Method and system for minimizing energy consumption of unmanned aerial vehicle assisted multi-cluster NOMA network

Abstract

The invention relates to the technical field of communication, and particularly discloses an energy consumption minimization method and system for an unmanned aerial vehicle assisted multi-cluster NOMA network, which constructs the unmanned aerial vehicle assisted multi-cluster NOMA network, and maximizes the total energy consumption of an unmanned aerial vehicle as a target and meets the constraint of the unmanned aerial vehicle assisted multi-cluster NOMA network as a condition. A first optimization problem for jointly optimizing the hovering point access sequence, the transmission power distribution, the transmission time length distribution and the total time length of the unmanned aerial vehicle is constructed and solved, and the hovering point access sequence, the transmission power distribution, the transmission time length distribution and the total time length which enable the total energy consumption of the unmanned aerial vehicle to be lowest are obtained. A simulation result verifies that the method is remarkably superior to a traditional reference scheme in the aspects of energy saving and system efficiency improvement, and meanwhile, the QoS requirements of all ground users can be effectively met.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a method and system for minimizing energy consumption of a drone-assisted multi-cluster NOMA network. Background Art

[0002] In recent years, drones have been increasingly used in low-altitude areas, covering a variety of tasks such as aerial inspections, video transmission, photography, and package delivery. With their excellent maneuverability and flexibility, drones play an important role in enhancing ground network capabilities, especially in scenarios where traditional infrastructure is missing or damaged, such as natural disasters or communication needs in remote areas. In addition, the high-quality Line-of-Sight (LoS) communication channel between drones and ground users significantly improves transmission efficiency and reliability, making drones show great potential in a wide range of application scenarios such as emergency response, real-time data collection, and monitoring.

[0003] The rapid development of mobile Internet and the widespread popularity of smart devices have driven the surge in mobile multimedia services, resulting in an exponential growth in mobile communication traffic. In addition, the rise of the Internet of Things has led to the deployment of a large number of IoT devices, all of which require wireless connections. This growing demand has put forward higher performance requirements for wireless communication systems to cope with the increasing traffic and user connection needs. Non-orthogonal multiple access has emerged as a key technology that can provide high-speed data transmission, low latency, high reliability, and large-scale connectivity. Non-orthogonal multiple access (NOMA) enables multiple terrestrial users to share communication resources simultaneously by decoding overlapping signals using serial interference cancellation, significantly improving spectrum efficiency and system throughput, and achieving seamless integration with the fifth generation mobile communication technology (5G). A large number of studies have shown that NOMA has great potential in improving spectrum efficiency, network throughput, and overall performance. For example, passive beamforming technology combined with reconfigurable intelligent surface (RIS) in multiple input multiple output (MIMO) NOMA networks effectively optimizes signal quality and reduces interference. In addition, the millimeter wave NOMA framework improves network security and performance through strategic user grouping and power allocation. In a dual-hop mobile edge computing system, the NOMA-based resource allocation scheme achieves a balance between latency and energy consumption.

[0004] Combining NOMA technology with drone-assisted communications represents a significant step forward in improving system performance. This combination has attracted widespread attention in the academic community, and multiple studies have demonstrated its significant advantages in improving the efficiency and reliability of communication systems. For example, studies have shown that NOMA technology can significantly enhance network security in drone-supported wireless networks, and an analytical framework for Coordinated Multipoint (CoMP) NOMA integrated drone cellular networks has demonstrated further improvements in system performance. In addition, hybrid aerial networks built by combining reconfigurable intelligent surfaces (RIS) with NOMA have also achieved optimization in key network parameters. It should be noted that the efficiency of NOMA can be more fully utilized when the number of users in each cluster is limited; however, when drones need to provide services to a large number of users, a single NOMA group may not be able to meet the needs. At this time, it is often necessary to divide users into multiple clusters and apply NOMA technology in each cluster separately. To address this problem, one paper proposed a method to optimize user clustering, routing, transmission power, hovering position and time to improve network performance, while another study focused on improving the quality of service (QoS) through similar optimization measures.

[0005] At the same time, a notable shortcoming in most existing studies is the neglect of energy consumption in UAV networks. Effective energy management is critical to operational efficiency, especially considering the physical limitations of UAVs. For example, rotary-wing UAVs typically have a flight time of only about 30 minutes, which requires frequent battery replacement or charging, highlighting the need for optimized energy management strategies. Some literature is dedicated to minimizing the total energy consumption in UAV-based patrol detection scenarios by optimizing task completion time, communication scheduling, computing resource allocation, and UAV trajectory. Similarly, some literature develops a framework to optimize the process of collecting data from IoT devices by using energy-constrained UAVs to maximize the amount of data while effectively managing energy usage.

[0006] In a multi-cluster environment, energy management in UAV-NOMA networks faces many complex challenges. One of the key issues is how to plan the optimal order for UAVs to visit each cluster to minimize flight time and energy consumption. This requires designing an efficient flight trajectory. However, this optimization faces a significant trade-off: the flight time can be shortened by increasing the flight speed of the UAV, thereby buying more time for data transmission and hovering within the cluster, but this will also lead to a significant increase in flight energy consumption. Therefore, striking a balance between reducing flight time and reducing energy consumption has become an important and understudied topic in UAV-assisted multi-cluster NOMA networks. Summary of the invention

[0007] The present invention provides a method and system for minimizing energy consumption of a drone-assisted multi-cluster NOMA network, and solves the technical problem of how to jointly optimize the hovering point access sequence, transmission power allocation and transmission duration of the drone to minimize the total energy consumption of the drone.

[0008] In order to solve the above technical problems, the present invention provides a method for minimizing energy consumption of a drone-assisted multi-cluster NOMA network, comprising the steps of:

[0009] Construct a drone-assisted multi-cluster NOMA network, where NOMA refers to non-orthogonal multiple access; the drone-assisted multi-cluster NOMA network includes drones, K ground users divided into M clusters, and the drone's flight altitude is fixed at H. The drone starts from a starting position q I Departure, along the hovering point visit sequence of the drone, arrive at the hovering point of each cluster in turn to provide NOMA service for the cluster, and finally return to the specified end point q F ;

[0010] With the goal of maximizing the total energy consumption of the UAV and satisfying the constraints of the UAV-assisted multi-cluster NOMA network, a joint optimization of the UAV's hovering point access sequence, transmission power allocation P, transmission time allocation T and total time T is constructed. 0 The first optimization problem;

[0011] Solve the first optimization problem and obtain the hovering point access sequence, transmission power allocation P, transmission time allocation T and total time T of the drone. 0 .

