Unmanned aerial vehicle green stereo coverage method based on non-orthogonal multiple access

Abstract

The application provides a kind of unmanned aerial vehicle green stereoscopic coverage method based on non-orthogonal multiple access, belongs to the energy efficiency optimization field of unmanned aerial vehicle communication technology.In the non-orthogonal multiple access unmanned aerial vehicle communication network, users are randomly distributed in a three-dimensional manner in a cellular cell, and unmanned aerial vehicle base stations are used to provide wireless coverage for users in the cell.Based on this model, the application provides a resource allocation algorithm for jointly optimizing user pairing, unmanned aerial vehicle trajectory and transmission power.This method can optimize the energy efficiency of the network and avoid waste of communication resources while meeting the user rate threshold.

Description

Technical Field

[0001] This invention belongs to the field of energy efficiency optimization of unmanned aerial vehicle (UAV) communication networks, and relates to a green three-dimensional coverage scheme for UAV communication networks based on non-orthogonal multiple access (NOMA). Specifically, it refers to jointly optimizing user matching, UAV trajectory and transmission power to maximize the energy efficiency of wireless communication networks. Background Technology

[0002] With the widespread adoption of the internet, communication networks have become an indispensable part of people's lives. From local area networks (LANs) to wide area networks (WANs), and now to the "Internet Plus" era, the development of communication networks has brought about tremendous changes and impacts. Today, people can use communication networks for various activities such as information transmission, data sharing, and remote work. In recent years, drone communication technology, as an emerging communication method, has the following advantages compared to traditional ground base stations:

[0003] 1) Flexibility: Drone base stations do not require a fixed network connection and can be deployed on either a mobile or fixed basis, offering greater flexibility. Compared to traditional ground base stations, drone base stations can be deployed and moved more quickly, adapting to various scenarios and needs.

[0004] 2) Wider Coverage: Drone base stations can be mounted on drones, leveraging the high altitude advantage of aircraft to achieve wider coverage and more stable signal transmission. This helps solve the problem of insufficient communication coverage and meets users' communication needs globally.

[0005] 3) Rapid Deployment: In environments such as natural disasters and emergencies, communication infrastructure is often severely damaged, leading to communication outages and hindering rescue efforts. The rapid deployment capability of drone base stations makes them a powerful tool for emergency communications. When traditional base stations are unable to operate normally, drone base stations can quickly take off and provide temporary communication support, contributing to the smooth progress of rescue operations.

[0006] 4) Improved communication speed: With the continuous growth of mobile communication data, the load on infrastructure is also increasing rapidly, often limiting communication speed. Drone base stations, as a supplement to mobile base stations, can further share the communication load, providing wider signal coverage, thereby improving data transmission speed and communication quality, and optimizing user experience.

[0007] 5) Line-of-sight wireless transmission: UAVs generally fly at high altitudes and when communicating with ground base stations or other UAVs, their channels usually maintain good line-of-sight (LoS) characteristics.

[0008] On the other hand, with the rapid development of wireless communication technology, especially the arrival of the 5G era, the demands for efficiency and capacity in wireless communication systems are increasing daily. Traditional Orthogonal Multiple Access (OMA) technology can no longer meet these needs. Therefore, NOMA technology is gradually emerging. With its higher spectral efficiency and greater system capacity, NOMA technology has become an important research direction for next-generation wireless communication systems. The basic principle of non-orthogonal multiple access technology is to allow multiple users to transmit signals on the same time, frequency, and channel simultaneously, thereby enabling simultaneous access for multiple users. Unlike traditional OMA technology, NOMA technology overcomes the limitations of orthogonal signals through non-orthogonal signal transmission, thus improving spectral efficiency and system capacity.

