Multi-unmanned aerial vehicle cooperative trajectory optimization method based on user priority

Abstract

The invention discloses a multi-unmanned aerial vehicle cooperative trajectory optimization method based on user priorities, and relates to the technical field of trajectory optimization when unmanned aerial vehicles perform data acquisition on ground users. Regions are divided according to the maximum coverage radius of unmanned aerial vehicle base stations, the optimal coverage points of the unmanned aerial vehicle base stations in the small regions are found according to the actual distribution condition of users, then it is assumed that a command center is deployed at an original point and there are U unmanned aerial vehicle base stations, the unmanned aerial vehicle base stations take off from the origin of coordinates, and a single unmanned aerial vehicle base station with the maximum coverage radius from the original point is found. And traversing the optimal coverage points of the unmanned aerial vehicle base stations in all the small areas, and returning to the shortest path of the original point. The shortest path is reasonably divided according to the number of the unmanned aerial vehicle base stations, and the trajectory of the unmanned aerial vehicle base stations is optimized on the basis of considering the user priority, the unmanned aerial vehicle energy consumption and the task completion time of the unmanned aerial vehicle base stations. Simulation analysis shows that under the condition that the user priority is considered, the method effectively balances the overall performance of energy consumption and task completion time of the unmanned aerial vehicle.

Description

Technical Field

[0001] The present invention belongs to the technical field of trajectory optimization when a UAV collects data from a ground user, and in particular, relates to a multi-UAV collaborative trajectory optimization method based on user priority. Background Art

[0002] It is not difficult to see from the communication repair sites in recent years that drones, with their advantages of flexibility and strong maneuverability, play a vital role in the restoration and guarantee of communications in various environments. Especially in harsh environments, all fixed base stations in the area have been destroyed, and the number of drone base stations is insufficient to provide full-time communication coverage for all users in the entire area, and ground users and the regional command center are disconnected. In this case, it is possible to consider the method of intermittently collecting ground user data through limited drone base stations to complete the transmission of important data to the regional command center. In particular, it is necessary to consider that for ground users with different tasks, the user priorities are different, and the data of important users should be transmitted back as soon as possible. The trajectory optimization problem under the collaborative conditions of multiple drones based on user priorities needs to comprehensively consider multiple factors such as drone flight altitude, user differentiated communication needs, user priority, signal transmission loss, signal data transmission rate, drone energy consumption, and drone task completion time. However, the current research has not fully considered the differences in user priorities in harsh environments, and has unified users into one category, and only optimized drone trajectories from a single perspective, which does not conform to the actual situation. For ground users with different tasks, the priority of the data to be transmitted is also different. The data with high priority plays a vital role in the decision-making of the regional command center. Therefore, when optimizing the trajectory of the drone base station, it is necessary to consider the user priority and make overall optimization from the perspective of drone energy consumption and drone task completion time.

[0003] In solving trajectory optimization problems, the particle swarm algorithm has the advantages of strong global search capability, easy to optimize problems together with other algorithms, easy to extend to multi-objective optimization, strong robustness, and simple and easy implementation. At the same time, it has the problems of premature convergence and falling into local optimality. Summary of the invention

[0004] The present invention aims to solve the problems raised in the background technology and proposes a multi-UAV collaborative trajectory optimization method based on user priority.

[0005] In order to achieve the purpose of the present invention, the present invention discloses a multi-UAV collaborative trajectory optimization method based on user priority, comprising the following steps:

[0006] Step 1: Divide the area into N small areas according to the maximum coverage radius of the drone base station. In each small area, find the optimal coverage position of the drone base station according to the differentiated communication needs and distribution of the four types of users. The serial number is the same as that of the small area.

[0007] Step 2: In each small area, the total priority of the users, the amount of data to be transmitted, and the required time are calculated according to the distribution of the users;

[0008] Step 3: Use the traveling salesman algorithm to find the shortest path starting from the origin, connecting the best coverage points of N small-area drone base stations, and returning to the origin;

[0009] Step 4: Obtain the total time t required for the shortest path based on the time required for the path and the time required for the amount of data to be transmitted at the coverage point. sum ;

[0010] Step 5: Preliminary division of the shortest path tasks based on the U drone base stations dispatched by the regional center. Each part needs to undertake t av =t sum / U amount of tasks during the time;

[0011] Step 6: Form a closed path for each part, calculate the path time and the maximum difference between different paths; discuss the attribution of the boundary points of adjacent parts, and finally obtain a path division method with the minimum maximum difference between different paths;

[0012] Step 7: Use the improved particle swarm algorithm for each small part divided in step 6. The specific steps are as follows:

[0013] Step 8: The particle swarm size is set to 200, i=0, and the position and velocity of each particle are randomly generated under the initialization condition;

[0014] Step 9: Remove duplicate particles, and obtain the fitness value of each particle according to the utility function u. Each particle obtains the individual optimum and the global optimum.

[0015] Step 10, after adding 1 to the number of iterations, determine whether it exceeds the specified maximum number of iterations, if so, execute step 12, if not, execute step 11;

[0016] Step 11: Keep the global optimal particle unchanged, select the 10 particles with the highest fitness in the particle swarm, simulate the crossover and mutation operations in the genetic algorithm, cross the 10 particles in pairs to generate 10 new particles, and make the 10 new particles mutate themselves with a probability of 0.5 to enhance the diversity of particles; the remaining particles use the particle swarm algorithm update formula to update the particle position and speed, and execute step 9;

[0017] Step 12: End the loop and output the global optimal path.

[0018] Furthermore, in step 2,

[0019] Table 1 Parameter settings for four types of ground users with different communication requirements

[0020]

[0021] The ground user information parameters are shown in Table 1. Users have different priorities due to their different tasks. The larger the user priority value, the more important it is, and the higher the priority of its information. The amount of data to be sent by each type of user in each time period is fixed.

