A heterogeneous unmanned aerial vehicle cluster task allocation method and system

Abstract

The application discloses a heterogeneous unmanned aerial vehicle cluster task allocation method and system, relates to unmanned aerial vehicle cluster scheduling technology, and comprises the following steps: establishing a heterogeneous unmanned aerial vehicle cluster task allocation model; performing task demand quantification and cluster scale estimation, and determining the maximum number of each type of unmanned aerial vehicle; and based on an improved genetic algorithm, a task sequence is solved with the minimum total number of scheduled unmanned aerial vehicles as a target. The improved genetic algorithm comprises the following steps: taking the reciprocal of the sum of non-zero genes of a chromosome as a fitness function; adopting real number coding; the mutation operation comprises a self-checking and repairing step based on a constraint order repair, checking in the order of a single-machine maximum time constraint, a task demand satisfaction constraint and a total load constraint, and actively repairing when a violation occurs, such as transferring an over-limit task to an idle unmanned aerial vehicle of the same type. The application significantly improves solving efficiency and reduces unmanned aerial vehicle use cost by pre-estimating to narrow the search space and combining an active repair mechanism.

Description

Technical Field

[0001] This application relates to the field of unmanned aerial vehicle (UAV) swarm scheduling technology, and in particular to a method and system for assigning tasks to heterogeneous UAV swarms. Background Technology

[0002] Unmanned aerial vehicles (UAVs) should typically operate in swarms rather than individually. UAV swarm resource allocation and task assignment are key technologies in UAV swarm warfare research, representing a complex and cutting-edge field involving multiple disciplines. To ensure UAV swarms can efficiently execute various tasks, effective task allocation is necessary based on the characteristics of the UAVs and the performance of their payloads. Finding the optimal task allocation solution while meeting mission requirements is crucial to fully utilize the capabilities of each UAV. Effectively utilizing UAVs with varying capabilities is essential. From a theoretical research perspective, UAV swarm task allocation is one of the major challenges in the field of UAV swarm cooperative systems research. UAV swarm task allocation is a key technology in UAV swarm mission planning and has become a research hotspot in industry.

[0003] The challenge in researching task allocation for UAV swarms lies in the large scale of heterogeneous UAV swarms, which causes the computational scale of task allocation to grow exponentially. Summary of the Invention

[0004] This application provides a heterogeneous UAV swarm task allocation method and system. For two types of coupled tasks, a function related to the completion percentage of task A and the completion probability of task B is established and quickly solved to achieve a reasonable allocation of resources required to complete tasks A and B.

[0005] This application provides a method for task allocation in a heterogeneous UAV swarm, used to generate task sequences for a swarm of UAVs with multiple task capabilities, including the following steps: A heterogeneous task allocation model for a heterogeneous drone swarm is established; in the heterogeneous task allocation model, the drone swarm includes at least: a first type of drone dedicated to performing a first task, a second type of drone integrating the capabilities of the first and second tasks, and a third type of drone dedicated to performing the second task. Based on the total time required to execute the first task at all task points and the maximum time for a single drone to execute the first task, determine the maximum total number of the first type of drones and the second type of drones required; based on the total payload required to execute the second task at all task points and the maximum payload for a single drone to execute the second task, determine the maximum total number of the second type of drones and the third type of drones required. Based on an improved genetic algorithm, the heterogeneous task allocation model is solved with the goal of minimizing the total number of scheduled drones to obtain the task sequence of the scheduled drones. The improved genetic algorithm includes: fitness function construction: the reciprocal of the sum of the non-zero real numbers of all gene positions on the chromosome is used as the fitness function so that the chromosome individual with a smaller total number of scheduled drones has a higher fitness.

[0006] This application provides a heterogeneous drone swarm task allocation system, including a processor and a memory. The memory stores a computer program, which, when executed by the processor, implements the steps of the heterogeneous drone swarm task allocation method described above.

