A resource scheduling method for swarm drones based on intelligent evolutionary algorithm

Through intelligent evolution algorithm and PGC-NSGAII algorithm, combined with monitoring and control modules, static and dynamic scheduling solutions are designed, and the efficiency of cluster drone scheduling in large-scale and dynamic environments is solved, and efficient resource scheduling and flight ban are achieved. Area evasion improves the performance of task completion.

CN115113647BActive Publication Date: 2024-11-05DALIAN MARITIME UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210813729.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-07-11
Publication Date
2024-11-05
Estimated Expiration
2042-07-11

AI Technical Summary

Technical Problem

The existing cluster drone scheduling method is limited in the face of large-scale drones and dynamic environments, making it difficult to efficiently schedule, especially under the comprehensive consideration of energy consumption, flight routes and no-fly zone constraints, which makes it difficult to achieve effective The task is completed.

Method used

Using a cluster UAV resource scheduling method based on intelligent evolution algorithm, by establishing a reconnaissance and tracking task processing model, combining monitoring, control and execution modules, static and dynamic scheduling schemes are designed, and sub-region coding and PGC-NSGAII algorithm are used to consider unmanned Energy consumption and dynamic environment, realize drone position prediction and route planning, and avoid no-fly areas.

Benefits of technology

It realizes efficient drone resource scheduling under dynamic environment and multi-constraint conditions, improves resource scheduling performance for task completion, can effectively overcome the harm caused by energy consumption and obstacles, and provides a better resource scheduling strategy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115113647B_ABST
    Figure CN115113647B_ABST
Patent Text Reader

Abstract

The present invention provides a cluster UAV resource scheduling method based on an intelligent evolutionary algorithm, which relates to the field of wireless communication technology, including: S1, establishing a cluster UAV reconnaissance and tracking task processing model; S2, designing a cluster UAV resource scheduling method based on the cluster UAV reconnaissance and tracking task processing model; S3, determining the UAV position according to the parameter prediction method in the scheduling method. Determine the UAV position according to the scheduling method. The present invention proposes a cluster UAV resource scheduling method based on intelligent evolution, which optimizes the threat level generated by the task target when the task is completed and the energy consumption of a single UAV by adding a preprocessing process, a group-based selection crossover method and a regional coding method. The present invention simultaneously schedules multiple UAVs equipped with radar equipment to perform target reconnaissance and target tracking tasks, and improves the processing efficiency of cluster UAVs for tasks through the present invention.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to a cluster drone resource scheduling method based on an intelligent evolutionary algorithm. Background Art

[0002] When using swarm drone technology to perform tasks, the large-scale number of swarm drones improves the efficiency of reconnaissance or tracking tasks. However, the scheduling of large numbers of drones is a thorny issue. How to efficiently schedule swarm drones is a prerequisite for efficiently completing reconnaissance and tracking tasks.

[0003] In the existing scheduling of swarm drones, most of them consider various restrictions faced by swarm drones separately, such as: the energy consumption of a single drone, the flight route of the drone is restricted by the area that cannot be vertically crossed (no-fly zone), the priority restrictions of the mission objectives, and the processing time of the mission objectives. However, the above restrictions do not exist alone, so they need to be considered together.

[0004] At present, the scheduling methods for swarm drones are mainly divided into several categories: (1) Scheduling methods based on trajectory planning. This method considers the obstacles encountered by drones on the path, plans multiple drone flight routes, and makes selections. This method is mostly offline and is not suitable for performing tasks in actual dynamic scenes. (2) Scheduling methods based on heuristic algorithms. Heuristic algorithms have relatively good properties in small-scale drone scheduling, but when the number of drones increases, the effectiveness of the algorithm will decrease. (3) Scheduling methods based on intelligent evolutionary algorithms. Intelligent evolutionary algorithms are suitable for large-scale combinatorial problems and NP (Non-deterministic Polynomial) difficult problems. They have unique advantages in large-scale swarm drone scheduling. However, due to the limitations of actual scenarios, the effectiveness of existing intelligent evolutionary algorithms is limited. Therefore, they need to be improved. Summary of the invention

[0005] In view of this, the purpose of the present invention is to propose a cluster UAV resource scheduling method based on an intelligent evolutionary algorithm to solve the problem of difficulty in the existing joint scheduling of multiple UAVs.