[0009] Unmanned aerial vehicle (UAV) communication networks utilizing NOMA technology can significantly increase the number of network accesses and coverage. Given the random distribution of ground user coordinates and altitudes, this invention aims to maximize energy efficiency by jointly optimizing user pairing, UAV trajectory, and transmission power, thereby achieving green wireless coverage. Summary of the Invention

[0010] To address the problems existing in current technologies, this invention provides a method for green three-dimensional coverage of unmanned aerial vehicles (UAVs) based on non-orthogonal multiple access (NOAMI) access, aiming to maximize the energy efficiency of UAV communication networks using NOAMI access. The UAV, flying at a fixed altitude, provides wireless communication services to several users within a ground area. The specific scheme is as follows: Figure 1 Based on this model, this invention provides a resource allocation algorithm that jointly optimizes user pairing, UAV trajectory, and transmission power, thereby maximizing the energy efficiency of wireless communication networks.

[0011] The technical solution adopted by this invention to solve the technical problem is as follows:

[0012] A method for green three-dimensional coverage by unmanned aerial vehicles (UAVs) based on non-orthogonal multiple access includes the following steps:

[0013] The first step is to build a system model:

[0014] 1) In a non-orthogonal multiple access (NMO) UAV communication network, a fixed-wing UAV flying at a fixed altitude acts as an aerial base station, providing wireless coverage to ground users. User altitudes are randomly distributed and represented using a three-dimensional Cartesian coordinate system. Any user u within the cell... k , The position can be represented as K represents the number of ground users. x(t) and y(t) are the two-dimensional coordinates of the aerial base station projected onto the ground at time t, respectively. Since optimizing continuous variables is difficult, the flight period T of the aerial base station is discretized into N time slots, each time slot having a length T. s Let T / N be the value. When T s When the time is sufficiently small, the location of the airborne base station can be considered constant within each time slot, thus discretizing the continuous-time variable.

[0015]

[0016] At the end of each service cycle, the aerial base station needs to return to its initial position. Let q represent the drone's trajectory; then q needs to satisfy...

[0017] q[0]=q[N].(2)

[0018] Furthermore, according to the laws of kinematics, the distance a ground station can travel in a time slot is limited by its velocity and acceleration at that moment. Therefore, q needs to satisfy the following constraints:

[0019] q[n+1]=q[n]+v[n]T s +a[n]T s 2 / 2,(3)

[0020] Where v[n] and a[n] are the velocity and acceleration of the fixed-wing UAV in the nth time slot, respectively. Due to the characteristics of the fixed-wing UAV, v[n] and a[n] also need to satisfy:

[0021]

[0022]

[0023] Among them, V min and V max Here, a represents the minimum and maximum speeds of the fixed-wing UAV, respectively. max This represents the maximum acceleration of the fixed-wing unmanned aerial vehicle.

[0024] Therefore, in the nth time slot, user u k Distance d from the airborne base station k [n] represents the following:

[0025]

[0026] When H is sufficiently high, the communication link between the airborne base station and the ground user can be approximated as line-of-sight transmission. Assuming the air-to-ground channel gain satisfies the free-space path loss model, the distance from the airborne base station to user u in the nth time slot... k The channel power gain is

[0027]

[0028] Where β0 is the power gain at a reference distance of 1m, it can be expressed as:

[0029]

[0030] Where c is the speed of light and f is the carrier frequency.

[0031] 2) Furthermore, due to limited spectrum resources, frequency band shortages may occur when the number of users in a cell is large. Therefore, NOMA technology is used to increase the number of network accesses. Specifically, each frequency band is shared by a group of two users, and each group uses orthogonal frequency division multiplexing. Within each time slot, the user with the larger channel gain in each group is considered a strong user, and the user with the smaller channel gain is considered a weak user. The strong and weak users in the m-th group are denoted as follows: and M represents the number of user groups, and the channel gains of the two groups in the nth time slot are denoted as follows: and The airborne base station pair in the nth time slot and The transmission power is denoted as follows: and According to NOMA, for both users to decode normally through Successive Interference Cancellation (SIC), the following constraints must be satisfied:

[0032]

[0033]