[0007] This application proposes a resource allocation method for heterogeneous UAV swarm operations, addressing the resource allocation problem between two coupled tasks (task A being considered a prerequisite task by task B). The method not only establishes a function related to the completion percentage of task A and the completion probability of task B, strengthening the coupling between tasks A and B, but also provides a fast solution to quickly determine the optimal allocation of resources needed to complete tasks A and B.

[0008] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0009] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a flowchart illustrating the heterogeneous UAV swarm task allocation method according to an embodiment of this application; Figure 2 This is a schematic diagram of the genetic algorithm encoding for the heterogeneous UAV cluster task allocation method in an embodiment of this application; Figure 3 This is a schematic diagram of the cross-operation of the heterogeneous UAV swarm task allocation method in the embodiments of this application; Figure 4 This is a schematic diagram of the mutation operation of the heterogeneous UAV cluster task allocation method in the embodiments of this application. Detailed Implementation

[0010] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0011] This application provides a method for task allocation in a heterogeneous UAV swarm, used to generate task sequences for a swarm of UAVs with multiple task capabilities, including the following steps: In step S101, a heterogeneous task allocation model for a heterogeneous drone cluster is established; in the heterogeneous task allocation model, the drone cluster includes at least: a first type of drone dedicated to performing a first task, a second type of drone integrating the capabilities of the first and second tasks, and a third type of drone dedicated to performing the second task.

[0012] In the task allocation model, this embodiment of the application pre-allocates tasks in stage two based on the resources required to execute task A and the amount of resources required to complete task B at each known target determined in stage one. This pre-allocates tasks to determine the heterogeneous UAV cluster composition and the task sequence of tasks to be executed by the UAVs performing the tasks.

[0013] In step S102, based on the total time required for all task points to execute the first task (Task A) and the maximum time for a single drone to execute the first task, the maximum total number of the first type of drones (Task A drones) and the second type of drones (Task A and Task B integrated drones) required is determined; based on the total payload required for all task points to execute the second task (Task B) and the maximum payload for a single drone to execute the second task, the maximum total number of the second type of drones and the third type of drones (Task B drones) required is determined.

[0014] In some embodiments of this application, the drone swarm consists of three types of drones, as shown in Table 1: Table 1 Performance parameters of various types of UAVs In step S103, based on the improved genetic algorithm, with the goal of minimizing the total number of scheduled drones, the heterogeneous task allocation model is solved to obtain the task sequence of the scheduled drones. The improved genetic algorithm includes fitness function construction: the reciprocal of the sum of the non-zero real numbers of all gene positions on the chromosome is used as the fitness function, so that the chromosome individual with a smaller total number of scheduled drones has a higher fitness.

[0015] In some embodiments, determining the maximum total number of the first type of drones and the second type of drones required, based on the total time required to perform the first task at all task points and the maximum time for a single drone to perform the first task, includes: In a specific embodiment of this application, the following parameters are defined, U k : Drone swarm; : Task point set; : Starting point; i,j: Task point indices, (i,j) represents the distance from task point i to task point j; k: UAV number; : The distance between task point i and task point j; The time it takes for the drone swarm to execute task A at task point i; The time that drone k performs mission A at mission point i; Maximum flight time of the drone; : The actual flight time of the drone k in performing the mission; : The area of ​​task point i; : Value coefficient of task point i; v :Speed ​​parameters of the drone; w The scan width of the drone reconnaissance sensor; The number of drones in a drone swarm; The number of mission points; : Binary decision variable, equal to 1 indicates that drone k travels from task point i to task point j to perform the task, otherwise it is 0.

[0016] For task A, the time from start to finish for the first task within a defined target area should be kept within a set end time. For example, in some specific examples, the time from start to finish of Task A within a target area should be kept within 5 minutes, that is, the maximum time for a single UAV to reconnoiter the same target area and perform Task A is 5 minutes.

[0017] The number of drones for the first mission is: The number of integrated drones for the second mission is The number of drones for the third mission is For any target: The maximum quantities of each type of drone required are determined as follows:

[0018] Corresponding to the above within, .