[0022] Furthermore, in step 4, the time for the UAV base station to complete the task includes the path flight time and the hovering coverage time; there are two ways for the UAV base station to complete the communication coverage, one is to cover while flying, and the other is to cover while hovering; considering the different urgency of the tasks undertaken by ground users, priority services are provided to users with high priority; when the UAV base station is flying to users with high priority, it adopts the method of flying while covering to provide communication services to ground users in the passing area;

[0023] First, the area is divided into small areas according to the maximum coverage radius of the drone base station. In the small area, the position and height of the drone base station are adjusted according to the distribution of users and differentiated communication needs, and the optimal coverage position of the drone base station in the small area is obtained; then the priority of the small area is calculated according to the priority of the users in the small area, and finally the trajectory of the drone base station is optimized according to the priority of the small area; when the drone base station flies to the small area with high priority, it adopts a communication coverage mode of flying and covering the small area it passes through, so as to reduce the total amount of data to be transmitted by users in the small area; if the flying time of the drone base station in the small area cannot completely transmit the user data, it is necessary to hover and cover at the optimal deployment point of the drone base station in the small area; therefore, the total time for the drone base station on the entire path to complete the task is the sum of the path flight time and the hovering time at the optimal coverage point:

[0024] t sum =t w +t s (1)

[0025] where t w is the path flight time, t s The time that the UAV base station hovers at the best coverage point; t w According to the optimized trajectory of the UAV base station, it is obtained from the geometric relationship; t s It is calculated based on the time taken by all users in a small area to transmit data and the flight time of the drone base station in the small area.

[0026] Furthermore, in step 6, considering the harsh regional environment, the longer the UAV base station stays in the air to complete the task, the greater the possibility of damage, and the time required to complete the task is the time taken by the UAV base station that takes the longest time. Therefore, it is very necessary to reasonably allocate the task volume of communication coverage in the entire area on the basis of determining the number of UAV base stations. The specific task allocation strategy includes calculating the overall task volume, preliminary task volume division and boundary point attribution discussion;

[0027] The calculation of the overall task volume is as follows: The calculation of the overall task volume can be regarded as the time for a single UAV base station to complete all tasks, that is, when the origin of the three-dimensional coordinates is defined as the center of the region, the time taken by the UAV base station to start from the origin, pass through the best coverage points of all small areas, and finally return to the origin of the coordinates; To ensure that the time to complete the task is the shortest, the first thing to do is to find the shortest time t required to complete the overall task volume. sum , which has been obtained in step 4;

[0028] Among them, the preliminary task volume division is as follows: when the regional center sends U drone base stations for coordination, the total task volume obtained above is initially divided into U parts according to the task equal division principle, and the time required for each part to complete the task is t av =t sum / U; When dividing the optimal deployment points of the UAV base station during task division, three points should be noted: First, considering that the points of the first and last two segments of the path are already connected to the starting point when the path is divided, and the middle path needs to be connected to the starting point after division, which takes more time than t av The first principle is to divide the two ends first and then the middle. Second, when the dividing point is at the optimal deployment position of the UAV base station, this point is divided into the part closer to the head or tail. Third, the discontinuity points between different parts are marked to facilitate the discussion of the boundary points below.

[0029] The discussion on the attribution of boundary points is as follows: each part forms a closed complete path, and the time required for each part to complete the task is obtained, and the maximum difference between the time required for different parts to complete the task is further obtained; then the attribution of boundary points of adjacent parts is discussed, and the division method that minimizes the maximum difference between the time required for different parts to complete the task is obtained, which is the final division method.

[0030] Furthermore, in step 9, considering the harshness of the regional environment, the user priorities are different for ground users with different tasks, and the data of important users should be transmitted back as soon as possible; therefore, when studying the multi-UAV collaborative trajectory optimization problem, it is necessary to balance the relationship between user priority, UAV energy consumption and UAV task completion time to achieve overall optimization; the optimized utility function needs to consider user priority, UAV energy consumption and the time it takes for the UAV base station to complete the task;

[0031] The specific steps of considering user priority are: giving priority to users with high priority, so that users with high priority have a shorter waiting time for service; introducing user satisfaction function as an indicator of the impact of user priority on the performance of drone trajectory optimization problem;

[0032] Considering the urgency of information in harsh regional environments, Y i As the user satisfaction in small area i, it is defined as:

[0033]

[0034] Among them, i refers to the serial number of the small area, C i Refers to the sum of the priorities of all users in small area i, t ai represents the time it takes for the drone base station a to reach the optimal deployment point in small area i, that is, the waiting time of small area i, t a represents the total time of the path where the drone base station a is located; when the priority of small area i is higher, C i The larger the value, the time t to reach small area i ai The smaller the value, the shorter the waiting time. a / t ai The larger the value, the higher the satisfaction level Y of small area i i The larger the value of i The larger the value, the more important the small area i is;

[0035] Then the satisfaction of all small areas in the region is:

[0036]

[0037] Where N represents the number of small areas;

[0038] The specific steps to consider the energy consumption of drones are as follows: The energy consumption of drones is an important issue to consider when drones are in motion, and the length of time a drone stays in the air is crucial in an emergency environment;

[0039] Considering that the motion energy consumption of the UAV base station is much greater than the hovering energy consumption and communication energy consumption, when reducing energy consumption is used as the optimization goal of the UAV base station deployment problem, research is conducted from the perspective of reducing the flight distance of the UAV base station;

[0040] Considering that both ground users and UAV base stations are moving, in order to better grasp their position changes, a time slicing method is used to divide a continuous time interval T into I sufficiently small discrete time periods, each of which is called a time slot, so that the position changes of ground users and UAV base stations in each small time slot can be ignored;

[0041] The dynamic deployment of drone base stations is prone to collision problems, so any two drones m and n (m≠n) should maintain a minimum safety distance when flying in the air within time slot t;

[0042]

[0043] Where S m [t] represents the location of the drone base station m at time slot t, d min Indicates the minimum safe distance between drone base stations to prevent collision;

[0044] In time slot t, the drone base station moves at a fixed speed V x The power consumption when moving in the horizontal direction is:

[0045]

[0046] Where P 0 , P 1 、v 1 、v 0 They represent the blade power, induced power, blade tip speed, and rotor induced speed of the UAV when it is hovering, respectively. 0 and ρ represent the drag ratio and air density of the UAV base station, respectively; s and A represent the solidity and blade area of ​​the UAV rotor, respectively;

[0047] When the horizontal speed of the drone base station V x =0 The energy consumption when hovering in the air is:

[0048] P h =P 0 +P 1 (6)

[0049] When the drone base station moves at a fixed speed V z The energy consumption when moving in the vertical direction is:

[0050] P z =P h +mg·V z (7)

[0051] Where m represents the weight of the drone base station, and g represents the acceleration of gravity;

[0052] The energy consumption of a dynamic UAV base station is mainly composed of motion energy, hovering energy and communication energy. The energy consumption of the UAV base station m in the time slot t is:

[0053]

[0054] Where P m is the transmission power of the UAV base station m;

[0055] When there are U drone base stations in the area, the total energy consumption of the drone base stations in time slot t is:

[0056]