[0012] Furthermore, the total energy consumption of the drone includes the hovering energy consumption E hov , UAV transmission energy consumption E tr and the propulsion energy consumption E of the UAV f .

[0013] Furthermore, the constraints of the drone-assisted multi-cluster NOMA network include the first to seventh constraints. The first constraint is that the drone's flight hovering point set must include the centroid coordinates of each cluster. The second constraint is that the drone's flight speed is not less than the maximum speed V max , the third constraint is the starting and ending positions of the given drone, the fourth constraint is that the throughput of each ground user is not less than the defined throughput threshold, the fifth constraint is that the decoding SINR of each ground user at least reaches the defined decoding SINR threshold, SINR refers to the signal-to-noise ratio, the sixth constraint is that the order of power allocation is consistent with the order of distance between the drone and the ground user, and the seventh constraint is that the total transmission power of all users in each cluster does not exceed the power upper limit P of the drone max .

[0014] Furthermore, the first optimization problem is solved to obtain the hovering point access sequence, transmission power allocation P, transmission time allocation T and total time T of the drone. 0 , specifically including the steps:

[0015] To minimize the flight propulsion energy consumption of the UAV f The goal is to optimize the flight speed and hovering point visit sequence of the UAV, that is, to optimize the flight time and flight trajectory of the UAV;

[0016] Based on the determined visit order of hovering points, the transmission energy consumption E of the UAV is minimized. tr and hovering energy consumption E hov The sum of the two is taken as the goal, and the transmission power allocation P and transmission time allocation T of the UAV are optimized. Based on the flight time of the UAV and the transmission time allocation T, the total time T of the UAV is obtained. 0 Furthermore, to minimize the UAV's flight propulsion energy consumption E f To achieve this goal, we optimized the flight speed and hovering point access sequence of the drone, including the following steps:

[0017] To minimize the flight propulsion energy consumption of the UAV f As the goal, taking the first to third constraints as constraints, the first optimization problem is transformed into a second optimization problem of optimizing the flight speed of the UAV and the order of visiting hovering points;

[0018] Based on the speed range of the drone [0,V max ], the optimal flight speed V of the UAV is determined by the golden section search method * , and determine that the optimal flight trajectory Q consists of line segments connecting the hovering points;

[0019] Rearrange the order of visiting the M hovering points to obtain an ordered sequence of waypoints make Define the access order π as (π 1 ,…,π M ), the flight energy consumption E f Expressed as express The propulsion energy consumption between |||| represents the Euclidean distance, and the second optimization problem is transformed into minimizing To meet the goal The third optimization problem is to optimize the visit sequence π of the hovering points under the constraint condition;

[0020] The traveling salesman problem algorithm is improved by constructing a weighted graph to represent the flight path of the drone and introducing a virtual node q′ whose distance to the starting and ending points is zero but whose distance to all other points is infinite.

[0021] The third optimization problem is solved using an improved traveling salesman problem algorithm.

[0022] Furthermore, the constructed weighted graph The vertex set Edge Set (μ i ,μ j ) represents any two different vertices μ i and μ j The edge between them, ω is the weight function, μ i ,μ j The weight between

[0023] The improved traveling salesman problem algorithm includes the following steps:

[0024] Constructing a weighted graph

[0025] Add a virtual node q′ to

[0026] Set the edge weight to: ω(μ 0 ,q′)=0,ω(μ M+1 ,q′)=0,ω(μ m ,q′)=∞,1≤m≤M;

[0027] Starting from the virtual node q′, execute the standard traveling salesman problem algorithm to determine the initial visit sequence π′;

[0028] Remove q′ and its related edges from π′; determine the final access order as

[0029] Furthermore, based on the determined visit order of the hovering points, the transmission energy consumption E of the UAV is minimized. tr and hovering energy consumption E hov The sum of the two is taken as the goal, and the transmission power allocation P and transmission time allocation T of the UAV are optimized, which specifically includes the following steps:

[0030] Based on the determined flight speed of the UAV and the order of hovering point visits, the transmission energy consumption E of the UAV is minimized. tr and hovering energy consumption E hovThe sum of is taken as the goal, and the fourth to seventh constraints are taken as constraints, and the first optimization problem is transformed into the fourth optimization problem of optimizing the transmission power allocation P and the transmission time allocation T of the UAV;

[0031] Decomposing the fourth optimization problem into a fifth optimization problem of optimizing transmission duration allocation T and a sixth optimization problem of optimizing transmission power allocation P;

[0032] The fifth optimization problem and the sixth optimization problem are solved by alternate iteration to obtain the transmission power allocation P and the transmission time allocation T of the UAV.

[0033] Furthermore, the optimization goal of the fifth optimization problem is to minimize the transmission energy consumption E of the UAV. tr and hovering energy consumption E hov The sum of , the constraint condition is the fourth constraint condition;

[0034] The solution to the fifth optimization problem is: τ m represents the transmission duration allocated to the mth cluster, v m,k represents the kth ground user s in the mth cluster m,k The throughput threshold, R m,k Indicates m,k The downlink achievable rate is min, which indicates the minimum value.

[0035] Furthermore, the optimization goal of the sixth optimization problem is to minimize the transmission energy consumption E of the UAV. tr and hovering energy consumption E hov The sum of , the constraints are the fourth constraint to the seventh constraint;

[0036] Solving the sixth optimization problem comprises the steps of:

[0037] Replace R in the fourth constraint m,k Transformed into R m,k is a non-concave function with respect to P, is a non-convex function;

[0038] Will At a given local point Perform a first-order Taylor expansion at

[0039] Rewrite the fourth constraint as: When k = 1, R m,1 Satisfy R m,1 τ m ≥v m,1 ;

[0040] Rewrite the fifth constraint as a linear constraint;

[0041] Based on the sixth constraint condition, the seventh constraint condition and the rewritten fourth constraint condition and the fifth constraint condition, the sixth optimization problem is transformed into a seventh optimization problem;

[0042] The seventh optimization problem is solved using a convex optimization problem solver to obtain the transmission power allocation P of the UAV.

