coefficients for users and UAVs, receivers and transmitters respectively. The model considered in this patent is an application scenario for information interaction between users in two cells. Compared with the single-cell scenario where the UAV is used as a base station for communication, the proposed scenario has a wider application range and has a higher system sum rate than the traditional model. For the non-convex problem P1 of multi-variable coupling, the present invention adopts alternating optimization, continuous convex approximation and variable substitution to solve it, decomposes P1 into two sub-problems P2 and P3, and obtains the optimal solution through iterative optimization. Compared with other resource allocation schemes of NOMA bidirectional relay networks without UAV assistance, the method provided by the present invention improves the safety rate and has better practicality and feasibility. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 The present invention provides a system model of a UAV-assisted NOMA bidirectional relay network in a preferred embodiment;

[0056] Figure 2 It is the convergence diagram of the present invention under the condition of maximum user transmission power and fixed position of the UAV;

[0057] Figure 3 It is the impact of the present invention on the two-way relay communication system and rate at different altitudes of the UAV;

[0058] Figure 4 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0059] The following will describe the technical solutions in the embodiments of the present invention in detail in conjunction with the accompanying drawings in the embodiments of the present invention. The described embodiments are only part of the embodiments of the present invention.

[0060] The technical solution of the present invention to solve the above technical problems is:

[0061] The present invention Figure 4 A UAV-assisted NOMA two-way relay communication network and a rate maximization method under hardware damage conditions are disclosed, which comprises the following steps:

[0062] Step 1: Establish a UAV-assisted NOMA two-way relay communication network and rate maximization model.

[0063] Step 2: Initialize the drone position, NOMA decoding order, user transmit power, drone transmit power, and minimum rate decision threshold;

[0064] Step 3: Using the given user transmit power and drone power, the drone position and NOMA decoding order are solved and updated;

[0065] Step 4: Using the given drone position and NOMA decoding order, solve the user transmit power and drone power allocation coefficient and update them;

[0066] Step 5: Determine whether the sum rate update converges, calculate the updated sum rate value, if the absolute value of the difference between the updated sum rate and the previous sum rate is not greater than the sum rate decision threshold, then determine that the sum rate converges, give the maximum sum rate value, and the method ends; if the absolute value of the difference between the updated sum rate and the previous sum rate is greater than the sum rate decision threshold, then save the newly calculated sum rate value as the sum rate value at this time, and go to step 2) to update the system parameters until the sum rate meets the convergence condition and gives the maximum sum rate;

[0067] Furthermore, in the first step, the UAV-assisted NOMA two-way relay communication network and rate maximization model (P1) is:

[0068]

[0069]

[0070] Where R represents the system and rate, S k To represent the rate from user k to user 2K-k+1, p k is the transmission power of user k, β k is the power allocation coefficient of the UAV allocated to user k, and the three-dimensional coordinates of the UAV can be expressed as [X U ,Y U ,H] T , a k,i and b k,i Represents the control coefficient of the SIC decoding order in uplink and downlink and is a binary variable. and They are the hardware damage coefficients of the user sending end and the drone receiving end, and they obey and n is additive white Gaussian noise and obeys P UAV is the UAV transmission power, β k is the power allocation coefficient allocated to user k, r min is the minimum rate constraint, k and 2K represent the kth user and the number of users respectively, i and j represent the ith user and the jth user respectively, δ 2 is the system noise power. The P1 objective function R represents the sum rate, which is the sum rate of all users. The maximum rate of each user is equivalent to the maximum value of the two-stage rate from the source user to the UAV and from the UAV to the target user. Constraint C1a is the user transmission power constraint, C1b is the UAV power allocation coefficient constraint, and C1c-C1h is the SIC decoding order coefficient restriction constraint. The alternating optimization method is used to convert the P1 optimization problem into three sub-problems for solution.

[0071] Furthermore, in the second step, the drone position is initialized to q=(X, Y, H), the drone transmit power allocation coefficient is β, the uplink decoding order coefficient is A, the downlink decoding order coefficient is B, the user transmit power is p, and the rate decision threshold is ε, where x, y, and H represent the horizontal coordinate, vertical coordinate, and flight altitude of the drone, respectively.