[0057] Considering the time it takes for the drone base station to complete the task, the specific steps are as follows: For the problem of collaborative trajectory optimization of multiple drones, an important factor is the time it takes to complete the task, including the path flight time and the hovering coverage time; this part has been explained in detail in steps 4-6, but it should be noted that the time it takes for the drone to complete the task is the execution time T of the drone base station that takes the longest time after the task volume is divided. max ;

[0058] Considering the harsh regional environment, the shorter the time for the drone base station to perform tasks in the air, the safer it is, the earlier the user data with higher priority is transmitted back, and the less energy consumption the drone consumes when completing the task, the better. Therefore, when constructing the optimization target, the present invention comprehensively considers performance indicators such as task completion time, user priority and drone energy consumption; after allocating the total task of communication coverage to U drone base stations, each drone base station optimizes the motion trajectory according to the task allocation situation;

[0059] The utility function u is defined as the optimization target as the ratio of user satisfaction to the sum of the time to complete the task and the energy consumption of the drone base station:

[0060]

[0061] where α 1 , α 2 is the weight coefficient and sums to 1, β 1 , β 2 and β 3 is a normalized function; the time to complete the task is the longest execution time of the drone base station T max ;

[0062] The number of small areas and the number of drone base stations in the area are N and U respectively. The problem is described as:

[0063]

[0064] The problem constraints include: the spatial distance between any two drone base stations must be greater than or equal to the minimum safe distance, the optimal coverage point of each small area of ​​the drone base station can only be allocated to one drone path, the height of the drone base station is restricted, and the number of users covered by each drone when hovering cannot exceed the maximum user load number N of the drone. u ,The communication coverage condition of the UAV base station for the user is that the user’s receiving signal to interference and noise ratio must exceed its communication threshold.

[0065] Furthermore, in step 11:

[0066] Particle Swarm Optimization (PSO) is a swarm intelligence algorithm, which is inspired by the cooperative flight pattern of bird flocks. The algorithm explores the optimal solution by simulating a group of particles moving in the solution space, in which each particle adjusts its position in the search space according to its own fitness evaluation criteria. In the particle swarm optimization algorithm, each particle will track and update two key positions according to the value of the fitness function: one is the optimal solution it has encountered in the past (personal optimality), and the other is the optimal solution encountered by all particles in the entire particle swarm (global optimality). Until the maximum number of iterations is reached or the iteration termination condition is met, the global optimal solution is output.

[0067] The particle swarm algorithm is used to optimize the trajectory. The position of each particle represents a path sequence, and the speed is used to guide the moving direction and distance of the particle in the search space. The first and last points of each path sequence are the coordinate origin and are fixed. The update formula of the particle speed and position is as follows:

[0068]

[0069] where x i 、v i are respectively the position and velocity of the ith particle; r 1 and r 2 is a random number between [0,1], f 1 and f 2 is the acceleration constant, ω is the inertia factor, p ib is the optimal position in the search history of the i-th particle, p gb is the optimal position in the search history of all particles;

[0070] The traditional particle swarm algorithm has the problem of slow convergence speed and is easy to fall into the local optimal value. The inertia factor ω is the key parameter that determines the convergence speed of the algorithm. Increasing ω will enhance the global search ability of the particle swarm, and decreasing ω will enhance the local search ability of the particle swarm. Introducing the method of dynamically adjusting parameters, the value of ω is linearly reduced during the iteration process, which enhances the local search ability of the particle swarm and accelerates the convergence speed of the algorithm; the ω decreasing formula is as follows:

[0071]

[0072] where ω max ,ω min Represent the maximum and minimum values ​​of the inertia factor, D represents the current number of iterations, and D max represents the maximum number of iterations. As the number of iterations D increases, ω gradually increases from ω max Decrease to ω min ;

[0073] Considering that the traditional particle swarm algorithm is prone to premature convergence, the global optimal particle in each iteration remains unchanged, and the 10 particles with the highest fitness are selected according to the fitness function, and the crossover and mutation operations in the genetic algorithm are simulated. The 10 particles are cross-crossed to generate 10 new particles, and the 10 new particles are mutated with a probability of 0.5 to form a new particle swarm. This method helps to enhance the diversity of the particle swarm, break the local optimal trap of the particle swarm algorithm, and enhance the global search ability of the algorithm;

[0074] The particle swarm algorithm, which has been optimized by dynamically adjusting parameters and maintaining particle diversity, effectively balances global and local searches, improving search efficiency and solution quality. These improvements enhance the algorithm's adaptability to complex problems, accelerate convergence, and effectively avoid the risk of premature convergence, making the algorithm more robust and flexible in trajectory optimization scenarios.

[0075] Compared with the prior art, the significant progress of the present invention is that: under harsh environmental conditions, in order to address the problem of different priorities of ground users, the optimal coverage points of UAV base stations in small areas are found by reasonably dividing the areas, the total task volume is rationally divided, and the UAV path is optimized. Taking into account user priorities, this method effectively balances the overall performance of UAV energy consumption and task completion time.

[0076] In order to more clearly illustrate the functional characteristics and structural parameters of the present invention, further description is given below in conjunction with the accompanying drawings and specific implementation methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0078] Figure 1 It is a system model diagram;

[0079] Figure 2 This is a partial schematic diagram of the drone base station providing communication coverage services to ground users;

[0080] Figure 3It is a schematic diagram of the task allocation strategy;

[0081] Figure 4 It is a schematic diagram of the best coverage points of drone base stations based on user distribution;

[0082] Figure 5 The shortest coverage path diagram is obtained based on the traveling salesman algorithm;

[0083] Figure 6 The three optimal path diagrams are obtained according to the TOUP method;

[0084] Figure 7 It is a schematic diagram comparing the utility function values ​​of different methods;

[0085] Figure 8 It is a schematic diagram comparing the utility function values ​​when the number of ground users is different;

[0086] Figure 9 It is a schematic diagram of the average waiting time of different categories of users;

[0087] Figure 10 This is a schematic diagram comparing the sum of ground user satisfaction using different methods;

[0088] Figure 11 This is a schematic diagram comparing the total energy consumption of drone base stations using different methods;

[0089] Figure 12 This is a schematic diagram comparing the time it takes to complete a task using different methods;

[0090] Figure 13 It is a flowchart of a multi-UAV collaborative trajectory optimization method based on user priority. DETAILED DESCRIPTION

[0091] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0092] like Figure 1 , Figure 13 As shown, a multi-UAV collaborative trajectory optimization method based on user priority includes the following steps:

[0093] Step 1: Divide the area into N small areas according to the maximum coverage radius of the drone base station. In each small area, find the optimal coverage position of the drone base station according to the differentiated communication needs and distribution of the four types of users. The serial number is the same as the serial number of the small area, such as Figure 4 As shown;

[0094] Step 2: Calculate the total priority of users, the amount of data to be transmitted, and the required time in each small area according to the distribution of users;

[0095] The ground user information parameters are shown in Table 1. Users have different priorities due to their different tasks. The larger the user priority value, the more important it is, and the higher the priority of its information. The amount of data to be sent by each type of user in each time period is fixed.

