Unmanned aerial vehicle distributed task allocation method based on time constraint

Abstract

The invention provides an unmanned aerial vehicle distributed task allocation method based on time constraint, and belongs to the technical field of unmanned aerial vehicle command decision. The method specifically comprises the steps that an unmanned aerial vehicle task beam, RPI values of all tasks by an unmanned aerial vehicle and an unmanned aerial vehicle sequence are generated, when the task execution of the unmanned aerial vehicle meets the time constraint, the RPI values are set to be infinite according to a PI algorithm, and when the task execution of the unmanned aerial vehicle does not meet the time constraint, the RPI values are set to be infinite; the unmanned aerial vehicles communicate with each other, for each task, the unmanned aerial vehicle with the lowest RPI value is reserved to execute the task, the other unmanned aerial vehicles delete the task, and consistent processing is achieved; for the unmanned aerial vehicle with the maximum number of deleted tasks, deleting all tasks needing to be executed by the unmanned aerial vehicle; for other unmanned aerial vehicles, conflict tasks are removed again, and corresponding RPI values and unmanned aerial vehicle sequences are updated; and if the task allocation result of each unmanned aerial vehicle is no longer changed, the task allocation of the unmanned aerial vehicles is ended, otherwise, a new round of task allocation is started for the tasks to which the unmanned aerial vehicles are not allocated.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicle (UAV) command and decision-making, and in particular relates to a UAV distributed task allocation method based on time constraints. Background Art

[0002] Unmanned aerial vehicles (UAVs), with their advantages of high flexibility, good concealment, and low cost, play a vital role in military and civilian applications such as reconnaissance and strike, information collection, environmental monitoring, and search and rescue. UAV task allocation aims to solve the problem of matching UAVs with tasks. Based on the presence of a distribution center, UAV task allocation methods can be categorized as centralized or distributed. Centralized task allocation relies on a central node to generate task allocation plans and distribute the results to each node in the swarm system. This approach suffers from single points of failure and high computational complexity. Distributed task allocation often resolves conflicts through communication, interaction, and negotiation among individuals within the swarm to maximize the collective interests of the group. It does not rely on global information and has therefore become the primary approach currently used in UAV task allocation research. The PI (performance impact) algorithm is one of the most effective methods for solving UAV task allocation, particularly in time-constrained UAV tasks in known environments, where each task requires a UAV to begin executing by a specified time, and a single UAV can complete multiple tasks consecutively.

[0003] The PI (Performance Impact) algorithm is one of the most effective methods for allocating drone tasks, particularly those within known environments and subject to time constraints. While the PI algorithm can effectively address the problem when the number of tasks is small, it suffers from a low task completion rate when the number of drones is large, resulting in unassigned tasks. Therefore, this paper improves the PI algorithm's task elimination strategy, increasing the number of tasks performed by each drone and improving the completion rate of drone task assignments. Summary of the Invention

[0004] In view of this, the present invention provides a method for allocating distributed tasks of unmanned aerial vehicles (UAVs) taking time constraints into consideration. This method can achieve the most reasonable task allocation and improve the completion rate of UAV task allocation.

[0005] The technical solutions for implementing the present invention are as follows:

[0006] A distributed task allocation method for UAVs based on time constraints. The specific process is as follows:

[0007] Step 1: Collect initial parameters, including the latest start time of the task;

[0008] Step 2: Generate UAV mission bundle p i , UAV RPI value for all tasks and drone sequence β i , where the RPI value generation process is as follows: when the UAV performs a task that meets the time constraint, the RPI value corresponding to the UAV performing the task is calculated according to the PI algorithm; when the time constraint is not met, the RPI value of the UAV performing the task is set to infinity;

[0009] Step 3: The drones communicate with each other. For each task, the drone with the lowest RPI value is retained to perform the task, and the remaining drones delete the task to achieve consistency processing. The number of tasks that each drone needs to delete is obtained. i , RPI value and drone sequence β i ;

[0010] Step 4: For the drone with the largest number of deleted tasks, delete all the tasks it needs to perform and update its corresponding RPI value and drone sequence;

[0011] Step 5: For the remaining drones, re-eliminate conflicting tasks and update their corresponding RPI values ​​and drone sequences;

[0012] Step 6: If the task allocation result of each drone does not change, the drone task allocation ends. Otherwise, a new round of task allocation starts for the tasks that have not been allocated to drones.

