A method, device and medium for planning multi-task collaborative reconnaissance of drone swarms

Abstract

This application relates to a method, device, and medium for planning multi-task collaborative reconnaissance for a swarm of unmanned aerial vehicles (UAVs). The method includes: utilizing a swarm of UAVs to conduct reconnaissance of a mission scenario; setting a benefit objective function based on the number of mission areas to be reconnaissanced in the mission scenario, the number of UAVs involved in decision-making, the reconnaissance value coefficient of the mission area, and the reconnaissance value coefficient of the mission area; setting a time objective function based on the number of mission areas actually reconnaissanced by the UAVs, the time available for the UAVs to transfer road sections, and the reconnaissance time allocated by the UAVs in the mission areas; constructing a multi-task collaborative reconnaissance model based on the benefit objective function, the time objective function, and pre-set constraints; and solving the multi-task collaborative reconnaissance model using genetic expression programming to obtain the UAVs' mission sequence and the reconnaissance time in each mission area. This method can improve the efficiency of multi-task reconnaissance.

Description

Technical Field

[0001] The present application relates to the field of drone reconnaissance technology, and in particular to a method, device, and medium for planning multi-task collaborative reconnaissance of a drone cluster. Background Art

[0002] With the increasing occurrence of natural and man-made disasters, such as earthquakes, nuclear leaks, and large-scale fires, these disasters not only cause irreparable property losses but also pose a serious threat to people's lives. With the recent development of drone technology and the rise of artificial intelligence, intelligent drone detection systems have gradually become a hot topic in drone research. Drones, due to their high cruising speed and low operating costs, are widely used in various environments. During land disaster relief and disaster area reconnaissance, drones are often used to conduct reconnaissance of specific areas, providing forward situational information to helicopters, ground vehicles, or personnel in the rear. Many related studies have proposed using drone swarms to perform multiple reconnaissance missions. Maximizing the benefits of drone reconnaissance in multiple mission areas primarily requires maximizing the total reconnaissance benefit (total intelligence gained) by minimizing the reconnaissance time of each mission area within a short period of time. The purpose of drone reconnaissance in a mission area is to obtain effective information and reduce uncertainty about the mission area. However, drone reconnaissance in a mission area generally operates in a complex and uncertain environment, and is also limited by the drone's flight time and the operating time of the reconnaissance payload it carries. This often makes it difficult to ensure complete information coverage of each mission area, resulting in low multi-mission reconnaissance efficiency. Summary of the Invention

[0003] Based on this, it is necessary to provide a UAV cluster multi-task collaborative reconnaissance planning method, equipment and medium that can improve the efficiency of multi-task reconnaissance in response to the above technical problems.

[0004] A multi-task collaborative reconnaissance planning method for a swarm of unmanned aerial vehicles (UAVs), the method comprising:

[0005] Obtain reconnaissance missions and mission scenarios; use a drone swarm to conduct reconnaissance of the mission scenario, and set a benefit objective function based on the number of mission areas to be reconnaissanced in the mission scenario, the number of drones involved in decision-making, the reconnaissance value coefficient of the mission area, and the reconnaissance value coefficient of the mission area; set a time objective function based on the number of mission areas actually reconnaissanced by the drones, the time the drones can use to transfer sections, and the reconnaissance time allocated by the drones in the mission areas;

[0006] Construct a multi-task collaborative reconnaissance model based on the benefit objective function, time objective function and pre-set constraints;

[0007] The multi-task collaborative reconnaissance model is solved according to gene expression programming. The module library is designed according to the number of mission areas to be reconnaissanced. The characters in each module library are connected into a row at the root node of the expression tree to generate the initial solution individual. The benefit objective function and the time objective function are normalized and set as the fitness evaluation function. The initial solution individual is evolved according to different evolutionary operations through pre-set evolutionary constraints to obtain the chromosome of the UAV's action sequence, that is, the UAV's task sequence and the reconnaissance time in each mission area.