[0096] Table 1 Parameter settings for four types of ground users with different communication requirements

[0097]

[0098] Step 3: Use the traveling salesman algorithm to find the shortest path starting from the origin, connecting N small-area drone base station optimal coverage points, and returning to the origin, such as Figure 5 As shown;

[0099] Step 4: Obtain the total time t required for the shortest path based on the time required for the path and the time required for the amount of data to be transmitted at the coverage point sum ;

[0100] The time it takes for the drone base station to complete a task includes the path flight time and the hovering coverage time. In the present invention, there are two ways for the drone base station to complete communication coverage, one is coverage while flying, and the other is hovering coverage. Taking into account the different urgency of the tasks undertaken by ground users, priority services are provided to users with high priority. When the drone base station is flying to users with high priority, it can provide communication services to ground users in the passing area by flying and covering.

[0101] like Figure 4 As shown in Figure 1, the area is first divided into small areas according to the maximum coverage radius of the drone base station. In the small area, the position and height of the drone base station are adjusted according to the distribution of users and differentiated communication needs to obtain the optimal coverage position of the drone base station in the small area (such as Figure 2 Medium 1 -O 3 ). Figure 2 Medium S I1 and S O1 They are the entry point and exit point of the small area with sequence number 1. The drone base station is at S I1 and S O1 When flying between small areas O 1Provide communication coverage for ground users in the small area, so that the total amount of data to be transmitted by users in the small area is reduced. If the UAV base station cannot fully transmit the user data in the small area during its flight time, it is necessary to hover and cover the optimal deployment point of the UAV base station in the small area. Assuming that small area 3 has a higher priority than small area 2, and small area 3 can only be reached through small area 2, the UAV base station will pass through S I2 , S O2 It passes through small area 2 and then reaches the optimal deployment point of the drone base station in small area 3 for hovering coverage. Therefore, the total time for the drone base station on the entire path to complete the task is the sum of the path flight time and the hovering time at the optimal coverage point:

[0102] t sum =t w +t s (1)

[0103] where t w is the path flight time, t s The time that the UAV base station hovers at the best coverage point. w According to the optimized trajectory of the UAV base station, it is obtained from the geometric relationship. s It is calculated based on the time taken by all users in a small area to transmit data and the flight time of the drone base station in the small area.

[0104] Step 5: Preliminary division of the shortest path tasks based on the U drone base stations dispatched by the command center. Each part needs to undertake t av =t sum / U amount of tasks during the time;

[0105] Step 6: Form a closed path for each part, calculate the path time and the maximum difference between different paths. Discuss the attribution of the boundary points of adjacent parts, and finally obtain the path division method with the smallest maximum difference between different paths;

[0106] Considering the harsh environment, the longer the drone base station stays in the air to complete the task, the greater the possibility of being destroyed, and the time required to complete the task is the time taken by the drone base station that takes the longest time. Therefore, it is very necessary to reasonably allocate the task volume of communication coverage in the entire area based on the determination of the number of drone base stations. The specific task allocation strategy is as follows:

[0107] (1) Calculation of the overall task volume: The calculation of the overall task volume can be regarded as the time it takes for a single UAV base station to complete all tasks, that is, when the origin of the three-dimensional coordinates is defined as the location of the command center, the time it takes for the UAV base station to start from the origin, pass through the best coverage points of all small areas, and finally return to the origin of the coordinates. To ensure that the task is completed in the shortest time, the first thing to do is to find the shortest time t required to complete the overall task volume. sum, which has been obtained in step 4.

[0108] (2) Preliminary task volume division: When the command center dispatches U drone base stations for coordination, the total task volume obtained above is preliminarily divided into U parts according to the task equal division principle. The time required for each part to complete the task is t av =t sum / U. When dividing the tasks, the main task is to divide the optimal deployment points of the UAV base station. Three points should be noted: First, considering that the points of the first and last two sections of the path are already connected to the starting point when the path is divided, and the middle path needs to be connected to the starting point after division, which may take longer than t av The first principle is to divide the two ends first and then the middle. Second, when the dividing point is at the optimal deployment position of the UAV base station, this point is divided into the part closer to the head or tail. Third, the discontinuity points between different parts are marked to facilitate the following discussion on the attribution of boundary points.

[0109] (3) Discussion on the attribution of boundary points: Form a closed complete path for each part, obtain the time required for each part to complete the task, and further obtain the maximum difference between the time required for different parts to complete the task. Then discuss the attribution of boundary points of adjacent parts, and obtain the division method that minimizes the maximum difference between the time required for different parts to complete the task, which is the final division method.

[0110] The task allocation strategy takes the coordination of three UAV base stations as an example. Figure 3 As shown:

[0111] Step 7: Use the improved particle swarm algorithm for each small part divided in step 6. The specific steps are as follows;

[0112] Step 8: The particle swarm size is set to 200, i=0, and the position and velocity of each particle are randomly generated under the initialization condition;

[0113] Step 9: Remove duplicate particles, and obtain the fitness value of each particle according to the utility function u. Each particle obtains the individual optimum and the global optimum.

[0114] Taking into account the particularity of the environment, for ground users with different tasks, the user priority is different, and the data of important users should be transmitted back as soon as possible. Therefore, when studying the problem of multi-UAV collaborative trajectory optimization, it is necessary to balance the relationship between user priority, UAV energy consumption and UAV task completion time to achieve overall optimization. Therefore, the utility function optimized by the present invention needs to be considered from the following aspects:

[0115] (1) User priority: Prioritize users with high priority, which is mainly reflected in the short waiting time for high-priority users. This chapter introduces the user satisfaction function as an indicator of the impact of user priority on the performance of the UAV trajectory optimization problem.

[0116] Considering the urgency of the information, the present invention will i As the user satisfaction in small area i, it is defined as:

[0117]

[0118] Where i refers to the serial number of the small area, C i Refers to the sum of the priorities of all users in small area i, t ai represents the time for the drone base station a to reach the optimal deployment point of small area i (i.e., the waiting time of small area i), t a represents the total time of the path where the drone base station a is located. When the priority of small area i is higher, C i The larger the value, the time t to reach small area i ai The smaller it is, the shorter the waiting time is. a / t ai The larger the value, the higher the satisfaction level Y of small area i i The larger the value of i The larger the value, the more important the small area i is.