[0013] Furthermore, the time constraint of the present invention is that the time when the UAV performs the task is later than the latest start time of the task.

[0014] Furthermore, the present invention removes conflicting tasks for the remaining drones. The specific process is as follows: let the task to be deleted for drone i be d i ;

[0015] Determine the task d that drone i needs to delete i Length |d i |>0 is true, if |d i |>0 is not true, the conflict task elimination of drone i ends, if |d i |>0, UAV i enters the conflict task elimination phase, which includes the following steps:

[0016] (1) Calculate the RPI value of drone i for all tasks according to the PI algorithm

[0017] (2) Calculate the RPI value and RPI value The maximum difference and the task number corresponding to the maximum difference are respectively denoted as g ◇ and t q ;

[0018] (3) Determine g ◇ >0 is true, if g ◇ >0 is established, t q From p i and d i , re-enter the drone mission elimination phase, if g ◇ >0 is not true, and the conflict task elimination of drone i ends.

[0019] Furthermore, when constructing task bundles based on the PI algorithm, the present invention aims to maximize the number of task assignments.

[0020] Furthermore, in the present invention, when there are at least two UAVs with the largest number of tasks, the UAV with the larger number is selected to delete the task.

[0021] Beneficial effects:

[0022] First, the PI algorithm determines whether to remove a task by comparing the RPI of each drone for that task. The basic principle is that the drone with the lowest RPI retains the task, while the remaining drones abandon it. Once a task is assigned to a drone, such as U1, it cannot be removed from U1's task queue unless another drone obtains a lower RPI than U1, resulting in some tasks being unable to be added to U1. By clearing and deleting the task bundle of the drone with the largest number of tasks, the present invention ensures that more tasks are added to drones.

[0023] Second, the RPI value of the drone for a task in the PI algorithm is closely related to the sequence of tasks being executed. Once a task is eliminated, the RPI value of the tasks after it will also change. By updating the RPI of all tasks, the present invention can ensure that the RPI values ​​of the remaining tasks are the latest, and can more accurately eliminate unnecessary tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0025] Figure 1 Flowchart of the distributed task allocation method for UAVs considering time constraints.

[0026] Figure 2 Schematic diagram of drone and mission distribution. DETAILED DESCRIPTION

[0027] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0028] It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments may be combined with each other; and, based on the embodiments in this disclosure, all other embodiments obtained by persons of ordinary skill in the art without creative work are within the scope of protection of this disclosure.

[0029] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on this disclosure, it should be understood by those skilled in the art that an aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects described herein can be used to implement an apparatus and / or practice a method. In addition, other structures and / or functionalities other than one or more of the aspects described herein can be used to implement this apparatus and / or practice this method.

[0030] like Figure 1 As shown, the embodiment of the present application is a method for allocating distributed tasks of drones based on time constraints, and the specific process is as follows:

[0031] Step 1: Given known environmental information, including the location, speed, and number of drones, the task location, number, speed, and the latest start time.

[0032] Step 2: Each UAV uses a greedy algorithm based on known environmental information to maximize the number of task assignments and constructs a task bundle using the PI algorithm based on the task time constraint.

[0033] The following takes the construction of the task bundle of UAV i under the initial condition as an example to generate the UAV task bundle p i , UAV RPI value for all tasks Drone Sequence β i For example, p1 represents the sequence of tasks that drone 1 can perform, such as {task 1, task 3}. ={RPI_1, RPI_2, RPI_3} represents the RPI values ​​obtained by drone 1 when it sequentially executes the tasks in p1. If drone 1's execution time for tasks 1 and 3 meets the time constraints, RPI_1 and RPI_3 are calculated using the PI algorithm. If the execution time for drone 2 does not meet the time constraints, the corresponding RPI values ​​are set to infinity. β1 represents drone 1's perceived task allocation. For example, β1 = {1, 2, 1} indicates that drone 1 believes that tasks 1 and 3 are assigned to drone 1, and task 2 is assigned to drone 2.

