Multi-unmanned aerial vehicle intelligent task collaboration method and task collaboration system under satellite assistance

The satellite-assisted multi-UAV system optimizes resource allocation and task offloading through a centralized training and decentralized execution framework, addressing resource insufficiencies and delays in traditional drone computing architectures, achieving enhanced service success rates.

CN119093995APending Publication Date: 2024-12-06BEIJING UNIV OF POSTS & TELECOMM
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411130760.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-16
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

Traditional computing architectures based on a single drone face challenges such as insufficient computing resources and network delays, especially in remote areas where dense ground infrastructure is lacking, hindering real-time network services and data analysis capabilities.

Method used

A satellite-assisted multi-UAV intelligent task coordination method that allocates computing and communication resources across multiple UAV clusters, utilizing a centralized training and decentralized execution architecture to optimize task offloading decisions through a critic and performer network trained by a satellite control center.

Benefits of technology

Improves the success rate of mission services by up to 40% compared to single-UAV systems and 30% compared to DDPG, enhancing the efficiency of data processing and resource allocation in drone edge computing scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119093995A_ABST
    Figure CN119093995A_ABST
Patent Text Reader

Abstract

The invention provides a satellite-assisted multi-unmanned aerial vehicle intelligent task collaboration method and task collaboration system, and relates to the technical field of mobile edge computing. According to the method, after receiving detection control information issued by a satellite control center, an unmanned aerial vehicle cluster head determines a local environment state of the unmanned aerial vehicle cluster head to an environment, and task unloading decisions of task unmanned aerial vehicles in the unmanned aerial vehicle cluster to corresponding local ground areas; determining a network resource allocation result of each task unmanned aerial vehicle and a task service reward corresponding to the corresponding network resource allocation result; and the satellite control center collects a local environment state, a task unloading decision and a task service reward, and trains the commentator network and the performer network configured by each unmanned aerial vehicle cluster head, so that each unmanned aerial vehicle cluster head updates the performer network. According to the method, computing resources and communication resources are reasonably allocated based on a multi-unmanned aerial vehicle task cooperation mode of an unmanned aerial vehicle edge computing architecture, and the task service success rate is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of mobile edge computing, and more specifically, to a satellite-assisted multi-UAV intelligent task coordination method and task coordination system. Background Art

[0002] With the large-scale application of technologies such as cloud computing, artificial intelligence, and the Internet of Things, the amount of data has shown an exponential growth. According to relevant institutions, there will be more than 100 billion terminal devices connected to the Internet. In order to solve the drawbacks of high latency and lack of real-time data analysis capabilities under traditional data processing methods, mobile edge computing technology has emerged. For example, in intelligent transportation systems, integrating edge computing into vehicle-mounted fog computing networks can effectively solve the limitations of traditional cloud computing frameworks. However, in some remote areas where dense ground infrastructure cannot be built, it is difficult to provide real-time network services. Therefore, there is an urgent need for a new network architecture to provide global communication coverage and popularize intelligent computing to meet the various service needs of all IoT terminal users. Due to the flexibility of drones, drone-based edge computing technology continues to develop and is widely used in a variety of computing scenarios to reduce task processing latency and meet user needs.

[0003] However, with the increasing number of user requests, the traditional computing architecture based on a single drone may face many problems such as insufficient computing resources and network delays. Summary of the invention

[0004] The purpose of the embodiments of the present application is to provide a satellite-assisted multi-UAV intelligent task coordination method and task coordination system, which is used to reasonably allocate computing resources and communication resources in a multi-UAV task coordination manner based on a UAV edge computing architecture, thereby improving the success rate of mission services.

[0005] In a first aspect, a satellite-assisted multi-UAV intelligent task coordination method is provided, which is applied to a task coordination system, wherein the task coordination system includes a satellite control center and a plurality of UAV clusters detecting a local ground area, each UAV cluster includes a plurality of task UAVs collecting task data from a corresponding local ground area in each time slot, and a UAV cluster head communicating with the plurality of task UAVs; different UAV cluster heads are connected to each other in communication and are all connected to the satellite control center, and the method may include:

[0006] For any drone cluster in any time slot, after receiving the detection control information sent by the satellite control center, the drone cluster head in the drone cluster determines its own local environmental state according to the mission data collected by the corresponding mission drones in the time slot;

[0007] The UAV cluster head processes the local observation state using the performer network configured by itself, and determines the task offloading decision of each task UAV in the UAV cluster to the corresponding local ground area;