[0008] In one embodiment, a multi-task collaborative reconnaissance model is constructed based on a benefit objective function, a time objective function, and preset constraints, including:

[0009] According to the benefit objective function, time objective function and pre-set constraints, a multi-task collaborative reconnaissance model is constructed as follows:

[0010]

[0011]

[0012]

[0013] Among them, μ1 and μ2 are adjustable parameters, H max represents the maximum reconnaissance benefit of multiple UAVs, T min represents the minimum reconnaissance time used by multiple drones as a whole, B represents the number of mission areas to be reconnaissanced in the mission scenario, D represents the number of drones participating in the decision-making in the mission scenario, Z represents the number of mission areas actually reconnaissanced by the drones, and the corresponding transfer section of the drone is Z+1. represents the time that drone d can be used to transfer road segment f, represents the reconnaissance time allocated to the d-th UAV in the i-th mission area, i∈M,l i represents the reconnaissance value coefficient of the i-th mission area, i∈M,Q i Indicates the area size of the i-th task area, i∈M, o i represents the effective scanning width of the UAV in the mission area i, v represents the flight speed of the UAV when performing the mission, H imin represents the minimum reconnaissance benefit that must be achieved when reconnaissance is conducted on the i-th mission area, E represents the path of the UAV, E={<i,j>|i,j∈M,i≠j}, d ij Represents the distance between two task areas, x ij The value is 1 or 0, 1 means the drone has moved from mission area i to mission area j, 0 means the drone has not moved from mission area i to mission area j, x i The horizontal coordinate of the absolute geographical location of mission area i, x jThe horizontal coordinate of the absolute geographical location of mission area j, y i The ordinate of the absolute geographical location of mission area i, y j The vertical coordinate representing the absolute geographical location of mission area j.

[0014] In one embodiment, the module library is designed using the number of task areas to be scouted, including: using the number s of task areas to be scouted to design the number of task area elements in the module library, and in the expression string of the genetic programming operation, each string cannot have more than s task area characters.

[0015] In one embodiment, the initial solution individuals are decision sequences of drones; the drone decision sequences include parameter sizes and sequence arrangements; the parameter sizes represent the reconnaissance time of each mission area; and the sequence arrangements represent the mission sequence of the drones. The benefit objective function and the time objective function are normalized and set as a fitness evaluation function, and the initial solution individuals are evolved according to different evolutionary operations using pre-set evolutionary constraints to obtain chromosomes of the drone action sequences, including:

[0016] Set the evolutionary generation of the initial solution individuals, and divide the evolutionary cycles according to the evolutionary generation. If an evolutionary cycle includes n generations, the first r generations are given initial parameters, and the order is optimized according to the evolutionary constraints and different evolutionary operations. The fitness of the optimized multiple order arrangements is evaluated according to the fitness evaluation function, and the optimal arrangement scheme with the highest fitness is retained. The parameter size of the next nr generations under the optimal arrangement scheme is optimized according to the evolutionary constraints and different evolutionary operations to obtain the optimal solution within a cycle. The optimal solutions within multiple evolutionary cycles are compared, and the optimal solution with the highest fitness is taken as the final solution, that is, the chromosome of the drone's action sequence.

[0017] In one embodiment, the evolutionary constraints include:

[0018] (1) The number of mission areas reconnaissance by each UAV is s, plus the round trip to the base, the encoding length of each string cannot exceed s+2;

[0019] (2) The reconnaissance time of each UAV in its mission area plus the transfer time between mission areas and between mission areas and bases cannot exceed the UAV's flight time;

[0020] (3) In the coding sequence of the UAV, the first and last codes must be the base, and there cannot be duplication of mission area codes;

[0021] (4) The maximum reconnaissance time of the UAV in each mission area cannot exceed the preset maximum reconnaissance time t max , the UAV's reconnaissance benefit in each mission area cannot be less than H imin ;H iminIt represents the minimum reconnaissance benefit that must be achieved when reconnaissance is conducted on the i-th mission area;

[0022] (5) All mission areas have drone reconnaissance, that is, the expression strings of all drones must contain the codes of all module libraries;

[0023] (6) If the reconnaissance benefit of a mission area exceeds a certain value, the drone will definitely leave. The certain value is related to the number of drones that have historically reconnaissanced the area and the drone reconnaissance time;

[0024] (7) Evolutionary algebraic restrictions.

[0025] In one embodiment, the evolutionary operations include selection, mutation, insertion, symmetry, local inversion and crossover recombination; selection means that before the start of each round, the two representative chromosomes with the highest fitness values ​​in the current evolutionary cycle are selected as excellent parents, which serve as the basis for subsequent evolutionary development; mutation acts on a single chromosome, and any gene on the chromosome is selected with a certain probability, and the code of the bit is regenerated when the mutation probability is met; insertion means randomly selecting a letter representing a mission area from the mission area list and inserting it into a randomly specified position of the chromosome of the drone mission area sequence; symmetry means randomly generating two numbers less than the length of the operable chromosome whose difference is an odd number not equal to 1 and symmetric about the center of the chromosome; inversion means randomly generating two numbers less than the length of the operable chromosome, indexing the chromosome segments corresponding to the two numbers on the chromosome, and performing character inversion; the crossover recombination targets two chromosomes, and is divided into one-point recombination and two-point recombination according to the number of chromosome crossover points. If one-point recombination occurs, the two chromosomes exchange the genes behind the recombination point. If it is a two-point recombination, the gene strings in the middle of the two points are exchanged, and the crossover points are randomly selected.