[0006] The technical means adopted by the present invention are as follows:

[0007] A cluster drone resource scheduling method based on intelligent evolutionary algorithm, comprising:

[0008] S1. Establish a cluster UAV reconnaissance and tracking task processing model;

[0009] S2. Design a cluster UAV resource scheduling method based on the cluster UAV reconnaissance and tracking task processing model;

[0010] S3. Determine the position of the UAV according to the parameter prediction method in the scheduling method.

[0011] Furthermore, S1 includes: analyzing the problems faced when cluster UAVs perform reconnaissance and tracking tasks, obtaining problem constraints, establishing a cluster UAV reconnaissance and tracking task processing model, considering the completion time of different types of tasks and the maximum energy consumption of a single UAV, and determining the threat level generated by the target when all tasks are completed and the flight distance of a single UAV as the optimization target; through the optimization target combined with the intelligent evolutionary algorithm, an idealized schedule is obtained to complete the allocation of all task targets.

[0012] Furthermore, the cluster UAV reconnaissance and tracking task processing model includes a monitoring module, a control module, an execution module and a timing module;

[0013] A monitoring module, wherein the monitoring module monitors the dynamic change information of the drone, wherein the dynamic change information includes the task location and the task quantity, and the monitoring module sends the dynamic change information to the control module; when a recurrence occurs in the scene or a tracking task is performed, the monitoring module sends a timing parameter to the control module;

[0014] A control module, wherein the control module receives dynamic change information and runs a cluster drone resource scheduling method, selects a corresponding static scheduling scheme and sends it to an execution module; the control module receives timing parameters and runs a cluster drone resource scheduling method, selects a corresponding dynamic scheduling scheme and sends it to an execution module;

[0015] An execution module, the execution module receives and executes the static scheduling scheme and the dynamic scheduling scheme;

[0016] A timing module records the system time, starts timing when the execution module starts executing and sends a time signal to the monitoring module.

[0017] Further, S2 includes the following steps:

[0018] S21, considering that different task objectives are subject to different coding restrictions and requirements, a regional coding method is established, wherein the coding method adopts integer coding;

[0019] S22, population preprocessing: select several individuals with excellent performance from all individuals in the population as the initial population of the algorithm;

[0020] S23, calculating the objective function value of the initial population according to the region coding method, and obtaining the non-dominated level of individuals in the initial population;

[0021] S24, selecting a parent population for crossover in the initial population according to the non-dominated level, and crossover the parent population using a grouping-based selection crossover method to obtain a crossover population;

[0022] S25, mutating the crossover population to obtain a progeny population;

[0023] S26, mixing the offspring population with the initial population to obtain a mixed population, and selecting individuals in the mixed population as the offspring population of the next generation according to the non-dominated level and the crowding degree;

[0024] S27, check whether the termination iteration number is met, and repeat S21 to S26 when the termination iteration number is not met. When the termination iteration number is met, output the initial population at this time.

[0025] Further, S21 includes the following steps:

[0026] S211, dividing the code into two parts, the first part of the code represents the drone serial number and task execution order required to perform the ordinary task, and the second part of the code represents the drone serial number and task execution order required for the tracking task.

[0027] Further, S22 includes the following steps:

[0028] S221, setting the termination condition, crossover parameter, mutation parameter and parameters required for population generation as parameters required for population preprocessing;

[0029] S222, randomly encoding and generating a number of feasible solution individuals in the global scope, and selecting an individual with a Pareto dominance level of 1 from the several feasible solution individuals and putting it into an external archive;

[0030] S223. When S222 does not satisfy the number of populations required to reach the initial population, S222 is repeated. When S222 satisfies the number of populations required to reach the initial population, the individuals in the external archive at this time are used as the initial population.

[0031] Further, S24 includes the following steps:

[0032] S241, take out the non-dominated level of each individual in the parent population;

[0033] S242, placing the individual with the best objective function value in each non-dominated level into the excellent set, and the remaining individuals into the ordinary set;

[0034] S243, randomly select two individuals from the excellent set for comparison, retain the individual with the best objective function and compare it with individuals in other excellent sets for several times, and put the retained individual after the comparison into the final set;

[0035] S244. Select the individual with the best objective function in the final set and cross-match it with the common set at the crossover probability to obtain a crossover population.