[0043] The present invention also provides an energy consumption minimization system for a drone-assisted multi-cluster NOMA network. The system applies the energy consumption minimization method for a drone-assisted multi-cluster NOMA network. The key lies in that the system includes a network construction module, a problem construction module and a problem solving module; the network construction module is used to construct a drone-assisted multi-cluster NOMA network, the problem construction module is used to construct a first optimization problem, and the problem solving module is used to solve the first optimization problem.

[0044] The energy consumption minimization method and system of the drone-assisted multi-cluster NOMA network provided by the present invention constructs a drone-assisted multi-cluster NOMA network, and takes maximizing the total energy consumption of the drone as the goal, and takes satisfying the constraints of the drone-assisted multi-cluster NOMA network as the condition, constructs a first optimization problem of jointly optimizing the hovering point access sequence, transmission power allocation, transmission time allocation and total time of the drone, and solves the first optimization problem, and obtains the hovering point access sequence, transmission power allocation, transmission time allocation and total time that minimize the total energy consumption of the drone. The simulation results verify that the present invention is significantly superior to the traditional benchmark solution in terms of energy saving and system efficiency improvement, and can effectively meet the QoS requirements of all ground users. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a flow chart of a method for minimizing energy consumption of a drone-assisted multi-cluster NOMA network provided by an embodiment of the present invention;

[0046] Figure 2 It is a UAV flight trajectory diagram optimized by an improved TSP algorithm provided by an embodiment of the present invention;

[0047] Figure 3 is a convergence process diagram of the golden section algorithm provided by an embodiment of the present invention;

[0048] Figure 4 is a convergence process diagram of Algorithm 3 provided in an embodiment of the present invention;

[0049] Figure 5 is a process diagram of optimizing the flight speed by using the golden section search method provided by an embodiment of the present invention;

[0050] Figure 6 is a graph showing the relationship between UAV propulsion energy consumption and flight time provided by an embodiment of the present invention;

[0051] Figure 7 is a graph showing the relationship between the flight speed and flight time of a UAV provided by an embodiment of the present invention;

[0052] Figure 8 is a comparison diagram of UAV propulsion energy consumption under different schemes provided by an embodiment of the present invention;

[0053] Fig. 9 is a diagram of optimized power allocation of each user equipment in different clusters provided by an embodiment of the present invention;

[0054] Fig.10 is the transmission and hovering energy consumption and throughput threshold v provided by the embodiment of the present invention 0 Relationship diagram;

[0055] Fig.11 4 is a graph showing the relationship between transmission and hovering energy consumption and the number of user devices provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The following specifically illustrates the implementation mode of the present invention in conjunction with the accompanying drawings. The embodiments are provided for illustrative purposes only and cannot be understood as limiting the present invention. The accompanying drawings are provided for reference and illustration only and do not constitute a limitation on the scope of patent protection of the present invention, because many changes can be made to the present invention without departing from the spirit and scope of the present invention.

[0057] The energy consumption minimization method of the drone-assisted multi-cluster NOMA network provided in the embodiment of the present invention is as follows: Figure 1 As shown in the flowchart, in this embodiment, the steps include:

[0058] Build a drone-assisted multi-cluster NOMA network;

[0059] With the goal of maximizing the total energy consumption of the UAV and satisfying the constraints of the UAV-assisted multi-cluster NOMA network, a joint optimization of the UAV's hovering point access sequence, transmission power allocation P, transmission time allocation T and total time T is constructed. 0 ;

[0060] Solve the first optimization problem and obtain the hovering point access sequence, transmission power allocation P, transmission time allocation T and total time T of the drone. 0 .

[0061] (1) Drone-assisted multi-cluster NOMA network

[0062] 11) Network Model

[0063] In the network model of UAV-assisted multi-cluster NOMA network (UAV-NOMA network), UAVs act as mobile base stations to provide services for K ground users, which are divided into M clusters. To achieve efficient user clustering, this paper uses the K-means algorithm with fast convergence characteristics to divide K ground users into M clusters. The cluster set is defined as C = {C 1 ,C 2 ,…,C M}. In the mth cluster C m There are N m A ground user, using Ω m Indicates that the total number of users in all clusters is K, satisfying The drone starts from the starting position q I Start, follow the trajectory to reach the hovering point of each cluster in turn, and finally return to the specified end position q F Specifically, the hovering point of each cluster m is set to its cluster center (excluding the horizontal position at altitude). During this process, the drone transmits only when hovering to reduce the impact of the Doppler effect. To meet the needs of large-scale connections and optimize spectrum efficiency, the drone uses NOMA technology to provide services to users in each cluster. In addition, the total duration T 0 Including transmission time T tr and flight time T f , that is, T 0 =T f +T tr The transmission duration allocated to the mth cluster is τ m ,satisfy Assume m,k represents the kth user in the mth cluster, which is close to the hovering point μ m The distance between them is denoted as d m,k The horizontal position of the drone at time t is Therefore, there is That is, the hovering point μ m Included in the drone at 0 to T 0 In each horizontal position in between.

[0064] Assume that the drone is flying at an altitude of H and user s m,k The location is Then d m,k Calculated as: |||| indicates the Euclidean distance. It is usually assumed that the mth C m In the example, the distance between each user and the drone satisfies

[0065] 12) Communication model

[0066] The air-to-ground communication channel is usually assumed to be a LoS channel. m,k The channel gain between the UAV and the where β 0 is the reference distance D 0 = average channel gain at 1 meter. In the NOMA system, the receiver uses serial interference cancellation technology to ensure fairness among ground users. Specifically, users with weaker channel conditions (weak users) are allocated more transmission power, while users with better channel quality (strong users) are allocated less power. Let p m,k Indicates user s m,k According to the distance from the ground user to the UAV, p m,k Need to meet In addition, the total transmission power of all ground users in each cluster must not exceed the power limit P of the UAV. max ,Right now In the NOMA system, each ground user first decodes the signal with a stronger channel and eliminates it from the received combined signal before decoding its own signal. m , users m,k The signal-to-noise ratio expression is:

[0067]

[0068] where σ 2 Represents the noise power at the receiving end.