[0119] Then the satisfaction of all small areas in the region is:

[0120]

[0121] Where N represents the number of small regions.

[0122] (2) UAV energy consumption: UAV energy consumption is an important issue that needs to be considered when the UAV is in motion. In addition, the length of time a UAV stays in the air is crucial in harsh environments.

[0123] Considering that the movement energy consumption of UAV base stations is much greater than the hovering energy consumption and communication energy consumption, when reducing energy consumption is used as the optimization goal of the UAV base station deployment problem, research can be conducted from the perspective of reducing the flight distance of the UAV base station.

[0124] Considering that both ground users and UAV base stations are moving, in order to better grasp their position changes, the time slicing method is used to divide a continuous time interval T into I sufficiently small discrete time periods, each of which is called a time slot, so that the position changes of ground users and UAV base stations in each small time slot can be ignored.

[0125] Collisions are prone to occur in the dynamic deployment of drone base stations, so any two drones m and n (m≠n) must maintain at least a minimum safety distance when flying in the air within time slot t.

[0126] ||S m [t]-S n [t]||≥d min (4)

[0127] Where S m [t] represents the location of the drone base station m at time slot t, d min Indicates the minimum safety distance between drone base stations to prevent collision.

[0128] In time slot t, the drone base station moves at a fixed speed V x The energy consumption when moving in the horizontal direction[4] is:

[0129]

[0130] Where P 0 , P 1 、v 1 、v 0 They represent the blade power, induced power, blade tip speed, and rotor induced speed of the UAV when it is hovering, respectively. 0 and ρ represent the drag ratio and air density of the UAV base station respectively, and s and A represent the rotor solidity and propeller disk area of ​​the UAV respectively.

[0131] When the horizontal speed of the drone base station V x =0 The energy consumption when hovering in the air is:

[0132] P h =P 0 +P 1 (6)

[0133] When the drone base station moves at a fixed speed V z The energy consumption when moving in the vertical direction is:

[0134] P z =P h +mg·V z (7)

[0135] Where m represents the weight of the drone base station, and g represents the acceleration of gravity.

[0136] The energy consumption of a dynamic UAV base station is mainly composed of motion energy, hovering energy and communication energy. That is, the energy consumption of a UAV base station m in time slot t is:

[0137]

[0138] Where P m is the transmission power of the UAV base station m.

[0139] When there are U drone base stations in the area, the total energy consumption of the drone base stations in time slot t is:

[0140]

[0141] (3) The time it takes for the drone base station to complete the task: For the problem of collaborative trajectory optimization of multiple drones, an important factor is the time it takes to complete the task, including the path flight time and the hovering coverage time. This part has been explained in detail in steps 4-6, but it should be noted that the time it takes for the drone to complete the task is the execution time T of the drone base station that takes the longest time after the task volume is divided. max .

[0142] Considering that in harsh environments, the shorter the time for drone base stations to perform tasks in the air, the safer it is, the earlier the data of users with higher priority is transmitted back, and the less energy consumption the drone consumes when completing the task, the better. Therefore, when constructing the optimization target, the present invention comprehensively considers performance indicators such as task completion time, user priority, and drone energy consumption. After allocating the total task of communication coverage to U drone base stations, each drone base station optimizes the motion trajectory according to the task allocation.

[0143] The utility function u is defined as the optimization target as the ratio of user satisfaction to the sum of the time to complete the task and the energy consumption of the drone base station:

[0144]

[0145] where α 1 , α 2 is the weight coefficient and sums to 1, β 1 , β 2 and β 3 is a normalized function. The time to complete the task is the longest execution time of the drone base station T max .

[0146] The number of small areas and the number of drone base stations in the area are N and U respectively. The problem can be described as:

[0147]

[0148] Constraints:

[0149] a) The spatial distance between any two drone base stations must be greater than or equal to the minimum safety distance;

[0150] b) The best coverage point of the drone base station in each small area can only be allocated to one drone path;

[0151] c) Limit the height of drone base stations;

[0152] d) The number of users covered by each drone when hovering cannot exceed the maximum user load N of the drone u ;

[0153] e) UAV base station coverage condition for user communication: the user's receiving signal-to-interference-noise ratio must exceed its communication threshold.

[0154] Step 10: After adding 1 to the number of iterations, determine whether it exceeds the specified maximum number of iterations. If so, execute step 12; if not, execute step 11;

[0155] Step 11: Keep the global optimal particle unchanged, select the 10 particles with the highest fitness in the particle swarm, simulate the crossover and mutation operations in the genetic algorithm, cross the 10 particles in pairs to generate 10 new particles, and make the 10 new particles mutate themselves with a probability of 0.5 to enhance the diversity of particles. The remaining particles use the particle swarm algorithm update formula to update the particle position and speed, and execute step 9;

[0156] Particle Swarm Optimization (PSO) is a swarm intelligence algorithm, which is inspired by the cooperative flight pattern of bird flocks. The algorithm explores the optimal solution by simulating a group of particles moving in the solution space, in which each particle adjusts its position in the search space according to its own fitness evaluation criteria. In the particle swarm optimization algorithm, each particle will track and update two key positions according to the value of the fitness function: one is the optimal solution it has encountered in the past (personal optimality), and the other is the optimal solution encountered by all particles in the entire particle swarm (global optimality). Until the maximum number of iterations is reached or the iteration termination condition is met, the global optimal solution is output.

[0157] When using the particle swarm algorithm to optimize the trajectory, the position of each particle represents a path sequence, and the speed is usually used to guide the movement direction and distance of the particle in the search space. In this chapter, the first and last points of each path sequence are the coordinate origin and are fixed. The update formula of the particle speed and position is as follows:

[0158]

[0159] where x i 、v i are the position and velocity of the ith particle respectively. 1 and r 2 is a random number between [0,1], f 1 and f 2 is the acceleration constant, ω is the inertia factor, p ib is the optimal position in the search history of the i-th particle, p gbis the optimal position in the search history of all particles.

[0160] (I) Dynamically adjust parameters

[0161] The traditional particle swarm algorithm has the problem of slow convergence and easy to fall into the local optimal value. The inertia factor ω is the key parameter that determines the convergence speed of the algorithm. Increasing ω will enhance the global search ability of the particle swarm, and decreasing ω will enhance the local search ability of the particle swarm. This chapter introduces a method of dynamically adjusting parameters, which linearly reduces the value of ω during the iteration process, enhances the local search ability of the particle swarm, and accelerates the convergence speed of the algorithm. The ω decreasing formula is as follows:

[0162]

[0163] where ω max ,ω min Represent the maximum and minimum values ​​of the inertia factor, D represents the current number of iterations, and D max represents the maximum number of iterations. As the number of iterations D increases, ω gradually increases from ω max Decrease to ω min .