[0026] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0027] Obtain reconnaissance missions and mission scenarios; use a drone swarm to conduct reconnaissance of the mission scenario, and set a benefit objective function based on the number of mission areas to be reconnaissanced in the mission scenario, the number of drones involved in decision-making, the reconnaissance value coefficient of the mission area, and the reconnaissance value coefficient of the mission area; set a time objective function based on the number of mission areas actually reconnaissanced by the drones, the time the drones can use to transfer sections, and the reconnaissance time allocated by the drones in the mission areas;

[0028] Construct a multi-task collaborative reconnaissance model based on the benefit objective function, time objective function and pre-set constraints;

[0029] The multi-task collaborative reconnaissance model is solved according to gene expression programming. The module library is designed according to the number of mission areas to be reconnaissanced. The characters in each module library are connected into a row at the root node of the expression tree to generate the initial solution individual. The benefit objective function and the time objective function are normalized and set as the fitness evaluation function. The initial solution individual is evolved according to different evolutionary operations through pre-set evolutionary constraints to obtain the chromosome of the UAV's action sequence, that is, the UAV's task sequence and the reconnaissance time in each mission area.

[0030] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0031] Obtain reconnaissance missions and mission scenarios; use a drone swarm to conduct reconnaissance of the mission scenario, and set a benefit objective function based on the number of mission areas to be reconnaissanced in the mission scenario, the number of drones involved in decision-making, the reconnaissance value coefficient of the mission area, and the reconnaissance value coefficient of the mission area; set a time objective function based on the number of mission areas actually reconnaissanced by the drones, the time the drones can use to transfer sections, and the reconnaissance time allocated by the drones in the mission areas;

[0032] Construct a multi-task collaborative reconnaissance model based on the benefit objective function, time objective function and pre-set constraints;

[0033] The multi-task collaborative reconnaissance model is solved according to gene expression programming. The module library is designed according to the number of mission areas to be reconnaissanced. The characters in each module library are connected into a row at the root node of the expression tree to generate the initial solution individual. The benefit objective function and the time objective function are normalized and set as the fitness evaluation function. The initial solution individual is evolved according to different evolutionary operations through pre-set evolutionary constraints to obtain the chromosome of the UAV's action sequence, that is, the UAV's task sequence and the reconnaissance time in each mission area.

[0034] The above-mentioned method, equipment and medium for multi-task collaborative reconnaissance planning of drone clusters, this application uses drone clusters to scout mission scenarios, and sets a benefit objective function based on the number of mission areas to be surveyed in the mission scenario, the number of drones participating in the decision-making, the reconnaissance value coefficient of the mission area, and the reconnaissance value coefficient of the mission area; uses the number of mission areas actually surveyed by the drone, the time the drone can use to transfer sections, and the reconnaissance time allocated by the drone in the mission area to set a time objective function, and then constructs a multi-task collaborative reconnaissance model based on the benefit objective function and the time objective function and pre-set constraints, which can convert the multi-task collaborative reconnaissance problem into a multi-objective optimization problem, and by setting constraints, the drone's reconnaissance mission can be optimized, which can ensure that the benefit is maximized in the shortest time, thereby improving the drone reconnaissance efficiency, and finally, uses genetic expression programming to solve the multi-task collaborative reconnaissance model, and continuously iterates and evolves the solution results through the adaptive evolution method, thereby improving the accuracy of the model solution and obtaining the optimal multi-task reconnaissance solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A flowchart of a multi-task collaborative reconnaissance planning method for a UAV swarm in one embodiment is shown;

[0036] Figure 2 is a schematic diagram of a portion of the evolution operation in one embodiment;

[0037] Figure 3 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0039] In one embodiment, Figure 1 As shown, a multi-task collaborative reconnaissance planning method for a UAV swarm is provided, comprising the following steps:

[0040] Step 102, obtain the reconnaissance mission and mission scenario; use the drone cluster to scout the mission scenario, and set the benefit objective function according to the number of mission areas to be surveyed in the mission scenario, the number of drones participating in the decision-making, the reconnaissance value coefficient of the mission area, and the reconnaissance value coefficient of the mission area; set the time objective function using the number of mission areas actually surveyed by the drone, the time the drone can use to transfer the road section, and the reconnaissance time allocated by the drone in the mission area.