[0069] When k=1, its SINR is:

[0070]

[0071] In addition, users with better channel conditions also need to accurately decode messages from users with weaker channels, so the constraint SINR m,k ≥η m,k , where η m,k Indicates user s m,k Therefore, user s m,k The downlink achievable rate is expressed as:

[0072] R m,k =B log 2 (1+SINR m,k ), (3)

[0073] Where B is the channel bandwidth.

[0074] 13) Energy consumption model

[0075] This paper divides the energy consumption of drones into two parts: flight propulsion energy consumption and transmission energy consumption, and analyzes them separately. The energy consumption of drones during flight is recorded as E f , and the energy consumption during transmission is recorded as E tr According to existing literature, the propulsion power consumption of a rotary-wing UAV in horizontal flight can be expressed as:

[0076]

[0077] in, Indicates the horizontal flight speed of the drone, V 0 and U tip They represent the average rotor induced speed and rotor tip speed of the UAV respectively. i and P 0 They represent the hovering induced power and rotor profile power of the UAV respectively. In addition, d 0 , s, A and ρ represent the drag coefficient of the drone, the solidity ratio of the rotor, the area of ​​the rotor disc and the air density respectively. Therefore, the propulsion energy consumption E of the drone is f It can be expressed as:

[0078]

[0079] On the other hand, the total transmission energy consumption of the UAV can be expressed as:

[0080]

[0081] In addition, the energy consumption of the drone during hovering is recorded as E hov , it is defined that when V(t) = 0, the drone is in hovering state. The energy consumption during hovering is calculated as:

[0082] E hov =(P 0 +P i )τ m , (7)

[0083] Finally, the total energy consumption of the system is defined as:

[0084] E total =E tr +E hov +E f . (8)

[0085] (2) Constructing the optimization problem

[0086] 21) Objective Function

[0087] In order to maximize resource efficiency, the optimization goal established in this method is to jointly optimize the UAV trajectory Q = {q(t)|t∈[0,T f]}, transmission power P = {p m,k |k∈[1,N m ], m∈[1,M]} and the transmission time allocation T={τ m |m∈[1,M]} and the total duration T 0 To minimize (min) the total energy consumption of the system. Therefore, the established objective function is: Total time T 0 = Transmission time T tr +Flight durationT f , transmission time T tr (i.e., the transmission time allocation T) has been used as the optimization variable, and the flight time T f Optimizing is equivalent to optimizing the flight speed.

[0088] 22) Constraints

[0089] According to the network model and communication model of the drone-assisted multi-cluster NOMA network, the constraints of the drone-assisted multi-cluster NOMA network include:

[0090]

[0091] q(0)=q I ,q(T 0 )=q F , (11)

[0092]

[0093] Among them, equations (9) to (15) are the first to seventh constraints respectively. Constraint (9) stipulates that the set of flying hovering points of the UAV must include the centroid coordinates of each cluster. Constraint (10) stipulates that the flying speed of the UAV is not less than the maximum speed V max Constraint (11) defines the starting and ending locations of the UAV. Constraint (12) ensures that the throughput of each ground user is not less than the defined threshold v m,k Constraint (13) ensures that the decoded SINR of each terrestrial user at least reaches the defined threshold η m,k Constraint (14) limits the order of power allocation to be consistent with the order of distance between the UAV and the ground users. Constraint (15) limits the total transmission power of all users in each cluster to not exceed the UAV's power upper limit P max .

[0094] 23) Optimization issues

[0095] By combining the objective function and the constraints, the first optimization problem (P1) is obtained.

[0096] Due to the high coupling of variables in the nonconvex constraints (12) and (13), and the complexity of the selectivity constraint (9), directly solving the nonconvex optimization problem (P1) is extremely challenging.

[0097] (3) Solving optimization problems

[0098] The present invention solves the first optimization problem through a two-stage method, the goal of which is to comprehensively consider the propulsion energy consumption and transmission energy consumption and minimize the total energy consumption of the UAV. In the first stage, the flight speed and hovering point access order of the UAV are optimized to reduce the flight propulsion energy consumption, and it is proved that the optimal flight trajectory Q consists of line segments connecting the hovering points. In the second stage, based on the determined hovering point access order (i.e., the flight path or flight trajectory), the transmission power allocation P and transmission time allocation T of the UAV are optimized to minimize the transmission energy consumption and hovering energy consumption of the UAV. That is, solving the first optimization problem, the hovering point access order, transmission power allocation P, transmission time allocation T and total time T of the UAV are obtained. 0 , specifically including the steps:

[0099] To minimize the flight propulsion energy consumption of the UAV f The goal is to optimize the flight speed and hovering point visit sequence of the UAV, that is, to optimize the flight time and flight trajectory of the UAV;

[0100] Based on the determined visit order of hovering points, the transmission energy consumption E of the UAV is minimized. tr and hovering energy consumption E hov The sum of the two is taken as the goal, and the transmission power allocation P and transmission time allocation T of the UAV are optimized. Based on the flight time of the UAV and the transmission time allocation T, the total time T of the UAV is obtained. 0 .

[0101] 31) Phase 1: Optimization of drone flight speed and hovering point access sequence

[0102] In the first stage, the goal is to minimize the flight propulsion energy consumption of the UAV. The first optimization problem is transformed into the second optimization problem:

[0103] (P2):

[0104] st (9),(10),(11)

[0105] To solve problem (P2), the present invention adopts a two-step optimization strategy. First, the optimal flight speed of the UAV is determined by the golden section search method, and it is strictly proved that the optimal flight trajectory consists of several line segments. Second, the visiting order of the UAV hovering points is optimized using the traveling salesman problem algorithm to further reduce the flight energy consumption.