[0034] Among them, P max This represents the maximum transmit power of the airborne base station. Within the nth time slot... and Signal-to-interference-to-noise ratio and They can be represented as follows:

[0035]

[0036]

[0037] Among them, B s For bandwidth, σ 2 This represents noise density. According to Shannon's formula, and Instantaneous rate in the nth time slot and They are respectively represented as

[0038]

[0039]

[0040] The effective throughput Q of all users within one flight cycle T can be expressed as:

[0041]

[0042] On the other hand, the energy consumption of an airborne base station mainly consists of three parts: transmit power energy consumption P max Fixed energy consumption P of airborne base stations Base And the propulsion energy consumption P of fixed-wing UAVs Pro [n], the energy consumption model for UAV propulsion is as follows:

[0043]

[0044] Where c1 and c2 are constants related to the parameters of the fixed-wing UAV and air density, respectively, and g is the acceleration due to gravity. The energy consumption of the airborne base station during one flight cycle T is then calculated as follows:

[0045]

[0046] The second step is to simplify the objective function and list the optimization problem:

[0047] Considering both throughput and energy consumption, the energy efficiency (EE) of an airborne base station over one flight cycle T (defined as the ratio of throughput to energy consumption, in bits / J) is as follows:

[0048]

[0049] The goal of the proposed solution is to first determine the pairings of each user in the NOMA (Normally Accessible Mapping) system, and then optimize the trajectory of the over-the-air (OTA) base stations. Transmission power of the air-based access point To maximize the system's energy efficiency. Once user pairings are determined (refer to step 3 for the user pairing algorithm), the optimization problem can be formulated as follows:

[0050] (P1)

[0051]

[0052]

[0053] q[0]=q[N],(19d)

[0054]

[0055]

[0056]

[0057]

[0058]

[0059] Where, η s and η w These represent the rate thresholds for strong and weak users, respectively. This problem is a non-convex optimization problem with coupled variables, and cannot be solved directly using convex optimization theory. Therefore, an alternating optimization algorithm is employed for solution.

[0060] The third step is to solve the optimization problem:

[0061] First, a user pairing algorithm is proposed based on the K-Means clustering algorithm and the Gale-Shapley algorithm. Second, the original optimization problem is decomposed into two subproblems by individually optimizing each variable. Using the relevant theory of convex optimization, the two subproblems are transformed into convex optimization problems respectively, and the two subproblems are solved alternately to obtain at least one local optimum of the original problem.

[0062] 1) User pairing algorithm

[0063] Based on the K-Means clustering algorithm and the Gale-Shapley algorithm, a user pairing algorithm is proposed. First, the K-Means clustering algorithm is used to divide ground users into two user clusters. However, since the number of users in the two clusters may differ in the K-Means classification results, some users are selected from the cluster with more users and assigned to the other cluster, ensuring that both clusters contain an equal number of users. Then, the Gale-Shapley algorithm is used to pair users from the two clusters. According to the NOMA principle, the greater the channel difference between the paired users, the greater the NOMA gain. Therefore, the degree of preference between the two users is determined by the distance between them; the greater the distance, the greater the preference; the closer the distance, the lower the preference. The specific algorithm is as follows:

[0064]

[0065] In this algorithm, the K-Means clustering algorithm is used to divide ground users into two clusters, and then the number of users in the two clusters is kept equal. Next, the Gale-Shapley algorithm is used to pair users from the two clusters one-to-one, so that each user group contains two users. The user pairing performed by this algorithm takes into account the distance between the two users in each group, ensuring channel differences between strong and weak users and fairness among all users.