[0036] Furthermore, a scheduling method in S2 includes:

[0037] S28. adopt a no-fly zone avoidance method to enable the drone to avoid unreachable areas during flight;

[0038] S281. Calculate the distances between the starting point A of the drone and the six vertices of the no-fly zone, record them as D1 to D6, and select the smallest distance D among D1 to D6. MIN ;

[0039] S282, Comparison D MIN The length of the no-fly zone is D O , select the set of vertices C that can be directly connected to A A , the selection method is: if D MIN <D O , then the two vertices with the smallest distance A among D1 to D6 are regarded as the directly connectable vertex set; if D MIN ≥D O , then the three vertices with the smallest distance A in D1~D6 are the set of directly connected vertices;

[0040] S283: Execute S282 for the end point A' where the drone crosses the no-fly zone, and obtain a set of directly connectable points C of A'. A’ ;

[0041] S284, take C A , C A’ If the intersection is empty, the vertex of the no-fly zone is used as the transfer point to fly around the edge of the no-fly zone;

[0042] S285, if C A , C A’ If the intersection of is not empty, then select a point with a small distance and fewer selected points in the intersection as a transfer point for detour.

[0043] Furthermore, S3 includes the following steps:

[0044] S31, determining the target space corresponding to each drone according to the static scheduling method in the cluster drone resource scheduling method;

[0045] S32, dividing the flight route of the UAV into several stages according to the mission objectives in the target space;

[0046] S33, according to the execution time of the scheduling plan, the flight speed of the drone and the speed-distance formula, in the route stage in S32, determine the stage I that the drone is currently in;

[0047] S34, taking the starting point of the stage I as the center of the circle O(o1,o2), taking the distance that the drone flies over the starting point of the stage I as the radius, and making an equation of the circle Γ(o1,o2);

[0048] S35, connecting the starting point and the end point of the stage I to obtain the straight line equation I(l) determined by the two points, solving the circle equation and the straight line equation and removing meaningless solutions, and the remaining solution is the predicted position of the drone.

[0049] Compared with the prior art, the present invention has the following advantages:

[0050] The present invention proposes a dual-mode model that combines static and dynamic modes. The model takes into account the energy constraints, dynamic environmental constraints and no-fly zone constraints faced by homogeneous multi-UAVs at the same time, so that the UAVs can efficiently overcome the hazards brought by the dynamic environment and obstacle blocking problems when facing the former.

[0051] The present invention solves the problems of dynamic scheduling of cluster UAVs and no-fly zone constraints by proposing a parameter prediction method and a no-fly zone avoidance method.

[0052] The present invention proposes a PGC-NSGAII algorithm based on embedded group selection crossover, wherein a group selection crossover (GBSC) method is used to improve the diversity of solutions. In order to improve the effectiveness of the algorithm, a step is added before the algorithm to preprocess the population.

[0053] The present invention verifies through multiple groups of multi-angle experiments that the method of the present invention can effectively obtain a better resource scheduling strategy. Compared with other methods, the method of the present invention better guarantees the resource scheduling performance of cluster drones to complete tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0055] Figure 1 A flow chart of a cluster UAV resource scheduling method is designed for the present invention.

[0056] Figure 2 This is a system model framework diagram of the present invention.

[0057] Figure 3 This is a preprocessing flow chart of the present invention.

[0058] Figure 4Flow chart of the crossover method for group selection of the present invention.

[0059] Figure 5 It is a schematic diagram of the triangulation of the no-fly zone of the present invention.

[0060] Figure 6 The diagram of flying around a no-fly zone when the intersection is empty is shown in the present invention.

[0061] Figure 7 This is a schematic diagram of flying around a no-fly zone when the intersection is not empty according to the present invention.

[0062] Figure 8 The comparison chart of the results of PGC-NSGAII and NSGAII algorithms when the total number of UAVs is 50 and the total number of missions is 150.

[0063] Fig. 9 The comparison chart of the results of PGC-NSGAII and NSGAII algorithms when the total number of UAVs is 50 and the total number of missions is 300.