[0164] 2. Maintaining Particle Diversity

[0165] Considering that the traditional particle swarm algorithm is prone to premature convergence, the global optimal particle in each iteration remains unchanged, and the 10 particles with the highest fitness are selected according to the fitness function, and the crossover and mutation operations in the genetic algorithm are simulated. The 10 particles are cross-crossed to generate 10 new particles, and the 10 new particles are mutated with a probability of 0.5. This method helps to enhance the diversity of the particle swarm, break the local optimal trap of the particle swarm algorithm, and enhance the global search ability of the algorithm.

[0166] The particle swarm algorithm, which has been optimized by dynamically adjusting parameters and maintaining particle diversity, effectively balances global and local searches, improving search efficiency and solution quality. These improvements enhance the algorithm's adaptability to complex problems, accelerate convergence, and effectively avoid the risk of premature convergence, making the algorithm more robust and flexible in trajectory optimization scenarios.

[0167] Step 12: End the loop and output the global optimal path;

[0168] The multi-UAV collaborative trajectory optimization algorithm (TOUP) based on user priority is experimentally simulated and analyzed on the simulation platform: MATLAB R2021a. Under the initial conditions, four types of heterogeneous users are randomly distributed in an area of ​​4km×4km. The settings of users and experimental simulation parameters are shown in Tables 1 and 2.

[0169] Table 2 Experimental simulation parameter settings

[0170]

[0171]

[0172] According to the maximum coverage radius of the drone base station, the area can be divided into 16 small areas. Under the initial conditions, the drone base stations all start from the coordinate origin. The performance of the TOUP algorithm proposed in this invention is compared with the following algorithms:

[0173] EATP algorithm: The algorithm takes into account the limited energy of the UAV and optimizes the trajectory of the UAV with the goal of reducing the energy consumption of the UAV under the premise of a known mission volume.

[0174] APOA algorithm: This is a new optimization method for generating time-optimal trajectories in a continuous-time polynomial framework. It achieves the goal of generating a minimum-time feasible trajectory by optimizing the segmented time and the polynomial in an alternating manner.

[0175] AEIA algorithm: This algorithm uses block coordinate descent (BCD) and successive convex approximation (SCA) techniques to optimize the trajectory of the drone with the goal of improving user throughput.

[0176] Figure 4 In order to divide the area according to the maximum coverage radius of the UAV base station, the optimal communication coverage point of the UAV base station is found in a small area according to the differentiated communication needs of users.

[0177] Figure 5 To find the shortest path starting from the origin, passing through all the best communication coverage points, and returning to the origin.

[0178] Figure 6 In order to divide the communication coverage tasks of the entire area according to the task volume, the TOUP algorithm is used to obtain three optimal paths.

[0179] Figure 7The figure is a schematic diagram of the relationship between the utility function value and the number of iterations when the number of ground users is 100. It can be seen from the simulation diagram that when the number of ground users is constant, the four types of algorithms gradually converge and stabilize with the increase of the number of iterations, and find the optimal path under the optimization target. In terms of utility function value, the TOUP algorithm is 23.1%, 24.8%, and 43.2% higher than the APOA, EATP, and AEIA algorithms, respectively. It can be seen from the figure that the TOUP algorithm is significantly better than other algorithms in terms of convergence speed. The TOUP algorithm takes the utility function under the three dimensions of user satisfaction, drone energy consumption, and task completion time as the optimization target, and has the best overall performance. The APOA and EATP algorithms take the task completion time and drone energy consumption as the optimization targets, respectively, and fail to consider the user's priority, so the overall performance is slightly lower. Moreover, considering that the drone base station's hovering time in harsh environments should not be too long, when optimizing the utility function, the weight of the task completion time is set higher than the weight of the drone energy consumption, so the performance of the APOA algorithm is slightly higher than that of the EATP algorithm. The AEIA algorithm optimizes user throughput without considering other factors, so the overall performance is low.

[0180] Figure 8 This is a schematic diagram comparing the utility functions of different algorithms when the number of ground users is different. From the simulation data, it can be seen that under the condition of a certain number of users, the TOUP algorithm is relatively optimal. The APOA and EATP algorithms take the time to complete the task and the energy consumption of the drone as the optimization targets respectively, and for the utility function, the time to complete the task has a greater weight than the energy consumption of the drone, so the performance of the APOA algorithm is higher than that of the EATP algorithm. However, the EATP algorithm also greatly reduces the time to complete the task to a certain extent, so the performance difference between the EATP algorithm and the APOA algorithm is not much. The AEIA algorithm only takes the user throughput as the optimization target, without considering other factors, so the overall performance is low.

[0181] Figure 9This is a schematic diagram comparing the average waiting time of different categories of users under different algorithm conditions. Taking 100 ground users as an example, in order to more intuitively reflect that the TOUP algorithm optimizes the drone trajectory problem based on user priority, the time for the drone base station to reach the best coverage point in a small area is defined as the time the user waits for service, and the average time for each category of users to wait for service is calculated. As can be seen from the figure, taking user class 1 as an example, the TOUP algorithm improves the average user waiting time by 22.5%, 23.7%, and 27.1% compared with the APOA, EATP, and AEIA algorithms, respectively. When optimizing the drone base station path, the TOUP algorithm first considers the user's priority. The larger the priority value, the higher the priority. Therefore, the average waiting time of user class 1 is the shortest, and so on. The other three algorithms do not consider user priority and can only plan the path based on their own optimization goals. Therefore, the user's waiting time is related to the user's distribution location, but not to the user's priority. Therefore, the TOUP algorithm has the best performance in reflecting user priority protection.

[0182] Figure 10 This is a comparative diagram of user satisfaction of different algorithms. The TOUP algorithm takes user satisfaction as part of the optimization goal, so it has the best performance, and its performance is improved by 36.6%, 30.2%, and 28.1% compared with the AEIA, EATP, and APOA algorithms, respectively. The other three algorithms do not consider the issue of user satisfaction, but the APOA algorithm considers the time issue and the EATP algorithm considers the energy consumption issue, which reduces the path time to a certain extent and improves user satisfaction, so the performance is slightly better than the AEIA algorithm.