[0072] Furthermore, in the third step, the user transmission power and the UAV power allocation coefficient are solved and updated by using the given UAV position and the uplink and downlink decoding order coefficients, which specifically includes:

[0073] Fix q, A and B, jointly optimize p and β, and the subproblem P2 is:

[0074]

[0075]

[0076] C2d:a k,i ,b k,i ∈{0,1}

[0077] C2e:a k,i +a i,k =1

[0078] C2f:b k,i +b i,k =1

[0079]

[0080] C2h:a k,i +a i,k =1

[0081] C2i:a k,k =0,b k,k =0

[0082] C2j:a k,i a i,h ≤a k,h

[0083] C2k:b k,i b i,h ≤ b k,h

[0084] C2l:S k ≥r min

[0085] The P1 objective function uses the auxiliary variable S k Replace, and get the objective function of P2; New constraints introduced based on auxiliary variables and

[0086] Subproblem P2 is a non-convex optimization problem. The objective function needs to be transformed to make subproblem P2 equivalent to P3:

[0087]

[0088]

[0089] C3e:Z i ≤logp max

[0090]

[0091] C3h:a k,i +a i,k =1

[0092]

[0093] C3k:S k ≥r min

[0094] Among them, M, V, and Z are introduced auxiliary variables. Since constraint C2a is a non-convex constraint, we can take the logarithm of both sides of the inequality of constraint C2a and move the terms, and it can be rewritten as: in Since there is a form of variable product, constraint C2a is a non-convex constraint. In order to solve this problem, variable substitution can be used. Then constraint C1 can be transformed into: because There is a form of convex function subtraction, so it is transformed by first-order Taylor expansion as follows: Simplifying the constraint C2b, it can be converted to This inequality is a convex constraint and can be computed using a convex optimization algorithm.

[0095] For constraint C2c, it can be equivalent to three constraints: a k,i ≤0, a k,i ≥1, constraint C2i, can be rewritten as Since there is still a difference in convex functions in this formula, it is still a non-convex constraint. Performing a first-order Taylor expansion at can be converted to: At this time, the left side of the inequality sign of the constraint is a convex function plus a convex function, which is still a convex function, so it is expressed as a convex constraint. At this time, problem (P3) and problem (P2) are equivalent; problem (P3) is a convex optimization problem, and its optimal solution can be solved using convex optimization tools.

[0096] Furthermore, fix p and β, jointly optimize q, A and B, and the second sub-problem P4 is:

[0097]

[0098]

[0099] C4c:a k,i ,b k,i∈{0,1}

[0100] C4d:a k,i +a i,k =1

[0101] Taking the logarithm of both sides of the inequality of constraint C4a and moving the terms, C4a can be rewritten as Since this constraint is still a non-convex constraint, variable substitution is used, Then C4a can be transformed into a convex constraint At the same time, the constraints and

[0102] By taking logarithms on both sides of the constraint C4b inequality, C4b can be rewritten as It is still a non-convex constraint, which can be transformed by variable substitution. get

[0103] Based on the above derivation process, the original problem is transformed into the following problem:

[0104]

[0105]

[0106] Problem (P4) is a convex optimization problem, and standard convex optimization methods can be used to find the optimal solution.

[0107] Furthermore, in the sixth step, compare |S (n+1) -S (n) | and the size of the rate decision threshold ε, where S (n+1) is the safe rate after iteration n+1 times, S (n) is the system sum rate after n iterations; if |S (n+1) -S (n) | is not greater than ε, the system and rate converge, giving the maximum system and rate, and then the process ends; if |S (n+1) -S (n) | is greater than ε, the newly calculated system sum rate is saved as the safe rate at this time, and go to step 3) until the system sum rate meets the conditions and the maximum system sum rate is given.

[0108] The present invention maximizes the sum rate of the UAV-assisted NOMA two-way relay network under the constraints of UAV transmission power allocation coefficient, NOMA decoding order, UAV layout, and user transmission power. The innovation of the present invention lies in that, compared with the traditional model, the impact of hardware damage on the communication system is taken into account. Compared with the traditional model, it has a higher system sum rate. In the step 5, the UAV layout and the uplink and downlink NOMA decoding order are jointly optimized. The present invention adopts alternating optimization, continuous convex approximation and variable replacement to design a resource allocation scheme. Compared with other resource allocation schemes of NOMA two-way relay networks without UAV assistance, the method provided by the present invention improves the safety rate and has better practicality and feasibility.