[0064] Fig.10 This is a comparison chart of the results of PGC-NSGAII and NSGAII algorithms when the total number of UAVs is 50 and the total number of missions is 150, and one tracking target is added.

[0065] Fig.11 This is a comparison chart of the results of PGC-NSGAII and NSGAII algorithms when the total number of UAVs is 50 and the total number of missions is 300, and one tracking target is added.

[0066] Fig.12 This is a comparison chart of the optimization of PGC-NSGAII and NSGAII on the ZDT1 function.

[0067] Fig.13 This is an optimization comparison chart between PGC-NSGAII and NSGAII on the ZDT2 function.

[0068] Fig.14 This is a comparison chart of the optimization of PGC-NSGAII and NSGAII on the ZDT3 function.

[0069] Fig.15 This is an optimization comparison chart between PGC-NSGAII and NSGAII on the ZDT4 function.

[0070] Fig.16 This is a comparison chart of the optimization of PGC-NSGAII and NSGAII on the ZDT5 function.

[0071] Fig.17 This is an optimization comparison chart between PGC-NSGAII and NSGAII on the ZDT6 function. DETAILED DESCRIPTION

[0072] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme 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, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0073] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0074] The present invention provides a cluster drone resource scheduling method based on an intelligent evolutionary algorithm, comprising:

[0075] S1. Establish a cluster UAV reconnaissance and tracking task processing model;

[0076] S1 includes: analyzing the problems faced when cluster UAVs perform reconnaissance and tracking tasks, obtaining problem constraints, establishing a cluster UAV reconnaissance and tracking task processing model, considering the completion time of different types of tasks and the maximum energy consumption of a single UAV, and determining the threat level of the target when all tasks are completed and the flight distance of a single UAV as the optimization target; deriving an idealized schedule through the optimization target combined with an intelligent evolutionary algorithm to complete the allocation of all task targets; the cluster UAV reconnaissance and tracking task processing model includes a monitoring module, a control module, an execution module and a timing module.

[0077] A monitoring module, wherein the monitoring module monitors the dynamic change information of the drone, wherein the dynamic change information includes the task location and the task quantity, and the monitoring module sends the dynamic change information to the control module; when a recurrence occurs in the scene or a tracking task is performed, the monitoring module sends a timing parameter to the control module;

[0078] A control module, wherein the control module receives dynamic change information and runs a cluster drone resource scheduling method, selects a corresponding static scheduling scheme and sends it to an execution module; the control module receives timing parameters and runs a cluster drone resource scheduling method, selects a corresponding dynamic scheduling scheme and sends it to an execution module;

[0079] An execution module, the execution module receives and executes the static scheduling scheme and the dynamic scheduling scheme;

[0080] A timing module records the system time, starts timing when the execution module starts executing and sends a time signal to the monitoring module.

[0081] S2. Design a cluster UAV resource scheduling method based on the cluster UAV reconnaissance and tracking task processing model;

[0082] S21. Considering that different task targets are subject to different coding restrictions and requirements, a regional coding method is proposed. Different task targets are coded in different regions. The coding uses integer coding.

[0083] S211. Divide the entire code into two parts. The first part of the code represents the UAV serial number and task execution order required to perform ordinary tasks, and the second part represents the UAV serial number and task execution order required for tracking tasks.

[0084] S22, population preprocessing: by selecting a number of individuals with excellent performance in the global scope as the initial population of the algorithm (hereinafter referred to as P0). The flowchart of preprocessing is as follows Figure 3 shown.

[0085] S221. Set the required parameters: termination conditions, crossover parameters, mutation parameters, and parameters required for population generation, etc.

[0086] S222. Randomly encode globally to generate certain feasible solution individuals. Among the generated feasible solution individuals, select relatively excellent individuals and put them into external archive.

[0087] S223, determine the termination condition, if the termination condition is not met, return to S222. If the termination condition is met, jump out of the loop, at this time, the individuals in the external archive are used as the initial population of the algorithm, and enter the next step of the algorithm.

[0088] S23. Calculate the objective function value of the population and obtain the non-dominated level of the individuals in the population.