[0183] Figure 11 This is a schematic diagram comparing the total energy consumption of drone base stations using different algorithms. When optimizing the trajectory of drone base stations, both the TOUP and EATP algorithms take drone energy consumption into consideration, and the TOUP algorithm converges faster than the EATP algorithm. However, the EATP algorithm only takes drone energy consumption as the optimization target, while the TOUP algorithm takes drone energy consumption as part of the optimization target, balancing the comprehensive optimization of user satisfaction, drone energy consumption, and task completion time. Therefore, the EATP algorithm is 7.2% better than the TOUP algorithm in terms of drone energy consumption. The APOA algorithm takes task completion time as the optimization target, which also reduces drone energy consumption to a certain extent. Therefore, the APOA algorithm has a slightly lower performance than the EATP algorithm. The AEIA algorithm takes user throughput as the optimization target and does not consider drone energy consumption, so its performance is poor.

[0184] Figure 12This is a comparative diagram of the task completion time of different algorithms. When optimizing the trajectory of the drone base station, both the TOUP and APOA algorithms take into account the time required to complete the task, but the APOA algorithm only takes the task completion time as the optimization target, and the TOUP algorithm takes the task completion time as part of the optimization target, achieving overall optimization in terms of user satisfaction, drone energy consumption, and task completion time. Therefore, the APOA algorithm is 3.9% better than the TOUP algorithm in terms of task completion time. The EATP algorithm takes drone energy consumption as the optimization target, and also reduces the task completion time to a certain extent, so the EATP algorithm has a slightly lower performance than the APOA algorithm. The AEIA algorithm takes the user throughput as the optimization target and does not consider the task completion time, so the performance is poor.

[0185] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0186] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A multi-UAV collaborative trajectory optimization method based on user priority, characterized in that: The steps include: Step 1: Divide the area into N small areas according to the maximum coverage radius of the drone base station. In each small area, find the optimal coverage position of the drone base station according to the differentiated communication needs and distribution of the four types of users. The serial number is the same as that of the small area. Step 2: Calculate the total priority of users, the amount of data to be transmitted, and the required time in each small area according to the distribution of users; Step 3: Use the traveling salesman algorithm to find the shortest path starting from the origin, connecting N small-area drone base station optimal coverage points, and returning to the origin; Step 4: Obtain the total time t required for the shortest path based on the time required for the path and the time required for the amount of data to be transmitted at the coverage point. sum ; Step 5: Preliminary division of the shortest path tasks based on the U drone base stations dispatched by the regional center. Each part needs to undertake t av =t sum / U amount of tasks during the time; Step 6: Form a closed path for each part, calculate the path time and the maximum difference between different paths; discuss the attribution of the boundary points of adjacent parts, and finally obtain a path division method with the minimum maximum difference between different paths; Step 7: Use the improved particle swarm algorithm for each small part divided in step 6. The specific steps are as follows: Step 8: The particle swarm size is set to 200, i=0, and the position and velocity of each particle are randomly generated under the initialization condition; Step 9: Remove duplicate particles, and obtain the fitness value of each particle according to the utility function u. Each particle obtains the individual optimum and the global optimum. Step 10, after adding 1 to the number of iterations, determine whether it exceeds the specified maximum number of iterations, if so, execute step 12, if not, execute step 11; Step 11: Keep the global optimal particle unchanged, select the 10 particles with the highest fitness in the particle swarm, simulate the crossover and mutation operations in the genetic algorithm, cross the 10 particles in pairs to generate 10 new particles, and make the 10 new particles mutate themselves with a probability of 0.5 to enhance the diversity of particles; the remaining particles use the particle swarm algorithm update formula to update the particle position and speed, and execute step 9; Step 12: End the loop and output the global optimal path.

2. The method for optimizing multi-UAV collaborative trajectories based on user priority according to claim 1 is characterized in that: In step 2, Table 1 Parameter settings for four types of ground users with different communication requirements The ground user information parameters are shown in Table 1. Users have different priorities due to their different tasks. The larger the user priority value, the more important it is, and the higher the priority of its information. The amount of data to be sent by each type of user in each time period is fixed.

3. The method for multi-UAV collaborative trajectory optimization based on user priority according to claim 1 is characterized in that: In step 4, the time it takes for the UAV base station to complete the task includes the path flight time and the hovering coverage time; There are two ways for drone base stations to complete communication coverage: one is to cover while flying, and the other is to cover while hovering. Considering the different urgency of tasks undertaken by ground users, priority services are provided to users with high priority. When the drone base station is flying to users with high priority, it adopts the method of covering while flying to provide communication services to ground users in the passing area. First, the area is divided into small areas according to the maximum coverage radius of the drone base station. In the small area, the position and height of the drone base station are adjusted according to the distribution of users and differentiated communication needs, and the optimal coverage position of the drone base station in the small area is obtained; then the priority of the small area is calculated according to the priority of the users in the small area, and finally the trajectory of the drone base station is optimized according to the priority of the small area; when the drone base station flies to the small area with high priority, it adopts a communication coverage mode of flying and covering the small area it passes through, so as to reduce the total amount of data to be transmitted by users in the small area; if the flying time of the drone base station in the small area cannot completely transmit the user data, it is necessary to hover and cover at the optimal deployment point of the drone base station in the small area; therefore, the total time for the drone base station on the entire path to complete the task is the sum of the path flight time and the hovering time at the optimal coverage point: t sum =t w +t s (1) where t w is the path flight time, t s The time for the drone base station to hover at the best coverage point; t w According to the optimized trajectory of the UAV base station, it is derived from the geometric relationship; t s It is calculated based on the time taken by all users in a small area to transmit data and the flight time of the drone base station in the small area.

4. The method for optimizing multi-UAV collaborative trajectories based on user priority according to claim 1, characterized in that: In step 6, it is very necessary to reasonably allocate the task volume of communication coverage in the entire area. The specific task allocation strategy includes calculating the overall task volume, preliminary task volume division and boundary point attribution discussion; The calculation of the overall task volume is as follows: The calculation of the overall task volume can be regarded as the time for a single UAV base station to complete all tasks, that is, when the origin of the three-dimensional coordinates is defined as the center of the region, the time taken by the UAV base station to start from the origin, pass through the best coverage points of all small areas, and finally return to the origin of the coordinates; To ensure that the time to complete the task is the shortest, the first thing to do is to find the shortest time t required to complete the overall task volume. sum , which has been obtained in step 4; Among them, the preliminary task volume division is as follows: when the regional center sends U drone base stations for coordination, the total task volume obtained above is initially divided into U parts according to the task equal division principle, and the time required for each part to complete the task is t av =t sum / U; When dividing the optimal deployment points of the UAV base station during task division, three points should be noted: First, considering that the points of the first and last two segments of the path are already connected to the starting point when the path is divided, and the middle path needs to be connected to the starting point after division, which takes more time than t av The first principle is to divide the two ends first and then the middle. Second, when the dividing point is at the optimal deployment position of the UAV base station, this point is divided into the part closer to the head or tail. Third, the discontinuity points between different parts are marked to facilitate the discussion of the boundary points below. The discussion on the attribution of boundary points is as follows: each part forms a closed complete path, and the time required for each part to complete the task is obtained, and the maximum difference between the time required for different parts to complete the task is further obtained; then the attribution of boundary points of adjacent parts is discussed, and the division method that minimizes the maximum difference between the time required for different parts to complete the task is obtained, which is the final division method.