[0066] 2) Fix the UAV's transmit power and optimize the UAV trajectory q

[0067] Given the user pairing and transmit power, (P1) becomes:

[0068] (P2)

[0069] st(18b)-(18g),(20b)

[0070] (P2) is a non-convex fractional programming problem, because and Since q is non-concave, it is difficult to solve directly. Therefore, a first-order Taylor expansion can be used to solve for q. An approximation is performed to obtain its global lower bound. The trajectory of the UAV in the previous iteration is defined as... Then we have:

[0071]

[0072] Among them, A m and B m They can be calculated separately as follows:

[0073]

[0074]

[0075] Similarly, The transformation is as follows:

[0076]

[0077] Among them, C m and D m They can be calculated as follows:

[0078]

[0079]

[0080] In conclusion and They are and The global lower bounds of and both are concave with respect to q, using and Will and By substitution, (20a) becomes:

[0081]

[0082] (19b) and (19c) respectively become

[0083]

[0084]

[0085] However, the denominator of the objective function (P2) is still non-convex, so we introduce an auxiliary variable. It needs to meet the following:

[0086]

[0087] Then (P2) can be approximated as

[0088] (P3)

[0089]

[0090]

[0091]

[0092]

[0093] (18d)-(18e),(18g).(31f)

[0094] The objective value of (P3) is the global lower bound of (P2), and it is a standard form of concave-convex fractional programming problem, which can be solved using the Dinkelbach algorithm to introduce the parameter μ. This algorithm, proposed by the German mathematician Dinkelbach in 1967, is widely used for optimization problems under fractional programming. This algorithm can equivalently transform the fractional programming problem into a set of equivalent optimization problems in affine form, including energy efficiency parameters. The optimal energy efficiency of the original fractional programming problem is the zero point of the equivalent optimization problem. Therefore, (P3) can be approximated as the following convex problem:

[0095]

[0096] st(18b)-(18c),(18h)-(18i),(32b) where μ is a constant that is updated to the global optimum after each iteration. (P4) is a standard convex optimization problem that can be solved using existing optimization tools (such as the CVX toolbox).

[0097] 3) Fix the drone trajectory and optimize the drone's transmission power P

[0098] Given the user pairings and drone trajectory, (P1) becomes:

[0099] (P5)

[0100] st(18b)-(18c),(18h)-(18i),(33b) where, The value of P is non-concave, and it can be approximated by a first-order Taylor expansion to obtain its global lower bound. The transmit power of the UAV to the strong user in the previous iteration is defined as... Then we have:

[0101]

[0102] Among them, E m F m and G m They can be calculated separately as follows:

[0103]

[0104]

[0105]

[0106] Then (P5) can be approximated as the following convex problem:

[0107] (P6)

[0108]

[0109] (18b),(18h)-(18i).(38c)

[0110] The objective value of (P6) is the global lower bound of (P5), and it is a standard convex optimization problem that can be solved using existing optimization tools (such as the CVX toolbox).

[0111] 5) Algorithm design based on continuous convex approximation

[0112]

[0113] In this algorithm, the optimized value of the objective function increases with the number of iterations, and the maximum power limits the upper limit of the optimization objective. Therefore, the algorithm based on block coordinate descent and continuous convex approximation can converge to a stable point.

[0114] In this way, we maximized the energy efficiency of the wireless communication network by combining user matching, drone trajectory, and transmission power. This enabled green wireless network coverage even with users distributed randomly in a three-dimensional manner.

[0115] The beneficial effects of this invention are as follows:

[0116] This invention provides a user pairing and resource allocation scheme for UAV communication networks based on non-orthogonal multiple access. By rationally designing user matching, UAV trajectory and transmission power, the energy efficiency of the wireless communication network is maximized under the constraints. Attached Figure Description

[0117] Figure 1 This is a communication network for unmanned aerial vehicles (UAVs) based on non-orthogonal multiple access.

[0118] Figure 2 This section presents a convergence analysis of a resource allocation algorithm based on continuous convex approximation.

[0119] Figure 3 This is a pairing demonstration of a user pairing algorithm proposed based on the K-Means clustering algorithm and the Gale-Shapley algorithm.

[0120] Figure 4 Comparison of the airborne base station trajectories with optimal energy efficiency and optimal speed at T=100s. Figure 4 (a) in the diagram is a comparison of top-down views. Figure 4 (b) in the figure is a comparison diagram of three-dimensional angles.