[0034] Step 3: UAV i sends the result and timestamp of task bundle construction in step 2 to neighboring UAVs. For the same task, the UAV with lower RPI retains the task, while other UAVs abandon the task. Therefore, after completing the above consistency processing, the number of tasks that UAV i needs to delete is calculated as θ i , RPI value Drone Sequence β i .

[0035] Step 4: According to the consistency processing results in step 3, determine the drone with the largest number of deleted tasks, that is, θ in step 3 i The drone corresponding to the maximum value; let the drone with the largest number of deleted tasks be numbered k.

[0036] Determine whether the current drone i is the drone with the most deleted tasks. If so, that is, k = i, then clear the task bundle p of drone i i , update the drone i and β i , the elimination of conflicting tasks of drone i is completed; if the current drone i is not the drone with the largest number of deleted tasks, that is, k≠i, go to step 5; if there are multiple corresponding drone numbers, select the drone with the larger drone number.

[0037] Step 5: Determine the mission d that drone i needs to delete i Length |d i |>0 is true. If |d i |>0 is not true, and the conflict task elimination of drone i ends. i |>0 holds, UAV i enters the conflict task elimination phase, which includes the following sub-steps.

[0038] Sub-step 5.1: Calculate the RPI value of drone i for all tasks according to the PI algorithm

[0039] Sub-step 5.2: Find the with the RPI value in step 2 The maximum difference and the task number corresponding to the maximum difference are respectively recorded as g ◇ and t q ;

[0040] Sub-step 5.3: Determine g ◇ >0. If g ◇ >0 is established, t q From p i and d i , repeat step 5, if g ◇ >0 is not true, and the conflict task elimination of drone i ends.

[0041] Step 6: If the task allocation results for each drone do not change, the algorithm converges and the drone task allocation ends. Otherwise, jump to step 2 and start a new round of task allocation.

[0042] Below is N a UAVs against N T The method of the present invention is described in detail as follows:

[0043] Step 1, N a Drone location Drone speed N T The location of the target Target speed Latest execution time

[0044] Step 2: Each UAV uses a greedy algorithm based on known environmental information to maximize the number of task assignments and build a task bundle based on the PI algorithm while meeting the task time constraints. Each UAV generates a task bundle p i , UAV RPI value for all tasks Drone Sequence β i ,

[0045] Step 3, drone i, The result of step 2 task bundle construction p i , timestamp time i Send to nearby drones Based on the PI algorithm, the consistency processing is completed and each drone Update the number of tasks that need to be deleted θ i , RPI value Drone Sequence β i ;

[0046] Step 4: Calculate the number of drones i, The set of tasks that need to be eliminated d i ,θ i . Find θ i The largest drone k. Drone k clears the task bundle, updates the RPI value and drone sequence, i.e.

[0047] Step 5: Determine other drones i, i≠k|d i |>0 is true. If |d i |>0 is not true, and the conflict task elimination of drone i ends. i|>0 holds, UAV i enters the conflict task elimination phase, which includes the following sub-steps.

[0048] Sub-step 5.1: Calculate the number of drones i according to the PI algorithm. RPI value of all tasks i≠k

[0049] Sub-step 5.2: Find the with the RPI value in step 2 The maximum difference and the task number corresponding to the maximum difference are respectively recorded as g ◇ and t q ;

[0050] Sub-step 5.3: Determine g ◇ >0. If g ◇ >0 is established, t q From p i and d i Remove from the list and repeat step 5. ◇ >0 is not true, and the conflict task elimination of drone i ends.

[0051] Step 6: If the drone i, If the task allocation result does not change, it means that the algorithm has converged and the drone task allocation is completed. Otherwise, jump to step 2 and start a new round of task allocation.