[0089] S24, selecting a parent population for crossover (hereinafter referred to as P1) according to the non-dominated level in S23, and then performing crossover using a grouping-based selection crossover method. The steps of the new crossover method are as follows Figure 4 As shown:

[0090] S241. Take out the non-dominated level of each individual in P1.

[0091] S242, the better performing ones in each non-dominated level are put into a set (here named C m The remaining individuals are placed in another set (here named C f gather).

[0092] S243, C m Two individuals in the set are randomly selected for comparison, the one with the higher fitness is retained and the others are accepted in C m The selected individuals in the set are compared, and the comparison will continue for several times. After the comparison is completed, the last retained individuals are put into a set (here named C b gather).

[0093] S244、C b The individual with the best fitness is selected from the set and combined with C with a certain probability. f Cross-matching of collections.

[0094] S25, mutate the individuals obtained in S24 to obtain the offspring population (hereinafter referred to as P3)

[0095] S26, P3 and P0 are mixed, and individuals are selected as P0 of the next generation according to the non-dominated level and crowding degree.

[0096] S27, check whether the termination condition is met, if not, repeat S21 to S26. If the termination condition is met, jump out of the loop and take out P0.

[0097] S28. A no-fly zone avoidance method is adopted to enable the UAV to avoid inaccessible areas during flight.

[0098] S281, such as Figure 5 As shown, calculate the distances between A and the six vertices of the no-fly zone, record them as (D1~D6), and select the smallest distance D among D1~D6. MIN .

[0099] S282, Comparison D MIN The length of the no-fly zone is D O , select the directly connected vertex set C of A according to the following definition A (No need to cross the no-fly zone, vertices that can be connected by a straight line from A): If D MIN <D O , then the two vertices with the smallest distance A among D1 to D6 are considered as the directly connectable vertex set. MIN >D O , then the three vertices in D1~D6 with the smallest distance A are regarded as the set of directly connected vertices.

[0100] S283, execute the steps in S282 on A' to obtain a set C of directly connectable points of A' A '.

[0101] S284, take C A ,C A ', if the intersection is empty, then take the vertex of the no-fly zone as the transfer point and fly around the edge of the no-fly zone (such as Figure 6 shown).

[0102] S285, if C A ,C A ' is not empty, then select a suitable point (with a small distance and few points) in the intersection as a transfer point for detour (such as Figure 7 (shown)

[0103] S3. Determine the position of the drone according to the scheduling method. This method uses the planned drone flight route and the drone flight speed combined with the speed-distance formula to estimate the parameters required for dynamic scheduling. The basic steps are:

[0104] S31. Determine the target space corresponding to each UAV (the set of tasks performed by each UAV) according to the scheduling plan.

[0105] S32. According to the mission objectives in the target space of the UAV, the flight route of the UAV is divided into several stages.

[0106] S33, according to the execution time of the scheduling plan, the flight speed of the drone and the speed-distance formula, in the route stage in S32, determine the stage the drone is currently in (hereinafter referred to as stage I).

[0107] S34, taking the starting point of stage I as the center of the circle Use the distance the drone flies past the starting point of Phase I as the radius and make the equation of the circle

[0108] S35, connect the starting point and the end point of stage I to obtain the equation of the straight line determined by the two points I(l). Solve the equation:

[0109]

[0110] The solution obtained, after removing the meaningless solutions, is the position of the drone predicted in this paper. Figure 5 In the figure, the position of the black asterisk is the position of the UAV predicted by the parameters. The determination of the position of the UAV means that the properties of the remaining targets are also determined.

[0111] This paper proposes a cluster UAV resource scheduling method based on intelligent evolution, combined with a new evolutionary algorithm PGC-NSSGAII (Preprocessing and Group Crossover Non-dominated SortGenetic Algorithm). By adding a preprocessing process, a group-based selection crossover method and a regional coding method on the basis of the traditional NSGAII algorithm, the threat level generated by the mission target when the mission is completed and the energy consumption of a single UAV are taken as optimization goals.

[0112] This embodiment conducts experiments in actual mission scenarios, and tests are performed under different scales of drone groups and mission numbers. The comparison algorithm used in this article is the NSGAII algorithm, which performs well among the current mainstream algorithms, for comparison experiments. The following comparison experiments are conducted for different mission objectives, as well as different drones and mission objective scales.