[0121] Figure 5 Comparison of instantaneous velocity and instantaneous acceleration for optimal energy efficiency and optimal data rate of airborne base stations at T=100s.

[0122] Figure 6 The signal power radiation diagram of the airborne base station to the user when T=100s and the flight time is 60s.

[0123] Figure 7 Comparison of throughput, energy efficiency, and average rate of different schemes at T=100s.

[0124] Figure 8 This study compares the throughput and energy efficiency of different schemes over a period of time. Detailed Implementation

[0125] To better understand the above technical solution, a detailed analysis is provided below in conjunction with the accompanying drawings and specific implementation methods.

[0126] The first step is to build a system model:

[0127] 1) such as Figure 1 This is a model of an unmanned aerial vehicle (UAV) communication network system based on non-orthogonal multiple access. The cell radius is set to 2km, the number of users K is set to 400, and they are randomly distributed within a cell at an altitude of 0-300m. The flight altitude of the airborne base station is H = 1000m, the carrier frequency is f = 4.9GHz, and the speed of light is c = 3 * 10^6 times. 8 m / s, noise power spectral density is σ 2 =10 -20.4 W / Hz. Maximum transmit power of the UAV, P max =5W, and the parameter related to the drone's rushing energy consumption is c1 = 1.84 * 10. -3 c2 = 4450. Fixed energy consumption P of the drone base station. Base =160W, maximum flight speed V max =100m / s, minimum flight speed V min =10m / s², maximum flight acceleration a max =10m / s 2 The acceleration due to gravity is g = 9.8 m / s². 2 Assume the base station's fixed power consumption is 270W. Set the strong user threshold to η. s =1.6 Mbit / s, weak user threshold set to η w =0.6 Mbit / s. In the comparison scheme, the ground base station is located in the center of a hexagonal cell at a height of 150m, with a path loss factor of 2.35. In this model, the drone base station utilizes NOMA to provide simultaneous wireless coverage to all ground users.

[0128] The second step is to determine the objective function, list the constraints, and list the optimization problems based on the specific parameter settings in the first step.

[0129] The third step is to solve the optimization problem.

[0130] The proposed user pairing algorithm is used to pair ground users. Next, by fixing other variables while individually optimizing each variable, the original optimization problem is decomposed into two subproblems, and each subproblem is transformed into a convex problem. Finally, based on iteratively solving the two subproblems, Algorithm 2 is proposed, the specific process of which is as follows:

[0131]

[0132]

[0133] Figure 2The convergence analysis of Algorithm 2's energy efficiency maximization scheme under different flight cycles is presented. The optimization target energy efficiency increases steadily with the number of iterations and eventually tends to remain constant, indicating that the algorithm is convergent. From the first to the third iteration, the optimization target increases rapidly. Then, from the fourth to the fifth iteration, the upward trend gradually slows down, and finally, after the sixth iteration, it remains almost constant, converging to a local optimum. Furthermore, the length of the cycle also affects the optimization target; a longer cycle allows the UAV more time to fly to the optimal position, and correspondingly, the optimization target increases.

[0134] Figure 3 This demonstrates the distribution of ground users and some user pairings. Users are randomly distributed within cells with sides of 2km and heights ranging from 0 to 300m. User pairings are determined using user location and Algorithm 1. Figure 2 As can be seen, users paired by Algorithm 1 maintained a relatively large distance, which ensured the channel differences between strong and weak users within each group, thus contributing to higher performance improvements. Furthermore, there were no instances of two closely adjacent users being paired, which also guaranteed fairness among users.