[0041] In a mission scenario, a swarm of drones can be selected to jointly reconnoiter a target, or several drones can be assigned to reconnoiter a mission area. After a certain level of reconnaissance, some drones will depart to assist other drones in reconnoitering other mission areas. The overall objective of multi-drone collaborative reconnaissance is to maximize the reconnaissance benefit of the swarm in the shortest possible time. This is a multi-objective optimization problem. The total time to complete a mission is affected by the decisions of the entire swarm. When a single drone completes its mission and returns to base, it wastes reconnaissance resources to a certain extent. Therefore, it is necessary to rationally schedule and allocate drone resources to minimize the time it takes for the swarm to complete the mission. To reduce computational complexity, the drones can be assumed to travel to and from the mission area and base at a constant speed. A cost-benefit objective function is defined by the number of mission areas to be reconnoitred in the mission scenario, the number of drones involved in decision-making, the reconnaissance value coefficient of the mission area, and the reconnaissance value coefficient of the mission area. A time objective function is defined by the number of mission areas actually reconnoitred by the drones, the time available for the drones to transfer to another area, and the reconnaissance time allocated by the drones in the mission area. The cost-benefit objective function and the time objective function are used to construct the objective function of the multi-task collaborative reconnaissance model. The final solution should be a spatiotemporal sequence of each drone's reconnaissance missions.

[0042] Step 104: construct a multi-task collaborative reconnaissance model based on the benefit objective function, the time objective function, and pre-set constraints.

[0043] A drone swarm (multiple drone swarms) starts from different bases to perform reconnaissance missions. The decision is related to the distance between the drone and each mission area. If we want to achieve the best possible result, we need to set multiple constraints in the solution process. The constraints include: (1) The drone’s flight time is certain, assuming it is T, and the time for drone d to perform the mission in the reconnaissance area i is , assuming that the drone d traverses Z mission areas, the corresponding drone transfer section is Z+1, and the total time to traverse these mission areas is The time it takes for the drone to travel to and from the mission area and the base is The constraints are expressed as follows:

[0044]

[0045]

[0046]

[0047] Where E represents the path of the UAV, E={<i,j>|i,j∈M,i≠j}, d ij Indicates the distance between two mission areas.

[0048] (2) Minimum reconnaissance benefit constraint for each UAV:

[0049]

[0050] (3) In the value coefficient constraint of the task area, the value coefficient of each task area should satisfy the following formula:

[0051]

[0052] (4) All mission areas should be traversed by drones:

[0053]

[0054] (5) Each UAV needs to depart from its base and return to its base after completing all mission areas:

[0055]

[0056] By setting constraints when solving the problem, the accuracy of the solution of the multi-task collaborative reconnaissance model can be improved.

[0057] Step 106, solve the multi-task collaborative reconnaissance model according to gene expression programming, design a module library based on the number of mission areas to be reconnaissanced, connect the characters in each module library into a row at the root node of the expression tree to generate an initial solution individual, normalize the benefit objective function and the time objective function and set them as the fitness evaluation function, and evolve the initial solution individual according to different evolutionary operations through pre-set evolutionary constraints to obtain the chromosome of the drone's action sequence, that is, the drone's mission sequence and the reconnaissance time in each mission area.

[0058] The gene expression module library is encoded according to the actual task, as shown in Table 1;

[0059] Table 1

[0060] name coding Mission Area 1 A Mission Area 2 B Mission Area 3 C Mission Area 4 D Mission Area 5 E Mission Area 6 F Mission Area 7 G Mission Area 8 H base I

[0061] The sequence of these code combinations represents the order in which the drones' reconnaissance missions are executed. Furthermore, the drones' reconnaissance time in different mission areas is used as a parameter for each string. Based on the model constraints, a parameter range for these reconnaissance times is established. Furthermore, during the genetic expression programming process for solving the multi-task collaborative reconnaissance model, a rule base is also established. This rule base contains the behavioral rules for multiple drones, and the code for the module library can be populated within the rule base. This rule base is very simple; it simply concatenates the characters from each module library into a single line at the root node of the expression tree to generate an initial solution, namely the drones' initial mission sequence and the reconnaissance time in each mission area during the mission.

[0062] When solving a multi-task collaborative reconnaissance model using gene expression programming, the objective function consists of a benefit objective function and a time objective function. The influencing factors of these two objective functions must be balanced. By setting two adjustable parameters, the weight of one function on the evolved individuals is neither too large nor too small. The two objective functions are normalized according to their ranges. Based on the distribution of fitness values ​​and the desired evolutionary direction, these objective functions are normalized and superimposed to serve as the fitness evaluation function used to evaluate the model solution during the evolutionary process. Evolutionary constraints are then set to improve evolutionary efficiency and accuracy, preventing unrealistic solutions.