[0109] This embodiment is a UAV-assisted NOMA two-way relay communication network and a rate maximization method. In a UAV-assisted NOMA two-way relay network, there are K = 10 ground users distributed in two cells. Cell 1 and cell 2 are circular ranges with a radius of 50m and a three-dimensional Cartesian coordinate system coordinate (100, 0, 0) and (-100, 0, 0) as the center, respectively. The initial coordinates of the drone are (0, 0, 200). The maximum transmission power of the ground user is p max =20dBm, the transmitting power of the drone is P UAV =30dBm. Reference channel power ρ 0 =-50dB, noise power spectrum density σ 2 =-174dBm / Hz, the flight altitude of the drone is H = 100m. Convergence threshold ε = 10 -4 .

[0110] In this embodiment, Figure 1 The present invention provides a system model in a UAV-assisted NOMA two-way relay communication network. In the figure, the UAV acts as an aerial relay to forward the signals of users in two cells. Both the ground user and the UAV are equipped with a single antenna. Figure 2 The sum rate convergence diagram obtained by using the method of this embodiment under the condition of maximum user transmission power and fixed drone position; Figure 3 Comparison diagram of the sum rate obtained by fixing the two-dimensional coordinates of the UAV and the method of this embodiment at different maximum transmission powers of the UAV;

[0111] from Figure 2 It can be seen from the figure that the method of this embodiment converges with the increase of the number of iterations. For the UAV with a maximum transmission power of P UAV =30dBm, the transmitting power of the ground user is p max =20dBm, the algorithm reaches the convergence condition after the third iteration, which shows that the method proposed in this paper has convergence and further verifies the effectiveness of the method proposed in this paper.

[0112] from Figure 3 It can be seen that when the number of ground users is 20, the sum rate increases with the increase of the maximum transmit power of the UAV. Under the same conditions, as the hardware damage coefficient increases, the system sum rate decreases, which shows that the UAV height and hardware damage will have a certain impact on the sum rate of the system. Compared with the case where the UAV two-dimensional coordinates are fixed, the sum rate obtained by the method of this embodiment is larger.

[0113] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0114] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0115] The above embodiments should be understood to be only used to illustrate the present invention and not to limit the protection scope of the present invention. After reading the contents of the present invention, technicians can make various changes or modifications to the present invention, and these equivalent changes and modifications also fall within the scope defined by the claims of the present invention.