[0052] Example:

[0053] This embodiment adopts a method for allocating distributed tasks to drones with consideration of time constraints provided by the present invention, and improves the completion rate of drone task allocation by improving the conflicting task elimination mechanism of the PI algorithm. The specific execution steps are as follows:

[0054] Step 1: Consider three drones performing 10 tasks. Each task requires one drone to execute before the latest start time. Each drone can perform multiple tasks. The drone's speed is [100-200] m / s. The latest start time of each task is randomly generated between [0, 300] s. The drones are evenly distributed along the X-axis, and the task location information is randomly generated within a 10 km × 10 km square, as shown in the following example: Figure 2 shown.

[0055] In step 2, each UAV uses a greedy algorithm based on its position, speed, and latest start time to maximize the number of assigned tasks. Under the condition of meeting the task time constraints, a task bundle is constructed based on the PI algorithm. The constructed task bundle is as follows:

[0056] p1=[7 2 5 9 3 1 10 6 4 8];

[0057] p2=[8 4 2 7 10 1 3 9 5 6];

[0058] p3=[8 4 2 5 9 6 1 3 10 7];

[0059] β1=[1 1 1 1 1 1 1 1 1 1];

[0060] β2=[2 2 2 2 2 2 2 2 2 2];

[0061] β3=[3 3 3 3 3 3 3 3 3 3];

[0062] Step 3: Send the result of task bundle construction in step 2, timestamp, and number of deleted tasks to neighboring drones, complete consistency processing based on the PI algorithm, and update the number of tasks to be deleted θ i , RPI value Drone Sequence β i Since a task can only be completed by one drone alone, according to the PI algorithm, for the same task, the drone with lower RPI retains the task and other drones give up the task. Therefore, after consistency processing, the drone task bundle p i , drone sequence β i as follows:

[0063] p1=[7 2 5 9 3 1 10 6 4 8];

[0064] p2=[8 4 2 7 10 1 3 9 5 6];

[0065] p3=[8 4 2 5 9 6 1 3 10 7];

[0066] β1=[1 2 2 3 3 1 1 3 1 2];

[0067] β2=[1 2 2 3 3 1 1 3 1 2];

[0068] β3=[1 2 2 3 3 1 1 3 1 2];

[0069] The number of tasks that the drone needs to delete are 6, 7, and 7 respectively.

[0070] Step 4: Since the number of tasks eliminated by drone 3 is large and the sequence number of drone 3 is large, the invalid RPI value in the task sequence is the largest. Therefore, drone 3 eliminates all tasks, and the drone task bundle p i , drone sequence β i as follows:

[0071] p1=[7 2 5 9 3 1 10 6 4 8];

[0072] p2=[8 4 2 7 10 1 3 9 5 6];

[0073] p3 = [];

[0074] β1=[1 2 2 3 3 1 1 3 1 2];

[0075] β2=[1 2 2 3 3 1 1 3 1 2];

[0076] β3=[0 0 0 0 0 0 0 0 0 0];

[0077] Step 5. Since the number of tasks to be deleted in drone 1 is d1=6>0, follow sub-steps 5.1-5.3 to remove tasks in sequence. Similarly, the number of tasks to be deleted in drone 2 is d2=7>0, follow sub-steps 5.1-5.3 to remove tasks in sequence. According to step 5, after a task is deleted, step 5.1 is recalculated. Step 5.2 Recalculate g ◇ If the value of g ◇ > 0 is not true, the elimination of conflicting tasks of drones is completed, so not all tasks in d1 or d2 are necessarily deleted. After eliminating conflicting tasks, the drone task bundle p i , drone sequence β i as follows:

[0078] p1 = [7 9 1 6];

[0079] p2 = [2 10 3];

[0080] p3 = [];

[0081] β1=[1 2 2 3 3 1 1 3 1 2];

[0082] β2=[1 2 2 3 3 1 1 3 1 2];