[0063] In the above-mentioned multi-task collaborative reconnaissance planning method for drone clusters, this application uses drone clusters to scout mission scenarios, and sets a benefit objective function based on the number of mission areas to be surveyed in the mission scenario, the number of drones participating in decision-making, the reconnaissance value coefficient of the mission area, and the reconnaissance value coefficient of the mission area; uses the number of mission areas actually surveyed by drones, the time drones can use to transfer sections, and the reconnaissance time allocated by drones in mission areas to set a time objective function, and then constructs a multi-task collaborative reconnaissance model based on the benefit objective function and the time objective function and pre-set constraints. It can convert the multi-task collaborative reconnaissance problem into a multi-objective optimization problem, and by setting constraints, it can optimize the reconnaissance mission of the drone, which can ensure that the benefit is maximized in the shortest time, thereby improving the drone reconnaissance efficiency. Finally, the multi-task collaborative reconnaissance model is solved by using genetic expression programming, and the solution results are continuously iteratively evolved by the adaptive evolution method, thereby improving the accuracy of the model solution and obtaining the optimal multi-task reconnaissance solution.

[0064] In one embodiment, a multi-task collaborative reconnaissance model is constructed based on a benefit objective function, a time objective function, and preset constraints, including:

[0065] According to the benefit objective function, time objective function and pre-set constraints, a multi-task collaborative reconnaissance model is constructed as follows:

[0066]

[0067]

[0068]

[0069] Among them, μ1 and μ2 are adjustable parameters, H max represents the maximum reconnaissance benefit of multiple UAVs, T minrepresents the minimum reconnaissance time used by multiple drones as a whole, B represents the number of mission areas to be reconnaissanced in the mission scenario, D represents the number of drones participating in the decision-making in the mission scenario, Z represents the number of mission areas actually reconnaissanced by the drones, and the corresponding transfer section of the drone is Z+1. represents the time that drone d can be used to transfer road segment f, represents the reconnaissance time allocated to the d-th UAV in the i-th mission area, i∈M,l i represents the reconnaissance value coefficient of the i-th mission area, i∈M,Q i Indicates the area size of the i-th task area, i∈M, o i represents the effective scanning width of the UAV in the mission area i, v represents the flight speed of the UAV when performing the mission, H imin represents the minimum reconnaissance benefit that must be achieved when reconnaissance is conducted on the i-th mission area, E represents the path of the UAV, E={<i,j>|i,j∈M,i≠j}, d ij Represents the distance between two task areas, x ij The value is 1 or 0, 1 means the drone has moved from mission area i to mission area j, 0 means the drone has not moved from mission area i to mission area j, x i The horizontal coordinate of the absolute geographical location of mission area i, x j The horizontal coordinate of the absolute geographical location of mission area j, y i The ordinate of the absolute geographical location of mission area i, y j The vertical coordinate representing the absolute geographical location of mission area j.

[0070] In one embodiment, the module library is designed using the number of task areas to be scouted, including: using the number s of task areas to be scouted to design the number of task area elements in the module library, and in the expression string of the genetic programming operation, each string cannot have more than s task area characters.

[0071] In a specific embodiment, multiple drones are required to select mission areas, and each drone may have multiple mission areas. First, consider some possible limitations of multiple drones reconnaissance of each mission area. For example, a drone may not traverse all mission areas. Assuming there are 8 mission areas, and assuming that a drone can traverse at most s mission areas, when assembling the mission areas in the module library through gene expression programming, a chromosome contains at most s mission area elements in the module library. Therefore, in the expression string of the genetic programming operation, each string cannot contain more than s mission area characters. As shown in Table 1, the gene expression module library is encoded according to the actual task.

[0072] In one embodiment, the initial solution individuals are decision sequences of drones; the drone decision sequences include parameter sizes and sequence arrangements; the parameter sizes represent the reconnaissance time of each mission area; and the sequence arrangements represent the mission sequence of the drones. The benefit objective function and the time objective function are normalized and set as a fitness evaluation function, and the initial solution individuals are evolved according to different evolutionary operations using pre-set evolutionary constraints to obtain chromosomes of the drone action sequences, including:

[0073] Set the evolutionary generation of the initial solution individuals, and divide the evolutionary cycles according to the evolutionary generation. If an evolutionary cycle includes n generations, the first r generations are given initial parameters, and the order is optimized according to the evolutionary constraints and different evolutionary operations. The fitness of the optimized multiple order arrangements is evaluated according to the fitness evaluation function, and the optimal arrangement scheme with the highest fitness is retained. The parameter size of the next nr generations under the optimal arrangement scheme is optimized according to the evolutionary constraints and different evolutionary operations to obtain the optimal solution within a cycle. The optimal solutions within multiple evolutionary cycles are compared, and the optimal solution with the highest fitness is taken as the final solution, that is, the chromosome of the drone's action sequence.