[0106] ①Determination of optimal flight speed

[0107] The present invention first considers designing the minimum energy flight trajectory between any two points, for example, from μi to μj. The path discretization method is used to divide the trajectory of the drone into L segments, thereby obtaining L+1 waypoints. The position information of the drone at the starting point of segment l is recorded as The flight time of this segment is denoted as δ l . Assume that for any l, ||q l+1 -q l ||≤d max , where d max is chosen to be small enough so that the flight speed of the drone can be approximately considered constant within each segment. max , L is chosen to be large enough so that in From μ i to μ j The upper bound of the total flight distance required. Therefore, the flight speed of the UAV in the lth segment can be expressed as Total flight time for in Therefore, the propulsion energy consumption of the first segment can be approximately expressed as:

[0108]

[0109] in, Therefore, the total propulsion energy consumption of L segments is

[0110] It should be noted that, given q I and q l+1 Under the conditions, depending on Where V l is the only variable, and its value range is in the interval [0,V max ]. Therefore, this section uses the golden section search method to obtain The minimum value of .

[0111] The golden section search method achieves its goal by gradually narrowing the search interval until it is small enough. max ] on the unimodal function The search interval is updated by comparing x 1 and x 2 This is done by looking at the function value at . During each iteration, the updated bounds of the interval can be expressed as follows:

[0112]

[0113] The specific search process of the golden section search method is shown in Table 1 below:

[0114] Table 1

[0115]

[0116] Through the golden section search method, it can be concluded that for any straight line segment l, its optimal flight speed V* is the same in all segments, that is, This uniform optimal speed is independent of the starting and ending positions of the road segment. The time complexity of the golden section search method depends on the required accuracy εg, not the size of the initial interval, so the time complexity is By adopting the golden section search method, the optimal speed of each UAV flight segment can be efficiently determined, thereby minimizing energy consumption. Next, we will continue with Proposition 1 to further determine the optimal flight trajectory of the UAV.

[0117] Proposition 1: To minimize flight propulsion energy consumption, the UAV should fly in a straight line between the specified starting point μi and the end point μj.

[0118] Proof: This proposition is proved by contradiction. Assume that the optimal flight trajectory designed to minimize propulsion energy consumption contains a curved path between points μi and μj. Then, it is always possible to connect the waypoints ql and ql +1 Any curved path between is replaced by a straight path and flies along this path at the optimal speed V*, thereby obtaining an alternative trajectory with lower energy consumption. in is a constant. Minimize The premise is to set ||q l+1 -q l || drops to the lowest value. Obviously, the shortest path between two points is a straight line. Therefore, the drone should take the path connecting q l and q l+1 The straight line segment, with the optimal speed V * Flying, thus completing the proof.

[0119] Based on Proposition 1, the most energy-efficient flying mode for a drone is to fly at the starting point μ i and the end position μ j Keep constant speed V along the straight path * The method shows that energy consumption is closely related to the flight distance, thus highlighting the importance of the shortest flight trajectory. To achieve this goal, this paper adopts a modified traveling salesman problem algorithm to determine the optimal order for the drone to visit the hovering points.

[0120] ②Access order optimization

[0121] By rearranging the order of visiting the M hovering points, an ordered sequence of waypoints can be obtained. Where (π 1 ,…,π M ) is a permutation of (1,…,M). Let The access order π is defined as (π 1 ,…,π M ), flight energy consumption E f It is expressed as: Therefore, the second optimization problem (P2) can be reformulated as the third optimization problem:

[0122] (P3):

[0123]

[0124] To clarify the problem (P3), we construct a weighted graph To optimize the flight path of the drone, represents the vertex set, E represents the edge set, and ω is the weighting function. For any two different vertices μ i and μ j ,in There exists an edge (μ i ,μ j ). Define the weight function ω→R + (positive real numbers), in particular,

[0125] By constructing a weighted graph G, it is found that problem (P3) is a traveling salesman problem based on graph G, which focuses on reaching a specific target point without returning to the starting point. To efficiently solve problem (P3), this paper proposes an improved traveling salesman problem algorithm, which introduces a virtual node q′, which is configured to have a zero distance from the starting and ending locations, but an infinite distance from all other points.

[0126] When solving the problem using the improved Traveling Salesman Problem (TSP) algorithm, first, according to the traditional strategy of the TSP algorithm, the UAV starts from the virtual node q′, traverses all points, and finally returns to the node q′. I and q F The distance from q′ to q is zero, and minimizing the total flight distance of the UAV requires choosing I and from q F Finally, by removing the virtual nodes and their associated edges, the target visit order of the UAV is obtained.

[0127] The detailed process of the improved traveling salesman problem algorithm is summarized in Algorithm 2 as shown in Table 2 below.

[0128] Table 2

[0129]

[0130] 32) Phase 2: Optimization of UAV transmission power and transmission duration allocation

[0131] After optimizing the UAV's flight propulsion energy consumption, the optimal flight speed and access order can be obtained. Based on the results of this optimization phase, the next step is to minimize the energy consumption of communication and hovering. The fourth optimization problem is stated as follows:

[0132] (P4):

[0133] st(12)-(15).

[0134] Problem (P4) is non-convex due to the coupling between P and T, making it challenging to solve directly. To simplify problem (P4), it is divided into two sub-problems, namely, transmission duration optimization and transmission power optimization. In the transmission duration optimization, a closed-form solution can be derived. For the non-convex power optimization problem, it is approximated as a convex problem by applying the first-order Taylor approximation. Subsequently, this embodiment proposes an algorithm to effectively handle these two sub-problems through iterative solution.

[0135] ①Transmission duration optimization

[0136] Given the transmission power P, the problem (P4) can be decomposed into the fifth optimization problem:

[0137] (P5):

[0138] st (12).

[0139] To solve this problem, we will obtain a closed-form solution through Proposition 2 and give the proof process.

[0140] Proposition 2: For drones, shorter transmission time means lower energy consumption. Therefore, under the condition of satisfying the constraints, the solution of problem (P5) can be expressed in the form of a closed-form solution:

[0141]

[0142] Proof: Since the objective function is τ m is an increasing function of , so minimize τ m It helps to reduce the objective function value. In addition, considering the constraint (12), τm The optimal value of is the minimum value that satisfies this constraint. This shows that the closed-form expression in equation (19) provides the optimal transmission time T, which is consistent with minimizing τ m In line with the principle of reducing energy consumption.