5. The method for optimizing multi-UAV collaborative trajectories based on user priority according to claim 1 is characterized in that: In step 9, considering the harshness of the regional environment, the user priorities are different for ground users with different tasks, and the data of important users should be transmitted back as soon as possible; the relationship between user priority, drone energy consumption and drone task completion time is balanced to achieve overall optimization; the optimized utility function needs to consider user priority, drone energy consumption and drone base station task completion time; The specific steps of considering user priority are: giving priority to users with high priority, so that users with high priority have a shorter waiting time for service; By introducing the user satisfaction function as an indicator of the impact of user priority on the performance of the UAV trajectory optimization problem; Considering the urgency of information in harsh regional environments, Y i As the user satisfaction in small area i, it is defined as: Among them, i refers to the serial number of the small area, C i Refers to the sum of the priorities of all users in small area i, t ai represents the time when the UAV base station a reaches the optimal deployment point of small area i, that is, the waiting time of small area i, t a represents the total time of the path where the drone base station a is located; when the priority of small area i is higher, C i The larger the value, the time t to reach small area i ai The smaller the value, the shorter the waiting time. a / t ai The larger the value, the higher the satisfaction level Y of small area i i The larger the value of i The larger the value, the more important the small area i is; Then the satisfaction of all small areas in the region is: Where N represents the number of small areas; The specific steps to consider the energy consumption of drones are as follows: Since the motion energy consumption of drone base stations is much greater than the hovering energy consumption and communication energy consumption, when reducing energy consumption is used as the optimization goal of the drone base station deployment problem, research is conducted from the perspective of reducing the flight distance of drone base stations; Using the time slicing method, a continuous time interval T is divided into I sufficiently small discrete time periods, each of which is called a time slot, so that the position changes of ground users and UAV base stations in each small time slot can be ignored; The dynamic deployment of drone base stations is prone to collision problems, so any two drones m and n (m≠n) should maintain a minimum safety distance when flying in the air within time slot t; ||S m [t]-S n [t]||≥d min (4) Where S m [t] represents the location of the drone base station m at time slot t, d min Indicates the minimum safe distance between drone base stations to prevent collision; In time slot t, the drone base station moves at a fixed speed V x The power consumption when moving in the horizontal direction is: Among them, P0, P1, v1, and v0 represent the blade power, induced power, blade tip speed, and rotor induced speed of the drone when it is hovering, respectively; d0 and ρ represent the drag ratio and air density of the drone base station, respectively; s and A represent the rotor solidity and blade disc area of ​​the drone, respectively; When the horizontal speed of the drone base station V x =0 The energy consumption when hovering in the air is: P h =P0+P1 (6) When the drone base station moves at a fixed speed V z The energy consumption when moving in the vertical direction is: P z =P h +mg·V z (7) Where m represents the weight of the drone base station, and g represents the acceleration of gravity; The energy consumption of a dynamic UAV base station is mainly composed of motion energy, hovering energy and communication energy. The energy consumption of the UAV base station m in the time slot t is: Where P m is the transmission power of the UAV base station m; When there are U drone base stations in the area, the total energy consumption of the drone base stations in time slot t is: Considering the time it takes for the UAV base station to complete the task, the specific steps are as follows: For the problem of collaborative trajectory optimization of multiple UAVs, an important factor is the time it takes to complete the task, including the path flight time and the hovering coverage time; the time it takes for the UAV to complete the task is the execution time T of the UAV base station that takes the longest time after the task volume is divided. max ; When constructing the optimization target, performance indicators such as task completion time, user priority, and drone energy consumption are comprehensively considered; after allocating the total communication coverage task to U drone base stations, each drone base station optimizes the motion trajectory according to the task allocation; The utility function u is defined as the optimization target as the ratio of user satisfaction to the sum of the time to complete the task and the energy consumption of the drone base station: Among them, α1 and α2 are weight coefficients and their sum is 1, β1, β2 and β3 are normalized functions; the time to complete the task is the execution time T of the longest drone base station max ; The number of small areas and the number of drone base stations in the area are N and U respectively. The problem is described as: The problem constraints include: the spatial distance between any two drone base stations must be greater than or equal to the minimum safe distance, the optimal coverage point of each small area of ​​the drone base station can only be allocated to one drone path, the height of the drone base station is restricted, and the number of users covered by each drone when hovering cannot exceed the maximum user load number N of the drone. u ,The communication coverage condition of the UAV base station for the user is that the user’s receiving signal to interference and noise ratio must exceed its communication threshold.

6. The method for optimizing multi-UAV collaborative trajectories based on user priority according to claim 1, characterized in that: In step 11: The particle swarm algorithm is used to optimize the trajectory. The position of each particle represents a path sequence, and the speed is used to guide the moving direction and distance of the particle in the search space. The first and last points of each path sequence are the coordinate origin and are fixed. The update formula of the particle speed and position is as follows: where x i 、v i They represent the position and velocity of the ith particle respectively; r1 and r2 are random numbers between [0,1], f1 and f2 are acceleration constants, ω is the inertia factor, and p ib is the optimal position in the search history of the i-th particle, p gb is the optimal position in the search history of all particles; The method of dynamically adjusting parameters is introduced to linearly reduce the value of ω during the iteration process, enhance the local search ability of the particle swarm, and accelerate the convergence speed of the algorithm; the ω decreasing formula is as follows: where ω max ,ω min Represent the maximum and minimum values ​​of the inertia factor, D represents the current number of iterations, and D max represents the maximum number of iterations. As the number of iterations D increases, ω gradually increases from ω max Decrease to ω min ; The global optimal particle that appears in each iteration remains unchanged, and the 10 particles with the highest fitness are selected according to the fitness function. The crossover and mutation operations in the genetic algorithm are simulated, and 10 particles are cross-crossed in pairs to generate 10 new particles. The 10 new particles are mutated with a probability of 0.5 to form a new particle group.

Images