Claims

1. A UAV-assisted NOMA bidirectional relay communication network and rate maximization method, characterized in that: The following steps are involved: Step 1) Establish a UAV-assisted NOMA two-way relay communication network and rate maximization model under hardware damage conditions; Step 2), initialize the drone position, NOMA decoding order, user transmit power, drone transmit power and minimum rate decision threshold; Step 3) Using the given user transmission power and drone power, the drone position and NOMA decoding order are solved and updated; Step 4), using the given drone position and NOMA decoding order, solve the user transmission power and drone power allocation coefficient and update them; Step 5), judge the convergence of the sum rate update, calculate the updated sum rate value, if the absolute value of the difference between the updated sum rate and the previous sum rate is not greater than the sum rate decision threshold, then judge the sum rate convergence, give the maximum sum rate value, output the drone position, NOMA decoding order coefficient, user transmission power and drone power allocation coefficient, and the method ends; if the absolute value of the difference between the updated sum rate and the previous sum rate is greater than the sum rate decision threshold, then save the newly calculated sum rate value as the sum rate value at this time, and go to step 3) to update the system parameters until the sum rate meets the convergence condition and gives the maximum sum rate; The step 1) establishes a sum rate maximization model of a drone-assisted NOMA bidirectional relay network as follows: Where R represents the sum rate of the system, S k To represent the rate from user k to user 2K-k+1, p k is the transmission power of user k, p represents p k The vector composed of UAV is the maximum transmission power of the UAV, β k is the power allocation coefficient allocated by the UAV to user k, and β represents β k The three-dimensional coordinates of the drone can be expressed as [X U ,Y U ,H] T , H is the height of the drone, X U and Y U is the horizontal coordinate position of the UAV, a k,i 、a i,k 、a i,h 、a k,h , b k,i , b i,k , b i,h , b k,h represents the control coefficient of the SIC decoding order in the uplink and downlink, and is a binary variable, A represents a k,i The matrix of B represents b k,i The matrix, h k (X,Y), Represents the drone in [X U ,Y U ,H] T The channel gain at and They are the hardware damage coefficients of the user sending end and the drone receiving end, and they obey and and Represent the hardware damage coefficients of the transmitter and receiver respectively, n is additive white Gaussian noise, and obeys β k is the power allocation coefficient allocated to user k, r min is the minimum rate constraint, k and 2K represent the kth user and the number of users respectively, i and j represent the ith user and the jth user respectively, δ 2 is the system noise power; the P1 objective function R represents the sum rate, which is the sum rate of all users. The maximum rate of each user is equivalent to the maximum value of the two-stage rate from the user to the UAV and from the UAV to the target user. Constraint C1a is the user transmission power constraint, C1b is the UAV power allocation coefficient constraint, and C1c-C1h is the SIC decoding order coefficient restriction constraint. The alternating optimization method is used to convert the P1 optimization problem into three sub-problems for solution; In the step 3), using the given drone position and uplink and downlink decoding order coefficients, the user transmission power and drone power allocation coefficients are solved and updated, specifically including: Fix q, A and B, jointly optimize p and β, and the subproblem P2 is: C2d:a k,i ,b k,i ∈{0,1} C2e:a k,i +a i,k =1 C2f:b k,i +b i,k =1 C2h:a k,i +a i,k =1 C2i:a k,k =0,b k,k =0 C2j:a k,i a i,h ≤a k,h C2k:b k,i b i,h ≤b k,h C2l:S k ≥r min The P1 objective function uses the auxiliary variable S k Replace, and get the objective function of P2; New constraints introduced based on auxiliary variables and The subproblem P2 is a non-convex optimization problem, and the objective function needs to be transformed to make the subproblem P2 equivalent to P3: C3e:Z i ≤logp max C3h: a k,i +a i,k =1 C3k:S k ≥r min Among them, M, V, and Z are introduced auxiliary variables. Since constraint C2a is a non-convex constraint, we can take the logarithm of both sides of the inequality of constraint C2a and move the terms, which can be rewritten as: in ρ0 is the channel gain when the reference distance is one meter, and σ is the replaced variable. Since there is a product of variables, constraint C2a is a non-convex constraint. In order to solve this problem, variable substitution is used. Let Then constraint C1 can be transformed into: because There is a form of convex function subtraction, so it is transformed by first-order Taylor expansion as follows: Simplifying the constraint C2b, it can be converted to This inequality is a convex constraint and can be calculated using a convex optimization algorithm; For constraint C2c, it can be equivalent to three constraints: a k,i ≤0, a k,i ≥1, constraint C2i, can be rewritten as Since there is a difference in convex functions in this formula, it is still a non-convex constraint; Performing a first-order Taylor expansion at can be converted to: At this time, the left side of the inequality sign of the constraint is in the form of a convex function plus a convex function, which is still a convex function, so it is expressed as a convex constraint; at this time, problem (P3) and problem (P2) are equivalent; problem (P3) is a convex optimization problem, and its optimal solution can be solved using convex optimization tools; The fixed p and β, joint optimization of q, A and B, sub-problem 2 P4 is: C4c:a k,i ,b k,i ∈{0,1} C4d:a k,i +a i,k =1 Taking the logarithm of both sides of the inequality of constraint C4a and moving the terms, C4a can be rewritten as Since this constraint is still a non-convex constraint, variable substitution is used, Then C4a can be transformed into a convex constraint At the same time, the constraints and By taking logarithms on both sides of the constraint C4b inequality, C4b can be rewritten as It is still a non-convex constraint, which can be transformed by variable substitution. get Based on the above derivation process, the original problem is transformed into the following problem: Problem (P4) is a convex optimization problem, and standard convex optimization methods can be used to find the optimal solution.

2. A UAV-assisted NOMA bidirectional relay communication network and rate maximization method according to claim 1, characterized in that: In the step 2), the drone position is initialized to q0=(X, Y, H), the drone transmit power allocation coefficient is β0, the uplink decoding order coefficient is A0, the downlink decoding order coefficient is B0, the user transmit power is p0, and the rate decision threshold is ε, where X, Y, and H represent the horizontal coordinate, vertical coordinate, and flight altitude of the drone, respectively.

3. A UAV-assisted NOMA two-way relay communication network and rate maximization method according to claim 1, characterized in that: In the step 5), compare |S (n+1) -S (n) | and the size of the rate decision threshold ε, where S (n+1) is the safe rate after iteration n+1 times, S (n) is the system sum rate after n iterations; If |S (n+1) -S (n) | is not greater than ε, the system and rate converge, give the maximum system and rate, output the optimized drone position, NOMA decoding order coefficient, user transmission power and drone power allocation coefficient, and end; if |S (n+1) -S (n) | is greater than ε, the newly calculated system sum rate is saved as the safe rate at this time, and go to step 3) until the system sum rate meets the conditions and the maximum system sum rate is given.

Images