[0143] ②Transmission power optimization

[0144] Given the transmission time T, problem (P4) can be decomposed into the sixth optimization problem:

[0145] (P6):

[0146] st (12)-(15).

[0147] Due to the existence of constraints (12) and (13), this problem is difficult to solve directly. Therefore, the continuous convex approximation method is used to approximate the non-convex constraints to convex constraints. For constraint (12):

[0148]

[0149] It further expands to:

[0150]

[0151] because The existence of R m,k is a non-concave function with respect to P. To handle non-convex terms For a non-convex function at a given local point Perform a first-order Taylor expansion at and obtain its upper bound:

[0152]

[0153] This will is approximated as a concave function. Therefore, constraint (12) can be approximated as:

[0154]

[0155] In particular, when k = 1, R m,1 Need to meet:

[0156]

[0157] Similarly, constraint (13) is transformed into the following form:

[0158]

[0159] Convert it into a linear constraint:

[0160]

[0161] Therefore, the sixth optimization problem (P6) can be approximated as the seventh optimization problem:

[0162] (P7):

[0163] st(14),(15),(23),(24),(26).

[0164] Problem (P7) is a standard convex optimization problem and can be solved with the help of a convex optimization problem solver (CVX solver).

[0165] ③Overall algorithm and complexity analysis

[0166] Based on the above discussion, in order to solve problem (P4), it is decomposed into two sub-problems: one can be expressed in the form of a closed-form solution, and the other is transformed into a convex optimization problem. Algorithm 3 shown in Table 3 summarizes the detailed process of solving problem (P4) using an iterative algorithm.

[0167] Table 3

[0168]

[0169] The core step of Algorithm 3 is to solve the convex optimization problem, so its complexity is mainly affected by the number of optimization variables. Under the given solution accuracy condition, its complexity can be expressed as

[0170] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this embodiment is not limited here.

[0171] (4) Analysis of simulation results

[0172] 41) Simulation settings

[0173] A large number of simulation experiments are carried out in this embodiment to verify the effectiveness of the proposed algorithm. First, consider a drone-assisted multi-cluster NOMA network, which contains K = 20 user devices, which are uniformly and randomly distributed in an area of ​​1.0 × 1.0 km 2 The relevant parameters are set as follows: B = 1MHz, H = 100m, σ 2 =-110dBm,β 0 =-60dB. Assume that the throughput threshold and decoding signal-to-noise ratio requirements of all user equipment are the same, that is, v m,k =v 0 =3bit / Hz,η m,k =η0 =1.0, The maximum flight speed of the drone is set to V max =30m / s. For the parameters related to the energy consumption of the rotary wing UAV, set d 0 =0.6,U tip =120,A=0.503,ρ=1.225,s=0.05,V 0 =4.03,P 0 =79.8563,P i =88.5279. The initial and final positions of the drone are set to q I =[95,270,H] T and q F =[750,890,H] T Specifically, the drone is I Start flying, pass through each cluster center, and finally return to q F .

[0174] The K-means algorithm is used to divide the randomly distributed K user devices into five different clusters, such as Figure 2 As shown. The center point of each cluster represents the geometric center of all ground user locations within the cluster to ensure fairness within the cluster. Figure 2 The optimized UAV flight trajectory using the improved TSP algorithm is also shown. Figure 2 The effectiveness of Algorithm 2 is verified, and the trajectory that can minimize the flight energy consumption during the flight of the UAV is demonstrated.

[0175] 42) Convergence analysis

[0176] Figure 3 The performance of Algorithm 1 at different numbers of iterations is shown. Figure 3 The results show that the energy consumption decreases significantly with the increase in the number of iterations and stabilizes around the 14th iteration. The curve highlights the effectiveness of the golden section search method in quickly converging to the minimum energy consumption. The peak in energy consumption in the second iteration can be attributed to the inherent mechanism of the method to evaluate extreme points within the search interval to determine the search direction, that is, the method searches for the minimum by alternately evaluating the left and right segments of the interval. During this process, it is likely that the extreme values ​​on the higher energy consumption side were tested, resulting in the observed peak, and this evaluation is a typical phenomenon in the golden section search method. In addition, Figure 4 The convergence of Algorithm 3 is shown. It can be observed that both the outer and inner iterations converge quickly, which verifies the effectiveness of the proposed algorithm.

[0177] 43) UAV propulsion energy consumption performance analysis

[0178] Figure 5 The flight performance of the UAV at a standard flight altitude of 100 meters was analyzed, and the relationship between flight speed and energy consumption was demonstrated. Figure 5 The results show that Algorithm 1 successfully achieved convergence and achieved the lowest flight energy consumption during the flight. Figure 6 and Figure 7 This further reveals the inherent trade-off between drone speed and operating time, as well as its impact on propulsion energy consumption. This example introduces the following two benchmark schemes for comparison: a drone flight time minimization scheme, in which the drone flies at maximum speed to shorten the total operating time; and a drone power minimization scheme, in which the drone flies at maximum endurance speed to reduce propulsion power consumption. Figure 6 The curve of the UAV flight time minimization scheme shows a sharp upward trend, which shows that increasing the flight speed to minimize the running time will significantly increase the propulsion energy consumption. In contrast, the UAV power minimization scheme shows lower power consumption in the initial stage, but due to the extension of the running time, it leads to additional energy consumption. Figure 7 The optimized drone speed changes over time, and different strategies balance energy efficiency and runtime by maintaining or adjusting flight speed. The proposed method can effectively handle this trade-off problem, significantly reducing energy consumption by balancing flight speed and runtime, and its performance is better than the above two benchmark solutions.

[0179] Figure 8 The flight energy consumption performance of the drone under three strategies is compared: Algorithm 2, random access order, and cluster index access order. The bar chart shows the energy consumption comparison of Algorithm 1, the drone flight time minimization scheme, and the power minimization scheme. It is worth noting that the improved TSP method achieves the lowest energy consumption in all scenarios, fully demonstrating its excellent performance in optimizing the drone flight path and improving energy efficiency. Figure 8 The results further show that the structured path planning provided by the improved TSP algorithm significantly enhances the energy saving effect, especially when combined with the golden section search method.