[0135] Figure 4 (a) Figure 4 Figure (b) shows the optimized airborne base station trajectory for this example at T=100s, and also provides a comparison with the optimal speed scheme. As can be seen from the figure, the airborne base station trajectory under the optimal energy efficiency condition is approximately circular. This is because the curvature of a circular trajectory is minimal at all points, avoiding frequent acceleration / deceleration of the UAV and thus reducing energy consumption. In contrast, the optimal speed scheme, without limiting energy consumption, causes the airborne base station to linger longer in areas with higher communication rates and accelerate away from areas with lower communication rates. This increases throughput to some extent, but also significantly increases the UAV's propulsion energy consumption.

[0136] Figure 5 The figure shows the instantaneous velocity and acceleration at T=100s for optimal energy efficiency and optimal speed. As can be seen from the figure, under optimal energy efficiency, the acceleration remains consistently low, while the velocity remains relatively constant, thus the UAV maintains low propulsion energy consumption to maximize energy efficiency. Under optimal speed, the UAV's velocity fluctuates significantly between 10m / s and 100m / s, while the acceleration remains mostly around 10m / s. 2 Frequent acceleration and deceleration keep the drone's propulsion power at a high level, thus reducing energy efficiency.

[0137] Figure 6The diagram shows the signal power radiation received by a user with a flight time of 60 seconds at T=100s. It can be seen that the ground projection of the drone is around (4000, 4000). Users in this area are closer to the drone, have a higher channel gain, and are therefore considered strong users, receiving a lower transmission power from the drone. Conversely, users near (0, 0) are farther from the drone, have a lower channel gain, and are considered weak users, receiving a higher transmission power from the drone. The diagram also shows that the energy is mainly concentrated around (0, 0), which to some extent ensures fairness among users.

[0138] Figure 7 The throughput, energy efficiency, and average rate of different schemes are shown at T=100s. To ensure consistency of variables, the user rate threshold is set to 0.6 Mbit / s. We considered the following schemes:

[0139] ●Solution based on UAV and non-orthogonal multiple access: In this scheme, UAVs are used as airborne base stations, non-orthogonal multiple access is adopted, user pairing is determined by Algorithm 1, and the UAV trajectory and transmission power are optimized to maximize energy efficiency (denoted as NOMA-UAV).

[0140] ●Solution based on UAV and orthogonal multiple access: In this scheme, UAVs are used as airborne base stations, orthogonal multiple access is adopted, and the UAV trajectory and transmission power are optimized to maximize energy efficiency (denoted as OMA-UAV).

[0141] ●Solution based on terrestrial base stations and non-orthogonal multiple access: In this scheme, terrestrial base stations are used to communicate with terrestrial users. Non-orthogonal multiple access is adopted. User pairing is determined by Algorithm 1, and the base station transmission power is optimized to maximize energy efficiency (denoted as NOMA-BS).

[0142] ●Solution based on terrestrial base stations and orthogonal multiple access: In this scheme, terrestrial base stations are used to communicate with terrestrial users. Orthogonal multiple access is adopted, and the base station transmission power is optimized to maximize energy efficiency (denoted as OMA-BS).

[0143] As shown in the figure, the scheme based on UAVs and non-orthogonal multiple access (NOMA) is optimal in terms of both throughput, energy efficiency, and average data rate. Furthermore, the performance of the airborne base station scheme is significantly better than that of the ground base station scheme. In addition, the energy efficiency of NOMA-BS is slightly lower than that of OMA-BS. This is because, for the ground base station scheme, the channel gain difference between strong and weak users in NOMA is smaller, resulting in greater interference from strong users to weak users, which increases energy consumption to some extent.

[0144] Figure 8The throughput efficiency of the UAV-based and non-orthogonal multiple access (NOMA) schemes is significantly better than other schemes across different periods. Furthermore, the throughput efficiency of both UAV-based and non-orthogonal multiple access (NOMA) schemes increases with the flight period T. This is because UAVs are maneuverable, and an increased flight period allows them more time to fly near higher speed locations. Theoretically, increasing T also increases the energy consumption of the UAV-based scheme. When T is sufficiently large, throughput and energy consumption reach equilibrium, and throughput efficiency saturates. The throughput efficiency of the two terrestrial base station-based schemes does not change with the period T.