[0008] The UAV cluster head determines the network resource allocation result of each task UAV and the task service reward corresponding to the corresponding network resource allocation result according to the task offloading decision of each task UAV and the local environment state;

[0009] The satellite control center uses the local environment status, corresponding task offloading decision and corresponding task service reward sent by each UAV cluster head received in different time slots as training data;

[0010] The satellite control center uses the training data to train the critic network and the performer network configured for the corresponding drone cluster head, and sends the trained performer network to each drone cluster head, so that each drone cluster head returns to execute the steps through the trained performer network: for any drone cluster in any time slot, the drone cluster head in the drone cluster determines its own local environmental state to the environment according to the mission data collected by the corresponding mission drones in the time slot.

[0011] In a possible implementation, the multiple drone clusters are determined based on the communication range of each drone cluster head, and the multiple mission drones in each drone cluster are used to jointly complete image acquisition of a local ground area.

[0012] In a possible implementation, the task data includes the task data size of each task drone in the time slot, the number of CPU cycles required for each task drone to process one bit of data, and the delay threshold for each task drone to process the corresponding task;

[0013] The local environment state includes a set of distances between each task drone and the drone cluster head in the drone cluster head communication area, a set of task data sizes of each task drone, a set of CPU resources required for each bit of data of each task drone, a set of task delay thresholds of each task drone, a set of task queues of each task drone in this time slot, a current task queue of the drone cluster head in this time slot, and the difference between the time slot in which the current task queue has been completely processed and this time slot.

[0014] In one possible implementation, the task offloading decision includes a first decision to execute the task on the task drone, a second decision to execute the task on the drone cluster head, and a third decision to migrate the task to other drone cluster heads for execution.

[0015] In one possible implementation, the network resources include communication resources and computing resources;

[0016] The UAV cluster head determines the network resource allocation result of each task UAV and the task service reward corresponding to the corresponding network resource allocation result according to the task offloading decision of each task UAV and the local environment state, including:

[0017] According to the task offloading decision of each task drone and the local environment state, the communication resources of each task drone are allocated according to the frequency division multiplexing method;

[0018] Prioritize the allocation of computing resources to the tasks uploaded by the corresponding task drone in the same time slot, and then reallocate computing resources to execute the tasks migrated from other drone cluster heads according to the time sequence of uploading the migration tasks, and obtain the task processing delay of the drone cluster head itself.

[0019] Based on the task processing delay Determine the task service reward corresponding to the network resource allocation result.

[0020] In a possible implementation, if the task offloading decision is the first decision, then the task processing delay The computational delay for each mission drone to process the task Delays associated with waiting for computing resources of and;

[0021] If the task offloading decision is the second decision, then the task processing delay The computational delay for each mission drone to process the task Waiting delay for computing resources The time delay between the task being uploaded from the task drone to the drone cluster head The return delay between the UAV cluster heads after the task calculation is completed of and;

[0022] If the task offloading decision is the third decision, the task processing delay The computational delay for each mission drone to process the task Waiting delay for computing resources The time delay between the task being uploaded from the task drone to the drone cluster head The return delay between the UAV cluster heads after the task calculation is completed The delay caused by task migration between the UAV cluster head The sum.

[0023] In a possible implementation, the task service reward is expressed as:

[0024] Among them, r represents the task service reward, α and β represent the proportion of the two items, which are between 0 and 1. represents the derivative of the sum of service delays of all incoming tasks in time slot t, represents the task service success rate in time slot t, K t represents all the tasks of the system in time slot t, t i is the delay threshold for the i-th task UAV to process the corresponding task.

[0025] In one possible implementation, if but otherwise

[0026] In a possible implementation, the satellite control center uses the training data to train the critic network and the performer network of the corresponding drone cluster head configuration, including:

[0027] The local environment states and corresponding task offloading decisions sent by each UAV cluster head received at different time slots are used as input data of the critic network, and the critic network is trained by guiding the corresponding task service rewards;

[0028] The output of the critic network is used to guide the training of the performer network.

[0029] In a second aspect, a task coordination system is provided, which may include: a satellite control center and a plurality of drone clusters detecting a local ground area;

[0030] Each drone cluster includes a plurality of mission drones that collect mission data from a corresponding local ground area in each time slot, and a drone cluster head that is communicatively connected to the plurality of mission drones; different drone cluster heads are communicatively connected to each other and are all connected to the satellite control center;

[0031] Each drone cluster head in the multiple drone clusters and the satellite control center are respectively used to execute the corresponding method described in any one of the first aspects.