[0113] A comparison of the two algorithms is Figure 8 Fig. 9 (The black diamond in the figure is the result of the PGC-NSGAII algorithm, and the black is the result of the NSGAII algorithm. The results of the two algorithms show that in terms of distance optimization and threat optimization, the results of PGC-NSGAII are closer to the lower left than those of NSGAII, that is, the distance and threat values ​​of the former are smaller than those of the latter, which shows that PGC-NSGAII has more advantages than NSGAII in multi-UAV dynamic environment scheduling. When the total number of UAVs in the two algorithms is 50 and the total number of tasks is 150, the results are compared as follows Figure 8 When the total number of drones and the total number of tasks of the two algorithms is 50 and 300, the results are compared. Fig. 9 Because a single tracking task requires four drones to cooperate, in order to compare the resource scheduling efficiency of the two algorithms for clustered drones in a dynamic environment, the two algorithms increase the number of tracking tasks by 50 drones and 150 tasks. Fig.10 When the total number of drones and the total number of tasks of the two algorithms is 50 and 300, the results are compared as follows: Fig.11 In order to compare the optimization performance of the mathematical functions of the two algorithms, the two algorithms are compared in the test functions of the ZDT series, such as Figure 12-Figure 17 The PF in the figure represents the Pareto theoretical frontier of this optimization function. The black diamonds represent the results of PGC-NSGAII, and the black circles represent the results of NSGAII.

[0114] Depend on Figure 12-Figure 17It can be concluded from Table 1 that the improved algorithm has similar effects on the ZDT1~4 and ZDT6 test problems, and on the ZDT5 test problem, the solution set obtained by the improved algorithm is closer to the Pareto frontier. Therefore, it can be said that the improved algorithm can achieve a solution set similar to that of the algorithm before the improvement on the ZDT1~4 and ZDT6 test problems, and has a better effect on the ZDT5 test problem.

[0115] Table 1 Comparison of the two algorithms before and after improvement on the ZDT5 test function (where Num represents the number of non-dominated solutions, and Dis represents the average distance from the non-dominated solution to the Pareto Front)

[0116]

[0117] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A cluster drone resource scheduling method based on intelligent evolutionary algorithm, characterized in that: include: S1. Establish a cluster UAV reconnaissance and tracking task processing model; Based on the analysis of the problems faced by swarm drones when performing reconnaissance and tracking tasks, the problem constraints are obtained, and a swarm drone reconnaissance and tracking task processing model is established. Considering the completion time of different types of tasks and the maximum energy consumption of single drones, the threat level of the target when all tasks are completed and the flight distance of a single drone are determined as optimization targets; the idealized scheduling is obtained by combining the optimization target with the intelligent evolutionary algorithm to complete the allocation of all task targets; S2. Design a cluster UAV resource scheduling method based on the cluster UAV reconnaissance and tracking task processing model; S21, considering that different task objectives are subject to different coding restrictions and requirements, a regional coding method is established, wherein the coding method adopts integer coding; S22, population preprocessing: select several individuals with excellent performance from all individuals in the population as the initial population of the algorithm; S23, calculating the objective function value of the initial population according to the region coding method, and obtaining the non-dominated level of individuals in the initial population; S24, selecting a parent population for crossover in the initial population according to the non-dominated level, and crossover the parent population using a grouping-based selection crossover method to obtain a crossover population; S25, mutating the crossover population to obtain a progeny population; S26, mixing the offspring population with the initial population to obtain a mixed population, and selecting individuals in the mixed population as the offspring population of the next generation according to the non-dominated level and the crowding degree; S27, check whether the termination iteration number is met, if the termination iteration number is not met, repeat S21 to S26, if the termination iteration number is met, output the initial population at this time; S3. Determine the position of the UAV according to the parameter prediction method in the scheduling method.