[0180] 44) Analysis of UAV communication and hovering energy consumption performance

[0181] Fig. 9 The optimized power allocation for user devices in different clusters in the UAV-NOMA network is shown. It can be observed that the power allocation for each cluster is different due to the need to adjust the power based on the distance to ensure successful SIC decoding. User devices that are farther away from the UAV are allocated more power to effectively meet the decoding requirements. Fig.10 The three strategies are compared. 0The change trends of transmission energy consumption and hovering energy consumption are shown in Figure 3. The three strategies are: Algorithm 3, equal power allocation scheme, and TDMA (time division multiple access) scheme. Algorithm 3 shows a moderate and stable energy consumption growth; in contrast, the energy consumption growth of the equal power allocation scheme is steeper, indicating that its energy allocation efficiency is low when the throughput demand increases. The energy consumption growth of the TDMA scheme is between the two. Although it is more efficient than the equal power allocation scheme, it is still not as good as the proposed Algorithm 3 scheme. The results show that Algorithm 3 can more effectively manage the energy consumption problem caused by the increase in throughput demand when utilizing the NOMA scheme.

[0182] Fig.11 The communication and hovering energy consumption performance of Algorithm 3, the equal power allocation scheme, and the TDMA scheme are compared and analyzed when the number of user devices increases. The energy consumption of all strategies increases with the increase in the number of user devices. However, the proposed scheme has the lowest increase in energy consumption when the number of user devices increases, showing excellent efficiency. In contrast, the equal power allocation scheme shows a significant linear increase in energy consumption, indicating that uniformly allocating power to each user will significantly improve the overall communication and hovering energy consumption. The TDMA scheme allocates the maximum allowed power within the specified time slot of each user, which leads to a significant increase in its communication and hovering energy consumption as the number of users increases. However, although its energy consumption growth is more controllable than that of the equal power allocation scheme, it is still higher than that of the proposed algorithm. Overall, Algorithm 3 can effectively optimize communication and hovering energy consumption, and its performance is better than that of traditional schemes and strategies.

[0183] Based on the above-mentioned UAV-assisted multi-cluster NOMA network energy consumption minimization method, this embodiment also provides a UAV-assisted multi-cluster NOMA network energy consumption minimization system, which includes a network construction module, a problem construction module and a problem solving module; the network construction module is used to construct a UAV-assisted multi-cluster NOMA network, the problem construction module is used to construct a first optimization problem, and the problem solving module is used to solve the first optimization problem.

[0184] The embodiments described herein may be implemented in a computing system that includes a backend component (e.g., as a data server), or a computing system that includes a middleware component (e.g., an application server), or a computing system that includes a frontend component (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0185] The computer programs for implementing the methods and systems of the present invention may be written in any combination of one or more programming languages ​​and stored in a computer-readable storage medium. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer programs are executed by the processor, the functions / operations specified in the flow charts and / or block diagrams are implemented. The computer programs may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0186] Computer readable storage medium can be a tangible medium that can contain or store a computer program for use by an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. Computer readable storage medium can include but is not limited to electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the above. Alternatively, computer readable storage medium can be a machine readable signal medium. More specific examples of machine readable storage medium can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, compact disk read-only memories (CD ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above.

[0187] In summary, the embodiments of the present invention aim to minimize the overall energy consumption of a multi-cluster UAV-NOMA network by jointly optimizing the flight trajectory, power allocation, and transmission duration of the UAVs. The main contributions of the present invention can be summarized as follows:

[0188] (1) A new multi-cluster UAV-NOMA network framework is proposed to meet the user decoding signal to interference plus noise ratio requirements with limited resources. The framework significantly improves the energy efficiency of the network by optimizing the allocation of UAV flight trajectory, transmission power and transmission duration. The problem is formalized as a non-convex optimization problem with selectivity constraints, which is challenging to solve.

[0189] (2) To address this challenge, the present invention proposes a two-stage algorithm. The first stage combines the improved traveling salesman problem method with the golden section search algorithm to focus on reducing propulsion energy consumption. The second stage effectively solves the non-convex constraints and minimizes transmission energy consumption by using alternating optimization techniques combined with a double-loop iterative algorithm of continuous convex approximation.

[0190] (3) Through comprehensive simulation verification, the present invention shows significant advantages in energy saving compared with the baseline solution. The simulation results further reveal the trade-off between flight time and energy consumption, indicating that optimizing the flight trajectory can effectively achieve higher energy efficiency.

[0191] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. The energy consumption minimization method of UAV-assisted multi-cluster NOMA network is characterized by: Includes steps: Construct a drone-assisted multi-cluster NOMA network, where NOMA refers to non-orthogonal multiple access; the drone-assisted multi-cluster NOMA network includes drones, K ground users divided into M clusters, and the drone's flight altitude is fixed at H. The drone starts from a starting position q I Departure, along the hovering point visit sequence of the drone, arrive at the hovering point of each cluster in turn to provide NOMA service for the cluster, and finally return to the specified end point q F ; With the goal of maximizing the total energy consumption of the UAV and under the condition of satisfying the constraints of the UAV-assisted multi-cluster NOMA network, a first optimization problem of jointly optimizing the hovering point access sequence, transmission power allocation P, transmission time allocation T and total time T0 of the UAV is constructed; Solve the first optimization problem to obtain the hovering point access sequence, transmission power allocation P, transmission time allocation T and total time T0 of the drone.

2. The method for minimizing energy consumption of a drone-assisted multi-cluster NOMA network according to claim 1, characterized in that: The total energy consumption of the UAV includes the hovering energy consumption E hov , UAV transmission energy consumption E tr and the propulsion energy consumption E of the UAV f .

3. The method for minimizing energy consumption of a drone-assisted multi-cluster NOMA network according to claim 2, characterized in that: The constraints of the drone-assisted multi-cluster NOMA network include the first to seventh constraints. The first constraint is that the drone's flight hovering point set must include the centroid coordinates of each cluster. The second constraint is that the drone's flight speed is not less than the maximum speed V. max , the third constraint is the starting and ending positions of the given drone, the fourth constraint is that the throughput of each ground user is not less than the defined throughput threshold, the fifth constraint is that the decoding SINR of each ground user at least reaches the defined decoding SINR threshold, SINR refers to the signal-to-noise ratio, the sixth constraint is that the order of power allocation is consistent with the order of distance between the drone and the ground user, and the seventh constraint is that the total transmission power of all users in each cluster does not exceed the power upper limit P of the drone max .