[0032] The satellite-assisted multi-UAV intelligent task coordination method provided in the embodiment of the present application is for any UAV cluster in any time slot. After the UAV cluster head in the UAV cluster receives the detection control information sent by the satellite control center, it determines its own local environmental state to the environment according to the task data collected by the corresponding task UAVs in the time slot; the UAV cluster head processes the local observation state by using the performer network configured by itself to determine the task offloading decision of each task UAV in the UAV cluster for the corresponding local ground area; the UAV cluster head determines the network resource allocation result and the corresponding task unloading decision of each task UAV according to the task offloading decision and the local environmental state of each task UAV. The task service reward corresponding to the network resource allocation result; the satellite control center uses the local environment state, corresponding task offloading decision and corresponding task service reward sent by each drone cluster head received in different time slots as training data; the satellite control center uses the training data to train the critic network and performer network configured by the corresponding drone cluster head, and sends the trained performer network to each drone cluster head, so that each drone cluster head returns to the execution step through the trained performer network: for any drone cluster in any time slot, the drone cluster head in the drone cluster determines its own local environment state for the environment according to the task data collected by the corresponding task drones in the time slot. This method reasonably allocates computing resources and communication resources based on the multi-drone task collaboration mode of the drone edge computing architecture, and improves the success rate of task services. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0034] Figure 1 A schematic diagram of the structure of a task collaboration system provided in an embodiment of the present application;

[0035] Figure 2 A flowchart of a satellite-assisted multi-UAV intelligent task coordination method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. Unless otherwise defined, the technical terms or scientific terms used in this application should be the common meanings understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in this application do not indicate any order, quantity or importance, but are only used to distinguish different components. "Including" or "including" and other similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connected", "coupled" or "connected" and other similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0037] The satellite-assisted multi-UAV intelligent task coordination method provided in the embodiment of the present application can be applied to Figure 1 In the task collaboration system shown in Figure 1 As shown, the task coordination system may include:

[0038] A satellite control center and multiple drone clusters for detecting local ground areas, for example, drone cluster 1 for detecting local ground area Area1, drone cluster 2 for detecting local ground area Area2, ... .

[0039] Each drone cluster includes multiple mission drones that collect mission data from the corresponding local ground area in each time slot, and a drone cluster head that is communicatively connected to the multiple mission drones. For example, drone cluster 1 includes drone cluster head 1-1 and three mission drones 1-2.

[0040] Among them, different drone cluster heads are connected to each other through communication, and different drone cluster heads are connected to the satellite control center to receive the detection and control information sent by the satellite control center, and send the detected environmental information to the satellite control center (the local environmental status, corresponding task offloading decision and corresponding task service reward sent by each drone cluster head in different time slots).

[0041] It should be noted that multiple drone clusters are determined based on the communication range of each drone cluster head. Multiple mission drones in each drone cluster are used to jointly complete image acquisition of a local ground area, and the entire swarm collaborates to complete the processing of the collected data.

[0042] The satellite-assisted multi-UAV intelligent task coordination method of the present application divides the UAV group into multiple small clusters according to the communication range of the UAV cluster head. A small UAV cluster includes a UAV cluster head and multiple mission UAVs. The mission UAVs detect mountainous areas and collect image data. Afterwards, a centralized training and decentralized execution architecture is adopted to train the critic network (Critic Network) and the intelligent agent performer network (ActorNetwork) in the satellite control center. The UAV cluster head of each cluster decides the task offloading plan. The method pre-processes information such as images of steep mountains captured by UAVs, reduces data redundancy, and improves the success rate of mission services.

[0043] The preferred embodiments of the present application are described below in conjunction with the drawings in the specification. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and are not used to limit the present application. In addition, the embodiments and features in the embodiments of the present application may be combined with each other if there is no conflict.

[0044] Figure 2 A schematic diagram of a satellite-assisted multi-UAV intelligent task coordination method provided in an embodiment of the present application. Figure 2 As shown, the method may include:

[0045] Step S210: For any drone cluster in any time slot, after the drone cluster head in the drone cluster receives the detection control information sent by the satellite control center, it determines its own local environmental state based on the mission data collected by the corresponding mission drones in the time slot.