2. The method for swarm drone resource scheduling based on intelligent evolutionary algorithm according to claim 1 is characterized in that: The cluster UAV reconnaissance and tracking task processing model includes a monitoring module, a control module, an execution module and a timing module; A monitoring module, wherein the monitoring module monitors the dynamic change information of the drone, wherein the dynamic change information includes the task location and the task quantity, and the monitoring module sends the dynamic change information to the control module; when a recurrence occurs in the scene or a tracking task is performed, the monitoring module sends a timing parameter to the control module; A control module, wherein the control module receives dynamic change information and runs a cluster drone resource scheduling method, selects a corresponding static scheduling scheme and sends it to an execution module; the control module receives timing parameters and runs a cluster drone resource scheduling method, selects a corresponding dynamic scheduling scheme and sends it to an execution module; An execution module, the execution module receives and executes the static scheduling scheme and the dynamic scheduling scheme; A timing module records the system time, starts timing when the execution module starts executing and sends a time signal to the monitoring module.

3. The method for swarm drone resource scheduling based on intelligent evolutionary algorithm according to claim 1 is characterized in that: S21 includes the following steps: S211, dividing the code into two parts, the first part of the code represents the drone serial number and task execution order required to perform the ordinary task, and the second part of the code represents the drone serial number and task execution order required for the tracking task.

4. The method for swarm drone resource scheduling based on intelligent evolutionary algorithm according to claim 1 is characterized in that: S22 includes the following steps: S221, setting the termination condition, crossover parameter, mutation parameter and parameters required for population generation as parameters required for population preprocessing; S222, randomly encoding and generating a number of feasible solution individuals in the global scope, and selecting an individual with a Pareto dominance level of 1 from the several feasible solution individuals and putting it into an external archive; S223. When S222 does not satisfy the number of populations required to reach the initial population, S222 is repeated. When S222 satisfies the number of populations required to reach the initial population, the individuals in the external archive at this time are used as the initial population.

5. The method for swarm drone resource scheduling based on intelligent evolutionary algorithm according to claim 1 is characterized in that: S24 includes the following steps: S241, take out the non-dominated level of each individual in the parent population; S242, placing the individual with the best objective function value in each non-dominated level into the excellent set, and the remaining individuals into the ordinary set; S243, randomly select two individuals from the excellent set for comparison, retain the individual with the best objective function and compare it with individuals in other excellent sets for several times, and put the retained individual after the comparison into the final set; S244. Select the individual with the best objective function in the final set and cross-match it with the common set at the crossover probability to obtain a crossover population.

6. The method for swarm drone resource scheduling based on intelligent evolutionary algorithm according to claim 1 is characterized in that: A scheduling method in S2 includes: S28. adopt a no-fly zone avoidance method to enable the drone to avoid unreachable areas during flight; S281. Calculate the distances between the starting point A of the drone and the six vertices of the no-fly zone, record them as D1 to D6, and select the smallest distance D among D1 to D6. MIN ; S282, Comparison D MIN The length of the no-fly zone is D O , select the set of vertices C that can be directly connected to A A , the selection method is: if D MIN <D O , then the two vertices with the smallest distance A in D1~D6 are regarded as the directly connectable vertex set; if D MIN D O , then the three vertices with the smallest distance A in D1~D6 are the set of directly connected vertices; S283: Execute S282 for the end point A' where the drone crosses the no-fly zone, and obtain a set of directly connectable points C of A'. A’ ; S284, take C A , C A’ If the intersection is empty, the vertex of the no-fly zone is used as the transfer point to fly around the edge of the no-fly zone; S285, if C A , C A’ If the intersection of is not empty, then select a point with a small distance and fewer selected points in the intersection as a transfer point for detour.

7. The method for swarm drone resource scheduling based on intelligent evolutionary algorithm according to claim 1 is characterized in that: S3 includes the following steps: S31, determining the target space corresponding to each drone according to the static scheduling method in the cluster drone resource scheduling method; S32, dividing the flight route of the UAV into several stages according to the mission objectives in the target space; S33, according to the execution time of the scheduling plan, the flight speed of the drone and the speed-distance formula, in the route stage in S32, determine the stage the drone is currently in I ; S34, at the stage I The starting point is the center of the circle O , with drones flying over the stage I The distance from the starting point is the radius, and the equation of the circle is ; S35, connecting the above stages I The starting point and end point of the straight line are determined by the two points. , after solving the equation of the circle and the equation of the line and removing the meaningless solutions, the remaining solution is the predicted position of the drone.