[0145] The above-described embodiments are merely illustrative of the implementation methods of the present invention, but should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the protection scope of the present invention.

Claims

1. A method for UAV green three-dimensional coverage based on non-orthogonal multiple access, characterized in that, The described UAV green three-dimensional coverage method increases the number of network accesses and coverage range through the UAV communication network using NOMA technology. Given the random distribution of ground user coordinates and altitudes, and with the goal of maximizing energy efficiency, an optimization problem is identified. This problem is solved by jointly optimizing user pairing, UAV trajectory, and transmission power to achieve green wireless coverage. The method includes the following steps: Once user pairings are confirmed, the optimization issues are as follows: (19a) (19b) (19c) (19d) (19e) (19f) (19g) (19h) (19i) in, η s and η w These are the rate thresholds for strong and weak users, respectively; Q represents the trajectory of the over-the-air base station. ; The transmit power of the air-to-ground access point. ; M N represents the number of user groups and the number of time slots; the first m Strong users and weak users in a group are denoted as follows: and , and In the n The instantaneous rates of each time slot are respectively and ; P max For transmission power consumption, P Base Fixed energy consumption for airborne base stations; P Pro [ n [This refers to the propulsion energy consumption of fixed-wing unmanned aerial vehicles;] T s The length of each time slot, T s =T / N, T For flight cycle; v[ n ] and a[ n ] are respectively the first n The velocity and acceleration of a time-slot fixed-wing UAV; V min and V max These are the minimum and maximum speeds of the fixed-wing UAV, respectively. a max This represents the maximum acceleration of a fixed-wing unmanned aerial vehicle (UAV). and The first n Within each time slot, the airborne base station pairs and The transmission power; P max This is the maximum transmit power of the airborne base station; The specific steps for solving the optimization problem are as follows: A user pairing algorithm is constructed based on the K-Means clustering algorithm and the Gale-Shapley algorithm; each variable is optimized individually, decomposing the optimization problem into two sub-problems; these two sub-problems are then transformed into convex optimization problems, and solved alternately to obtain at least one local optimum of the optimization problem. The first sub-problem: Optimize the drone trajectory q while keeping the drone's transmit power constant; Given user pairings and transmit power, the optimization problem P1 approximates as the following convex problem: (32a) (32b) in, μ It is a constant, and the global optimal value is obtained through iterative updates; The second sub-problem: Optimize the drone's transmit power P while keeping the drone's trajectory fixed; Given the user pairings and drone trajectories, the optimization problem approximates the following convex problem: (38a) (38b) (38c)。 2. The method for UAV green three-dimensional coverage based on non-orthogonal multiple access as described in claim 1, Its features are, Includes the following steps: The first step is to build a system model: Step 1.1: In a non-orthogonal multiple access (MOA) UAV communication network, a fixed-wing UAV flying at a fixed altitude acts as an aerial base station to provide wireless coverage to ground users. The user altitudes are randomly distributed and represented using a three-dimensional Cartesian coordinate system; any user within the cell... u k , The position is represented as w k = [ x k ,y k ,z k ] T ∈ Indicates the number of ground users; x ( t ) and y( t ) are respectively t The two-dimensional coordinates of the aerial base station projected onto the ground at any given time; the flight trajectory of the aerial base station is represented as: (1) At the end of each service cycle, the aerial base station needs to return to its initial position, where q represents the drone's trajectory, which must satisfy: (2) And q needs to satisfy the following constraints: (3) Where, v[ n ] and a[ n ] are respectively the first n The velocity and acceleration of a time-slot fixed-wing UAV; In the n During each time slot, the user u k Distance to aerial base station d k [ n The following is represented: (6) Step 1.2: Use NOMA technology to increase the number of network accesses; Each frequency band is shared by two users in a group, and each group uses orthogonal frequency division multiplexing; within each time slot, the user with the higher channel gain in each group is considered a strong user, and the user with the lower channel gain is considered a weak user; m Strong users and weak users in a group are denoted as follows: and , , Indicates the number of user groups, and the two are in the th order. The channel gains for each time slot are denoted as follows: and ; will the first n Within each time slot, the airborne base station pairs and The transmission power is denoted as follows: and If both users cancel normal decoding through continuous interference, the following constraints must be met: (9) (10) in, P max This is the maximum transmit power of the airborne base station; Then all users in one flight cycle T Effective throughput within Represented as: (15) in, and In the n The instantaneous rate of each time slot is and ; The energy consumption of an airborne base station includes transmission power consumption. P max Fixed energy consumption of airborne base stations P Base and the propulsion energy consumption of fixed-wing UAVs P Pro [ n The energy consumption model for UAV propulsion is as follows: (16) in, c 1 and c 2 are constants related to parameters of fixed-wing UAVs and air density, respectively. g It is gravitational acceleration; therefore, the airborne base station in one flight cycle T The energy consumption within is: (17) The second step is to simplify the objective function and list the optimization problem: Considering both throughput and energy consumption, the energy efficiency of an airborne base station within one flight cycle T is determined, defined as the ratio of throughput to energy consumption, expressed in bits per liter (J). The formula is as follows: (18) The goal is to first determine the pairings of each user in NOMA, and then optimize the trajectory of the over-the-air base station. Transmission power of the air-based access point To maximize the system's energy efficiency; once user pairings are determined, the optimization problem is identified; The third step is to solve the optimization problem: Step 3.1, User Pairing Algorithm; The K-Means clustering algorithm is used to divide the ground users into two user clusters. Several users are selected from the user cluster with more users and assigned to the other user cluster, so that the two user clusters contain an equal number of users. The Gale-Shapley algorithm is used to pair the users in the two user clusters one by one, so that each user group contains two users. Step 3.2: Fix the UAV's transmit power and optimize the UAV trajectory q; solve using optimization tools; Step 3.3: Fix the drone trajectory and optimize the drone's transmission power P; Solve using optimization tools; Step 3.4 maximizes the energy efficiency of the wireless communication network by combining user matching, drone trajectory and transmission power; and achieves green wireless network coverage under the condition of three-dimensional random distribution of users.