[0046] Among them, the time slot set is T = {1, 2, ..., t}, the drone cluster head set is E = {1, 2, ..., e}, the set of all task drones in the system is I = {1, 2, ..., i}, and the set of task drones covered by each drone cluster head is

[0047] The task data may include the task data size of each task drone in the time slot, the number of CPU cycles required for each task drone to process one bit of data, and the delay threshold for each task drone to process the corresponding task; it is expressed as: task data where d i represents the mission data size of the i-th mission drone in this time slot, c i represents the number of CPU cycles required for the i-th task drone to process one bit of data in this time slot, t i Represents the delay threshold for processing the i-th mission UAV of this mission.

[0048] The local environment state can include the set of distances between each task drone and the drone cluster head in the drone cluster head communication area, the set of task data sizes of each task drone, the set of CPU resources required for each bit of data of each task drone, the set of task delay thresholds of each task drone, the set of task queues of each task drone in this time slot, the current task queue of the drone cluster head in this time slot and the difference between the time slot when the current task queue has been completely processed and this time slot. It is expressed as: local environment state in, represents the set of mission drones i under the communication area of ​​drone cluster head e in time slot t, represents the set of distances between the mission UAVs and the UAV cluster head in the UAV cluster head e communication area, A collection representing the size of the mission data of the mission drone, Represents the set of CPU resources required for each bit of data of the mission drone, represents the set of mission delay thresholds of the mission UAV, represents the set of task queues possessed by the UAV with the task in time slot t, represents the task queue of the UAV cluster head e in time slot t, where e and t represent the index of the UAV cluster head and the time slot index respectively. It represents the current task queue of the UAV cluster head in the tth time slot and the difference between the time slot t′ and the tth time slot when the current task queue has been completely processed.

[0049] Step S220: The UAV cluster head processes the local observation state using the performer network configured by itself, and determines the task offloading decision of each task UAV in the UAV cluster to the corresponding local ground area.

[0050] Among them, in the initial case, the network parameters in the performer network configured by the drone cluster head itself are randomly generated.

[0051] The task offloading decision may include a first decision to execute the task on the task drone, a second decision to execute the task on the drone cluster head, and a third decision to migrate the task to other drone cluster heads for execution.

[0052] The task offloading decision can be expressed as in It indicates whether the task UAV i processes the task on e′, and e′∈{E,i} indicates whether the offloading location is the task UAV local, offloading from the task UAV local or migrating to other UAV cluster heads.

[0053] The task offloading decision is implemented through the multi-agent deep reinforcement learning MADDPG algorithm. There are three locations for task offloading: processing locally on the task drone, offloading processing at the drone cluster head of the cluster, or migrating to the drone cluster head of another cluster for processing.

[0054] Step S230: The UAV cluster head determines the network resource allocation result of each task UAV and the task service reward corresponding to the corresponding network resource allocation result according to the task offloading decision of each task UAV and the local environment status.

[0055] In the specific implementation, the communication resources of each mission UAV are allocated according to the task offloading decision of each mission UAV and the local environmental status based on the frequency division multiplexing method;

[0056] Priority is given to allocating computing resources to the tasks uploaded and unloaded by the corresponding task drones in the same time slot for execution. Then, computing resources are reallocated to the tasks migrated from other drone cluster heads according to the time sequence of uploading the migration tasks. Specifically, with regard to the allocation of computing resources, for the tasks uploaded to the drone cluster head, there are tasks uploaded and unloaded by task drones in its own area. At the same time, there are also tasks migrated from drone cluster heads in other areas. For different tasks, the drone cluster head will give priority to the tasks unloaded by task drones in its own area in the same time slot, and then allocate computing resources according to the time sequence of uploading the migration tasks.

[0057] At the same time, the task processing delay of the UAV cluster head itself is obtained Based on task processing delay Determine the task service reward corresponding to the network resource allocation result.

[0058] Furthermore, if the task offloading decision is the first decision, the task processing delay The computational delay for each mission drone to process the task Delays associated with waiting for computing resources of and;

[0059] If the task offloading decision is the second decision, the task processing delay The computational delay for each mission drone to process the task Waiting delay for computing resources The time delay between the task being uploaded from the task drone to the drone cluster head The return delay between the UAV cluster heads after the task calculation is completed of and;

[0060] If the task offloading decision is the third decision, the task processing delay The computational delay for each mission drone to process the task Waiting delay for computing resources The time delay between the task being uploaded from the task drone to the drone cluster head The return delay between the UAV cluster heads after the task calculation is completed The delay caused by task migration between the UAV cluster head The sum.