[0019] In some embodiments, solving the heterogeneous task allocation model based on an improved genetic algorithm, with the objective of minimizing the total number of scheduled drones, includes: by , , Models were established to represent the number of drones for the first, second, and third missions, and the results were solved using a genetic algorithm.

[0020] in, For binary decision variables, for, This represents the total number of task points. The flight time required to travel from the takeoff point to mission point i. The flight time from mission point i back to the takeoff point

[0021] For any j ≠ 0, i ≠ j: but ; (3) In some embodiments, the improved genetic algorithm includes: Chromosome encoding: Real-number encoding is used. The chromosome length is determined by the maximum number of each type of UAV estimated in step S102 above. The real number on each gene position represents the mission point number that the corresponding UAV needs to go to. In a specific example, the chromosome length is 2·(max a + max b + max d), which is real-number encoding. The number on the gene position represents that the UAV needs to go to the corresponding real-number target to perform a reconnaissance or strike mission. A value of 0 means that it does not need to go to the target.

[0022] In some embodiments, the chromosome encoding includes: The gene segment corresponding to the first type of drone contains two gene sites, which represent the first task of two different executable task points; The gene segment corresponding to the second type of UAV contains two gene positions. The first position represents the target point of its mission. If the second position is the same as the first position, it means that the second mission with the maximum payload is performed at the target point. If it is 0, it means that the second mission with half the maximum payload is performed. The gene segment corresponding to the third type of drone contains two gene loci, with the same meaning as the gene segment of the second type of drone.

[0023] like Figure 2 As shown, in a specific example with 6 targets, gene fragments corresponding to the 1st, 2nd, and max a(b,c) drones from the A-mission drone, the AB-mission integrated drone, and the B-mission drone are selected respectively. Figure 2 The meanings of the solutions corresponding to the gene fragments are as follows: Drone 1A is to carry out mission A for target 1; Drone 2A is to carry out mission A for targets 2 and 4; Drone MAXA is to carry out mission A for targets 5 and 6. The No. 1 AB integrated UAV went to perform missions A and B of target 3 and launched 2 payloads; the No. 2 AB integrated UAV went to perform missions A and B of target 5 and launched 1 payload; the max b AB integrated UAV went to perform missions A and B of target 3 and launched 1 payload. Mission 1B drone proceeds to perform Mission B for Target 1 and launches 2 payloads; Mission 2B drone proceeds to perform Mission B for Target 2 and launches 2 payloads; Mission max c drone proceeds to perform Mission B for Target 1 and launches 1 payload.

[0024] In the gene segment encoding of the AB integrated UAV, since the A mission UAV can perform two different target A missions, the two genes of the A mission UAV can be different real numbers; since the AB integrated UAV and the B mission UAV can only perform single target missions, the first digit of the two genes represents the target number participating in the mission. If the second digit is the same as the first digit, it means that the attack mission launches two payloads. If the second digit is 0, it means that the attack mission launches one payload.

[0025] Improved construction of the fitness function in genetic algorithms: , which represents the sum of all real numbers on a single chromosome. The larger the number, the more drones are used; The smaller the value, the fewer drones are used. We construct the fitness function. This results in fewer drones. The larger the size, the higher the fitness.

[0026] The selection operation involves best-fit selection, using a roulette wheel selection method (a random sampling method where the probability of each individual entering the next generation is equal to the ratio of its fitness value to the sum of the fitness values ​​of all individuals in the entire population). The genetic algorithm then performs the selection operation and replicates the structure of the individual with the highest fitness in the current population to the next generation.

[0027] Interleaving operations, such as Figure 3 As shown, the characteristic of this chromosome is that the gene segments corresponding to drones of the same type are basically a whole, meaning that there is no functional difference between each drone, and crossover between the gene segments of different drones is meaningless. Therefore, we adopt single-point crossover, taking the node between the drone segment of mission A and the integrated drone segment AB as the crossover point, and performing crossover operation between the two parent bodies.