4. The method for minimizing energy consumption of a drone-assisted multi-cluster NOMA network according to claim 3, characterized in that: Solving the first optimization problem to obtain the hovering point access sequence, transmission power allocation P, transmission time allocation T and total time T0 of the drone specifically includes the following steps: To minimize the flight propulsion energy consumption of the UAV f The goal is to optimize the flight speed and hovering point visit sequence of the UAV, that is, to optimize the flight time and flight trajectory of the UAV; Based on the determined visit order of hovering points, the transmission energy consumption E of the UAV is minimized. tr and hovering energy consumption E hov The sum of the two is taken as the goal, the transmission power allocation P and the transmission time allocation T of the UAV are optimized, and the total flight time T0 of the UAV is obtained based on the flight time of the UAV and the transmission time allocation T.

5. The method for minimizing energy consumption of a drone-assisted multi-cluster NOMA network according to claim 4, characterized in that: To minimize the flight propulsion energy consumption of the UAV f To achieve this goal, we optimized the flight speed and hovering point access sequence of the drone, including the following steps: To minimize the flight propulsion energy consumption of the UAV f As the goal, taking the first to third constraints as constraints, the first optimization problem is transformed into a second optimization problem of optimizing the flight speed of the UAV and the order of visiting hovering points; Based on the speed range of the drone [0,V max ], the optimal flight speed V of the UAV is determined by the golden section search method * , and determine that the optimal flight trajectory Q consists of line segments connecting the hovering points; Rearrange the order of visiting the M hovering points to obtain an ordered sequence of waypoints make Define the access sequence π as (π1,…,π M ), the flight energy consumption E f Expressed as express The propulsion energy consumption between them, ∥∥ represents the Euclidean distance, and the second optimization problem is transformed into minimizing To meet the goal The third optimization problem is to optimize the visit sequence π of the hovering points under the constraint condition; The traveling salesman problem algorithm is improved by constructing a weighted graph to represent the flight path of the drone and introducing a virtual node q′ whose distance to the starting and ending points is zero but whose distance to all other points is infinite. The third optimization problem is solved using an improved traveling salesman problem algorithm.

6. The method for minimizing energy consumption of a drone-assisted multi-cluster NOMA network according to claim 5, characterized in that: Constructed weighted graph The vertex set Edge Set (μ i ,μ j ) represents any two different vertices μ i and μ j The edge between them, ω is the weight function, μ i ,μ j The weight between The improved traveling salesman problem algorithm includes the following steps: Constructing a weighted graph Add a virtual node q′ to Set the edge weights to: ω(μ0,q′)=0, ω(μ M+1 ,q′)=0,ω(μ m ,q′)=∞,1≤m≤M; Starting from the virtual node q′, execute the standard traveling salesman problem algorithm to determine the initial visit sequence π′; Remove q′ and its associated edges from π′; Determine the final access order as 7. The method for minimizing energy consumption of a drone-assisted multi-cluster NOMA network according to claim 6, characterized in that: Based on the determined visit order of hovering points, the transmission energy consumption E of the UAV is minimized. tr and hovering energy consumption E hov The sum of the two is taken as the goal, and the transmission power allocation P and transmission time allocation T of the UAV are optimized, which specifically includes the following steps: Based on the determined flight speed of the UAV and the order of hovering point visits, the transmission energy consumption E of the UAV is minimized. tr and hovering energy consumption E hov The sum of is taken as the goal, and the fourth to seventh constraints are taken as constraints, and the first optimization problem is transformed into the fourth optimization problem of optimizing the transmission power allocation P and the transmission time allocation T of the UAV; Decomposing the fourth optimization problem into a fifth optimization problem of optimizing transmission duration allocation T and a sixth optimization problem of optimizing transmission power allocation P; The fifth optimization problem and the sixth optimization problem are solved by alternate iteration to obtain the transmission power allocation P and the transmission time allocation T of the UAV.

8. The method for minimizing energy consumption of a drone-assisted multi-cluster NOMA network according to claim 7, characterized in that: The optimization goal of the fifth optimization problem is to minimize the transmission energy consumption E of the UAV. tr and hovering energy consumption E hov The sum of , the constraint condition is the fourth constraint condition; The solution to the fifth optimization problem is: τ m represents the transmission duration allocated to the mth cluster, v m,k represents the kth ground user s in the mth cluster m,k The throughput threshold, R m,k Indicates m,k The downlink achievable rate is min, which indicates the minimum value.

9. The method for minimizing energy consumption of a drone-assisted multi-cluster NOMA network according to claim 7, characterized in that: The optimization goal of the sixth optimization problem is to minimize the transmission energy consumption E of the UAV. tr and hovering energy consumption E hov The sum of , the constraints are the fourth constraint to the seventh constraint; Solving the sixth optimization problem comprises the steps of: Replace R in the fourth constraint m,k Transformed into R m,k is a non-concave function with respect to P, is a non-convex function; Will At a given local point Perform a first-order Taylor expansion at Rewrite the fourth constraint as: When k = 1, R m,1 Satisfy R m,1 τ m ≥v m,1 ; Rewrite the fifth constraint as a linear constraint; Based on the sixth constraint condition, the seventh constraint condition and the rewritten fourth constraint condition and the fifth constraint condition, the sixth optimization problem is transformed into a seventh optimization problem; The seventh optimization problem is solved using a convex optimization problem solver to obtain the transmission power allocation P of the UAV.

10. A drone-assisted multi-cluster NOMA network energy consumption minimization system, the system applying the drone-assisted multi-cluster NOMA network energy consumption minimization method according to any one of claims 1 to 9, characterized in that: The system includes a network construction module, a problem construction module and a problem solving module; the network construction module is used to construct a drone-assisted multi-cluster NOMA network, the problem construction module is used to construct a first optimization problem, and the problem solving module is used to solve the first optimization problem.

Images