[0083] β3=[0 0 0 0 0 0 0 0 0 0];

[0084] Step 6: Since the UAV task allocation is not completed, repeat the process of step 2 and repeat the above task allocation process until the task allocation results converge. The final task allocation result is:

[0085] p1 = [7 9 3 1];

[0086] p2 = [2 10 9 5];

[0087] p3 = [8 4];

[0088] The present invention can increase the number of time-constrained UAV task assignments and improve the completion rate of UAV task assignments. The reasons for taking the above measures are:

[0089] First, the PI algorithm compares each drone's RPI for a task and decides whether to remove it. The basic principle is that the drone with the lowest RPI retains the task, while the remaining drones abandon it. Once a task is assigned to a drone, such as U1, it cannot be removed from U1's task list unless another drone achieves a lower RPI than U1, resulting in some tasks being unable to be added to U1. By clearing the task bundle for the drone with the largest number of tasks in step 4, more tasks can be added to the drones.

[0090] Second, the RPI value of the drone for a task in the PI algorithm is closely related to the sequence of tasks being executed. Once a task is eliminated, the RPI value of the tasks after it will also change. By updating the RPI of all tasks in sub-step 5.1, the RPI values ​​of the remaining tasks can be guaranteed to be the latest, and unnecessary tasks can be eliminated more accurately.

[0091] In summary, the above are merely embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the scope of protection of the present invention.

Claims

1. A method for allocating distributed tasks of unmanned aerial vehicles based on time constraints, characterized in that: The specific process is as follows: Step 1: Collect initial parameters, including the latest start time of the task; Step 2: Generate UAV mission bundle p i , UAV RPI value for all tasks and drone sequence β i , wherein the RPI value generation process is as follows: when the UAV performs a task that meets the time constraint, the RPI value corresponding to the UAV performing the task is calculated according to the PI algorithm; when the time constraint is not met, the RPI value corresponding to the UAV performing the task is set to infinity; Step 3: The drones communicate with each other. For each task, the drone with the lowest RPI value is retained to perform the task, and the remaining drones delete the task to achieve consistency processing. The number of tasks that each drone needs to delete is obtained. i 、RPI value and drone sequence β i ; Step 4: For the drone with the largest number of deleted tasks, delete all the tasks it needs to perform, and update its corresponding RPI value and drone sequence; Step 5: For the remaining drones, re-eliminate the conflicting tasks and update their corresponding RPI values ​​and drone sequences; Step 6: If the task allocation result of each drone does not change, the drone task allocation ends. Otherwise, a new round of task allocation starts for the tasks of unassigned drones.

2. The method for allocating distributed tasks of unmanned aerial vehicles based on time constraints according to claim 1 is characterized in that: The time constraint is that the time when the drone performs the task is later than the latest start time of the task.

3. The method for allocating distributed tasks of unmanned aerial vehicles based on time constraints according to claim 2 is characterized in that: For the remaining drones, conflicting tasks are removed again. The specific process is as follows: suppose the task that drone i needs to delete is d i ; Determine the task d that drone i needs to delete i Length |d i |>0 is true, if |d i |>0 is not true, the conflict task elimination of drone i ends, if |d i |>0 is established, and drone i enters the conflict task elimination phase, which includes the following steps: (1) Calculate the RPI value of drone i for all tasks according to the PI algorithm (2) Calculate the RPI value With RPI value The maximum difference of and the task number corresponding to the maximum difference are recorded as and t q ; (3)Judgment Is it established? If Established, t q From p i and d i , re-enter the drone mission elimination phase, if Not true, the conflict task elimination of drone i ends.

4. The method for allocating distributed tasks of unmanned aerial vehicles based on time constraints according to claim 1 is characterized in that: When constructing task bundles based on the PI algorithm, the goal is to maximize the number of task assignments.

5. The method for allocating distributed tasks of unmanned aerial vehicles based on time constraints according to claim 1 is characterized in that: When there are at least two drones with the largest number of tasks, the drone with the larger number is selected for task deletion.

Images