3. The method for UAV green three-dimensional coverage based on non-orthogonal multiple access according to claim 2, characterized in that, In the first step of formula (3), due to the characteristics of fixed-wing UAVs, v[ n ] and a[ n It also needs to meet the following requirements: (4) (5) in, V min and V max These are the minimum and maximum speeds of the fixed-wing UAV, respectively. a max This represents the maximum acceleration of the fixed-wing unmanned aerial vehicle.

4. The method for UAV green three-dimensional coverage based on non-orthogonal multiple access according to claim 2, characterized in that, The user pairing algorithm in step 3.1 of the first step is as follows: Step 3.1.1, Input ; Step 3.1.2: Randomly select two coordinates from the ground user coordinates as initial cluster centers: { μ 1, μ 2}; Step 3.1.3, Initialize the user cluster: ; Step 3.1.4, repeat the following operations: calculate the distance between each user and each cluster; assign users to the nearest cluster; Step 3.1.5: Recalculate the cluster centers based on the users contained in each cluster until the cluster centers remain unchanged; if the number of users contained in two clusters is different, select several users farthest from the cluster center from the cluster with more users and assign them to the other cluster, so that the number of users contained in the two clusters is relatively equal. Step 3.1.6, select a cluster and repeat the following steps: Let the unpaired users in the cluster be denoted as u x To include another cluster that was not included u x The user who has been paired with the user and is furthest away from them is denoted as . u y ; if u y Not yet paired; accept pairing request. Or its preference for the elements requested for pairing is higher than u x If the existing pairing is cleared, the pairing request will be cancelled and the pairing request accepted; otherwise, the pairing request will be rejected. Step 3.1.7: All users have completed pairing, and the user pairings are output.

Images