[0074] In a specific embodiment, the multi-task reconnaissance model involves multiple drones. The decisions made by one drone not only affect its own fitness evaluation, but also those of other drones. Therefore, collaborative evolution of the drone swarm is required. After initialization, each drone has its own decision sequence. Each drone utilizes this decision-making mechanism to participate in the reconnaissance mission. After the mission is completed, the sum of the fitness function values ​​of all drones is calculated as the reconnaissance benefit. The drone's decision sequence includes parameter size (reconnaissance time for each mission area) and sequence (the order in which the drones are reconnaissance), encompassing both continuous and discrete variables. In this case, the drones first optimize the sequence, then optimize the parameters based on this optimization. This alternating process ensures efficient sequence optimization and improves the accuracy of the evaluation through a global assessment of reconnaissance time. A total of 6,000 generations are considered. Each evolutionary cycle consists of 200 generations. In each evolutionary cycle, the sequence optimization is performed for the first 150 generations, given the given initial parameters. After fitness evaluation, the optimal sequence is retained (each drone's decision sequence). The parameters under this sequence are optimized for the next 50 generations to obtain the optimal solution within one cycle. The optimal solutions within multiple evolutionary cycles are compared, and the optimal solution with the highest fitness is taken as the final solution, that is, the chromosome of the drone's action sequence.

[0075] In one embodiment, the evolutionary constraints include:

[0076] (1) The number of mission areas reconnaissance by each UAV is s, plus the round trip to the base, the encoding length of each string cannot exceed s+2;

[0077] (2) The reconnaissance time of each UAV in its mission area plus the transfer time between mission areas and between mission areas and bases cannot exceed the UAV's flight time;

[0078] (3) In the coding sequence of the UAV, the first and last codes must be base I, and there must be no duplication of mission area codes;

[0079] (4) The maximum reconnaissance time of the UAV in each mission area cannot exceed the preset maximum reconnaissance time t max , the UAV's reconnaissance benefit in each mission area cannot be less than H imin ;H imin It represents the minimum reconnaissance benefit that must be achieved when reconnaissance is conducted on the i-th mission area;

[0080] (5) All mission areas have drone reconnaissance, that is, the expression strings of all drones must contain the codes of all module libraries;

[0081] (6) If the reconnaissance benefit of a mission area exceeds a certain value, the drone will definitely leave. The certain value is related to the number of drones that have historically reconnaissanced the area and the drone reconnaissance time;

[0082] (7) Evolutionary algebraic restrictions.

[0083] In one embodiment, the evolutionary operations include selection, mutation, insertion, symmetry, local inversion and crossover recombination; selection means that before the start of each round, the two representative chromosomes with the highest fitness values ​​in the current evolutionary cycle are selected as excellent parents, which serve as the basis for subsequent evolutionary development; mutation acts on a single chromosome, and any gene on the chromosome is selected with a certain probability, and the code of the bit is regenerated when the mutation probability is met; insertion means randomly selecting a letter representing a mission area from the mission area list and inserting it into a randomly specified position of the chromosome of the drone mission area sequence; symmetry means randomly generating two numbers less than the length of the operable chromosome whose difference is an odd number not equal to 1 and symmetric about the center of the chromosome; inversion means randomly generating two numbers less than the length of the operable chromosome, indexing the chromosome segments corresponding to the two numbers on the chromosome, and performing character inversion; the crossover recombination targets two chromosomes, and is divided into one-point recombination and two-point recombination according to the number of chromosome crossover points. If one-point recombination occurs, the two chromosomes exchange the genes behind the recombination point. If it is a two-point recombination, the gene strings in the middle of the two points are exchanged, and the crossover points are randomly selected.

[0084] In a specific embodiment, Figure 2 As shown, it includes insertion, mutation, symmetry, local inversion and crossover recombination. The specific operation process is as follows:

[0085] (1) Insertion: Randomly select a letter representing a mission area from the mission area list and insert it into a randomly specified position of the chromosome of the drone mission area sequence. Figure 2 As shown in Figure 1, after insertion, the UAV's mission sequence changes from "IBFDCI" to "IBFDHCI". Note that if the insertion operation produces duplicate letters (violating the non-repetition constraint of the UAV mission area), reinsert them.