[0028] Mutation operations: These include real-valued mutation and self-checking and repair based on constraint order repair. In some implementations, the self-checking and repair based on constraint order repair includes: For the mutated chromosome, check whether it violates the model's constraints in a preset order. The preset order is as follows: maximum mission time constraint for a single UAV, mission point requirement satisfaction constraint, and total payload delivery constraint for each type of UAV. When any constraint violation is detected, a repair operation for the mutated part is triggered. The repair operation includes adjusting the task points of relevant gene loci according to the principle of minimum task time increment, or using unscheduled drone gene loci to share the over-limit tasks, so that the repaired chromosome meets all constraints.

[0029] The mutation operation consists of two parts, such as Figure 4 As shown, real-valued mutation and self-checking are performed. Specifically, first, all individuals in the population are judged whether to mutate based on the mutation probability. Then, the mutation bit of the mutated individuals is randomly selected and set to 0. Finally, the mutated individuals are self-checked. If it is found that the mutated individuals do not meet the requirements, the mutation operation is reversed.

[0030] In some embodiments, enabling unscheduled drone geometries to offload overloaded tasks includes: For drones that violate constraints due to total task execution time exceeding their maximum endurance, starting from the end of their task sequence, task points are transferred one by one to the corresponding gene positions of currently unscheduled drones with matching task types until the constraint is satisfied. For example, in some embodiments, if the total task time of a drone k exceeds its endurance limit, then starting from the end of its task sequence, the excess task points are "transferred" one by one to the corresponding gene positions of other currently idle drones of the same type. For example, if drone 1 of type 1 times out while executing the first task of objectives 2 and 4, a search is conducted to determine if there is a type 1 drone with a gene position of 0, and then the task of objective 4 is assigned to that idle drone, thereby correcting the constraint violation.

[0031] If the total strike payload required for a certain mission point is not met, priority will be given to assigning the mission to a reconnaissance / strike UAV that already has a reconnaissance mission at that mission point, or dispatching a new strike / reconnaissance UAV to undertake the task. The repair mechanism in this application ensures that the mutation operation always provides the population with repaired feasible solutions that meet the key constraints, guiding the search to proceed efficiently within the feasible region, thereby finding a globally optimal or near-optimal solution using fewer UAVs in a shorter time.

[0032] In some embodiments, the first mission is a reconnaissance mission, the second mission is a strike mission; the first type of UAV is a reconnaissance UAV, the second type of UAV is a reconnaissance and strike integrated UAV, and the third type of UAV is a strike UAV.

[0033] The method in this application employs a two-stage strategy of "pre-estimation-fine solution" to narrow the search space of the genetic algorithm from an infinite range to a finite and compact interval. Combined with a strategy to prevent invalid crossover, this improves the convergence speed of the algorithm by more than 30% compared to standard methods. The "self-checking and repair mechanism based on constraint order repair" proactively repairs infeasible solutions into feasible solutions, avoiding waste of computational resources and guiding the algorithm to escape local optima and find a task allocation scheme that uses fewer drones (and thus lowers costs).

[0034] The allocation results obtained by the method in this application are clear, and the mission sequence of each UAV is clear, which can be directly applied to actual command and control systems.

[0035] This application also proposes a heterogeneous UAV swarm task allocation system, including a processor and a memory. The memory stores a computer program, which, when executed by the processor, implements the steps of the aforementioned heterogeneous UAV swarm task allocation method.

[0036] It should be noted that, in the embodiments of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0037] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0038] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0039] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims. All of these forms are within the protection scope of this application.

Claims

1. A method for task allocation in a heterogeneous UAV swarm, used to generate task sequences for a swarm of UAVs containing multiple task capabilities, characterized in that, Includes the following steps: Establish a heterogeneous task allocation model for heterogeneous drone swarms; In the heterogeneous task allocation model, the drone swarm includes at least: a first type of drone dedicated to performing a first task, a second type of drone integrating the capabilities of the first and second tasks, and a third type of drone dedicated to performing the second task; Based on the total time required to execute the first task at all task points and the maximum time for a single drone to execute the first task, determine the maximum total number of the first type of drones and the second type of drones required; based on the total payload required to execute the second task at all task points and the maximum payload for a single drone to execute the second task, determine the maximum total number of the second type of drones and the third type of drones required. Based on an improved genetic algorithm, the heterogeneous task allocation model is solved with the goal of minimizing the total number of scheduled drones to obtain the task sequence of the scheduled drones. The improved genetic algorithm includes fitness function construction: the reciprocal of the sum of the non-zero real numbers of all gene positions on the chromosome is used as the fitness function, so that the chromosome individual with a smaller total number of scheduled drones has a higher fitness.