[0086] (2) Mutation: Mutation acts on a single chromosome, selecting a gene on the chromosome with a certain probability, and regenerating the code of the bit when the mutation probability is met. Figure 2 As shown in Figure 2, after mutation, the UAV's mission sequence changes from "IBFDCI" to "IBFECI". Note that if the mutation operation produces repeated letters (violating the non-repetitive constraint of the UAV mission area), the mutation is repeated.

[0087] (3) Symmetry, that is, to perform local symmetry operation on the chromosome of the UAV mission sequence; randomly generate two numbers smaller than the length of the operable chromosome (the head and tail are fixed and cannot be operated) with a difference of an odd number not equal to 1, and perform symmetry on the chromosome about its center. Figure 2 As shown in the figure, after symmetry, the UAV's task sequence changes from "IBFDCI" to "IBCDFI".

[0088] (4) Local inversion, that is, the chromosome of the UAV mission sequence is locally inverted; two numbers that are less than the length of the operable chromosome (the head and tail are fixed and cannot be operated) are randomly generated, and the difference between them is not 1. The chromosome segments corresponding to the two numbers are indexed on the chromosome to perform character inversion. Figure 2 As shown in the figure, after symmetry, the UAV's mission sequence changes from "IBFDCI" to "ICDFBI".

[0089] (5) Crossover recombination. Randomly select a chromosome from the drone memory pool list. As shown in the figure, randomly generate two numbers that are smaller than the length of the operable chromosome (the head and tail are fixed and inoperable) and are not equal and the difference is not 1. Index the chromosome segments corresponding to the two numbers on the chromosome. The chromosomes from the memory pool will be exchanged to the chromosomes for evolution according to the two numbers. Figure 2 As shown in Figure 1, the UAV’s mission sequence changes from “IBFDCI” to “IBFHAI”. Note that if the cross-recombination operation produces repeated letters (violating the non-repetitive constraint of the UAV mission area), the cross-recombination is repeated.

[0090] It should be understood that although Figure 1The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. In addition, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0091] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 3 As shown. The computer device includes a processor, memory, network interface, display screen and input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for multi-task collaborative reconnaissance planning of a drone cluster is implemented. The display screen of the computer device can be a liquid crystal display or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse.

[0092] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0093] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0094] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0095] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A multi-task collaborative reconnaissance planning method for UAV swarms, characterized by: The method comprises: Obtain a reconnaissance mission and mission scenario; use a drone swarm to conduct reconnaissance on the mission scenario, and set a benefit objective function based on the number of mission areas to be surveyed in the mission scenario, the number of drones involved in decision-making, the reconnaissance value coefficient of the mission area, and the reconnaissance value coefficient of the mission area; set a time objective function based on the number of mission areas actually surveyed by the drones, the time the drones can use to transfer sections, and the reconnaissance time allocated by the drones in the mission areas; Constructing a multi-task collaborative reconnaissance model according to the benefit objective function, the time objective function and pre-set constraints; The multi-task collaborative reconnaissance model is solved according to gene expression programming, a module library is designed using the number of mission areas to be reconnaissanced, characters in each module library are connected into a row at the root node of the expression tree to generate an initial solution individual, the benefit objective function and the time objective function are normalized and set as a fitness evaluation function, and the initial solution individual is evolved according to different evolutionary operations using pre-set evolutionary constraints to obtain the chromosome of the drone's action sequence, that is, the drone's mission sequence and the reconnaissance time in each mission area; Constructing a multi-task collaborative reconnaissance model according to the benefit objective function, the time objective function, and pre-set constraints includes: A multi-task collaborative reconnaissance model is constructed based on the benefit objective function, the time objective function and the preset constraints: Among them, μ1 and μ2 are adjustable parameters, H max represents the maximum reconnaissance benefit of multiple UAVs, T min represents the minimum reconnaissance time used by multiple drones as a whole, B represents the number of mission areas to be reconnaissanced in the mission scenario, D represents the number of drones participating in the decision-making in the mission scenario, Z represents the number of mission areas actually reconnaissanced by the drones, and the corresponding transfer section of the drone is Z+1. represents the time that drone d can be used to transfer road segment f, represents the reconnaissance time allocated to the d-th UAV in the i-th mission area, i∈M,l i represents the reconnaissance value coefficient of the i-th mission area, i∈M,Q i Indicates the area size of the i-th task area, i∈M, o i represents the effective scanning width of the UAV in the mission area i, v represents the flight speed of the UAV when performing the mission, H imin represents the minimum reconnaissance benefit that must be achieved when reconnaissance is conducted on the i-th mission area, E represents the path of the UAV, E={<i,j>|i,j∈M,i≠j}, d ij Represents the distance between two task areas, x ij The value is 1 or 0, 1 means the UAV has moved from mission area i to mission area j, 0 means the UAV has not moved from mission area i to mission area j, x i The horizontal coordinate of the absolute geographical location of mission area i, x j The horizontal coordinate of the absolute geographical location of mission area j, y i The ordinate of the absolute geographical location of mission area i, y j represents the ordinate of the absolute geographical location of mission area j, and T represents the flight time of the UAV.