2. The heterogeneous UAV swarm task allocation method as described in claim 1, characterized in that, The improved genetic algorithm includes: Chromosome encoding: Real number encoding is used. The chromosome length is determined by the estimated maximum number of each type of drone. The real number on each gene locus represents the mission point number that the corresponding drone will go to. Mutation operations include real-valued mutation and self-checking and repair based on constraint order repair.

3. The heterogeneous UAV swarm task allocation method as described in claim 2, characterized in that, The self-check and repair based on constraint order repair includes: For the mutated chromosome, check whether it violates the model's constraints in a preset order. The preset order is as follows: maximum mission time constraint for a single UAV, mission point requirement satisfaction constraint, and total payload delivery constraint for each type of UAV. When any constraint violation is detected, a repair operation for the mutated part is triggered. The repair operation includes adjusting the task points of relevant gene loci according to the principle of minimum task time increment, or using unscheduled drone gene loci to share the over-limit tasks, so that the repaired chromosome meets all constraints.

4. The heterogeneous UAV swarm task allocation method as described in claim 3, characterized in that, Utilizing unscheduled drone gene loci to share the burden of overloaded tasks includes: For drones that violate constraints due to total mission execution time exceeding their maximum endurance, starting from the end of their mission sequence, the mission points are transferred one by one to the corresponding gene positions of drones that are not currently scheduled and whose mission types match, until the constraints are satisfied.

5. The heterogeneous UAV swarm task allocation method as described in claim 2, characterized in that, Based on the total time required to execute the first task at all task points and the maximum time for a single drone to execute the first task, the maximum total number of the required first type of drones and the second type of drones is determined, including: Define the time from start to finish for the first task within a target area to be maintained at a set end time. Inside; The number of drones for the first mission is: The number of integrated drones for the second mission is The number of drones for the third mission is For any target: The maximum quantities of each type of drone required are determined as follows: 。 6. The heterogeneous UAV swarm task allocation method as described in claim 5, characterized in that, Based on an improved genetic algorithm, the heterogeneous task allocation model is solved with the objective of minimizing the total number of scheduled drones, including: by , , Models were established to represent the number of drones for the first, second, and third missions, and the results were solved using a genetic algorithm. in, For binary decision variables, for, This represents the total number of task points. The flight time required to travel from the takeoff point to mission point i. The flight time is the time it takes to fly from mission point i back to the takeoff point.

7. The heterogeneous UAV swarm task allocation method as described in claim 6, characterized in that, In the chromosome encoding: The gene segment corresponding to the first type of drone contains two gene sites, which represent the first task of two different executable task points; The gene segment corresponding to the second type of UAV contains two gene positions. The first position represents the target point of its mission. If the second position is the same as the first position, it means that the second mission with the maximum payload is performed at the target point. If it is 0, it means that the second mission with half the maximum payload is performed. The gene segment corresponding to the third type of drone contains two gene loci, with the same meaning as the gene segment of the second type of drone.

8. The heterogeneous UAV swarm task allocation method as described in claim 1, characterized in that, The first mission is a reconnaissance mission, the second mission is a strike mission; the first type of UAV is a reconnaissance UAV, the second type of UAV is a reconnaissance and strike integrated UAV, and the third type of UAV is a strike UAV.

9. A heterogeneous unmanned aerial vehicle (UAV) swarm task allocation system, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program, which, when executed by the processor, implements the steps of the heterogeneous UAV swarm task allocation method as described in any one of claims 1 to 8.

Images