2. The method according to claim 1, characterized in that The module library is designed based on the number of mission areas to be surveyed, including: The number of task area elements in the module library is designed using the number s of task areas to be detected. In the expression string of the genetic programming operation, each string cannot have more than s task area characters.

3. The method according to claim 1, characterized in that The initial solution individual is a decision sequence of the UAV; the UAV decision sequence includes parameter size and sequence arrangement; the parameter size represents the reconnaissance time of each mission area; the sequence arrangement represents the mission sequence of the UAV; the benefit objective function and the time objective function are normalized and set as a fitness evaluation function, and the initial solution individual is evolved according to different evolutionary operations through pre-set evolutionary constraints to obtain a chromosome of the UAV action sequence, including: The evolutionary generation of the initial solution individuals is set, and multiple evolutionary cycles are divided according to the evolutionary generation. If an evolutionary cycle includes n generations, the initial parameters are given for the first r generations, and the order is optimized according to the evolutionary constraints and different evolutionary operations. The fitness of the optimized multiple order arrangements is evaluated according to the fitness evaluation function, and the optimal arrangement scheme with the highest fitness is retained. The parameter sizes of the subsequent nr generations under the optimal arrangement scheme are optimized according to the evolutionary constraints and different evolutionary operations to obtain the optimal solution within a cycle. The optimal solutions within multiple evolutionary cycles are compared, and the optimal solution with the highest fitness is taken as the final solution, that is, the chromosome of the drone's action sequence.

4. The method according to claim 3, characterized in that The evolutionary constraints include: (1) The number of mission areas reconnaissance by each UAV is s, plus the round trip to the base, the encoding length of each string cannot exceed s+2; (2) The reconnaissance time of each UAV in its mission area plus the transfer time between mission areas and between mission areas and bases cannot exceed the UAV's flight time; (3) In the coding sequence of the UAV, the first and last codes must be the base, and there cannot be duplication of mission area codes; (4) The maximum reconnaissance time of the UAV in each mission area cannot exceed the preset maximum reconnaissance time t max , the UAV's reconnaissance benefit in each mission area cannot be less than H imin ;H imin It represents the minimum reconnaissance benefit that must be achieved when reconnaissance is conducted on the i-th mission area; (5) All mission areas have drone reconnaissance, that is, the expression strings of all drones must contain the codes of all module libraries; (6) If the reconnaissance benefit of a mission area exceeds a certain value, the drone will definitely leave. The certain value is related to the number of drones that have historically reconnaissanced the mission area and the drone reconnaissance time; (7) Evolutionary algebraic restrictions.

5. The method according to claim 1, wherein The evolutionary operations include selection, mutation, insertion, symmetry, local inversion and crossover recombination; the selection means that before the start of each round, the two representative chromosomes with the highest fitness values ​​in the current evolutionary cycle are selected as excellent parents, which serve as the basis for subsequent evolution and development; the mutation acts on a single chromosome, selects any gene on the chromosome with a certain probability, and regenerates the code of the bit when the mutation probability is met; the insertion means randomly selecting a letter representing a mission area from the mission area list and inserting it into a randomly specified position of the chromosome of the drone mission area sequence; the symmetry means randomly generating two numbers less than the length of the operable chromosome whose difference is an odd number not equal to 1 and symmetric about its center on the chromosome; the local inversion means randomly generating two numbers less than the length of the operable chromosome, indexing the chromosome segments corresponding to the two numbers on the chromosome, and performing character inversion; the crossover recombination targets two chromosomes, and is divided into one-point recombination and two-point recombination according to the number of chromosome crossover points. If one-point recombination occurs, the two chromosomes exchange the genes behind the recombination point. If it is a two-point recombination, the gene strings in the middle of the two points are exchanged, and the crossover points are randomly selected.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images