[0061] It is expressed as:

[0062]

[0063] Task service rewards can be expressed as:

[0064] Among them, r represents the task service reward, α and β represent the proportion of the two items, which are between 0 and 1. represents the derivative of the sum of service delays of all incoming tasks in time slot t, represents the task service success rate in time slot t, K t represents all the tasks of the system in time slot t, t i is the delay threshold for the i-th task UAV to process the corresponding task. but otherwise

[0065] Furthermore, in order to obtain the highest task service success rate, constraints can be set:

[0066]

[0067] C1 means is a 0-1 integer variable indicating whether the task of task UAV i is executed at l, l∈{E,i} indicates that the task offloading location is all UAV cluster heads or task UAV i itself, indicates that the task is processed at position l, while Indicates that the task is not executed at location l; C2 indicates that the task can only be processed in one place: locally by task UAV i, offloaded to the UAV cluster head in the cluster for processing, or migrated to the UAV cluster head in another cluster for processing.

[0068] Step S240: The satellite control center uses the local environment status, corresponding task offloading decision and corresponding task service reward sent by each UAV cluster head received in different time slots as training data, and trains the critic network and performer network configured for the corresponding UAV cluster head to obtain a trained critic network and a trained performer network.

[0069] In the specific implementation, the local environment status and corresponding task offloading decisions sent by each drone cluster head received at different time slots are used as the input data of the critic network, and the critic network is trained by guiding the corresponding task service rewards; the output results of the critic network are used to guide the training of the performer network.

[0070] Step S250: The satellite control center sends the trained performer network to each UAV cluster head, so that each UAV cluster head replaces its own performer network with the trained performer network, and returns to execute step S210.

[0071] Each drone cluster head only needs to make unloading decisions based on its own local observation status.

[0072] Experimental results show that the service success rate of this scheme is about 40% higher than that of the task processing algorithm with only a single UAV, and about 30% higher than that of DDPG.

[0073] In another embodiment provided in the present application, a computer-readable storage medium is also provided, which stores instructions. When the computer-readable storage medium is executed on a computer, the computer executes the satellite-assisted multi-UAV intelligent task coordination method described in any of the above embodiments.

[0074] In another embodiment provided in the present application, a computer program product comprising instructions is also provided, which, when executed on a computer, enables the computer to execute the satellite-assisted multi-UAV intelligent task coordination method described in any of the above embodiments.

[0075] Those skilled in the art will appreciate that the embodiments in the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt a complete hardware embodiment, a complete software embodiment, or a form of an embodiment combining software and hardware. Moreover, the present application may adopt a form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0076] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0077] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0078] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0079] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the embodiments of the present application.

[0080] Obviously, those skilled in the art can make various changes and modifications to the embodiments in the present application without departing from the spirit and scope of the embodiments in the present application. Thus, if these modifications and variations of the embodiments in the present application fall within the scope of the claims and their equivalents in the embodiments of the present application, the embodiments of the present application are also intended to include these modifications and variations.

Claims

1. A satellite-assisted multi-UAV intelligent task coordination method, characterized in that: The invention is applied to a task coordination system, wherein the task coordination system includes a satellite control center and a plurality of drone clusters detecting a local ground area, each drone cluster includes a plurality of task drones collecting task data from a corresponding local ground area in each time slot, and a drone cluster head communicating with the plurality of task drones; Different drone cluster heads are connected to each other in communication and are all connected to the satellite control center. The method includes: For any drone cluster in any time slot, after receiving the detection control information sent by the satellite control center, the drone cluster head in the drone cluster determines its own local environmental state according to the mission data collected by the corresponding mission drones in the time slot; The UAV cluster head processes the local observation state using the performer network configured by itself, and determines the task offloading decision of each task UAV in the UAV cluster to the corresponding local ground area; The UAV cluster head determines the network resource allocation result of each task UAV and the task service reward corresponding to the corresponding network resource allocation result according to the task offloading decision of each task UAV and the local environment state; The satellite control center uses the local environment status, corresponding task offloading decision and corresponding task service reward sent by each UAV cluster head received in different time slots as training data; The satellite control center uses the training data to train the critic network and the performer network configured for the corresponding drone cluster head, and sends the trained performer network to each drone cluster head, so that each drone cluster head returns to execute the steps through the trained performer network: for any drone cluster in any time slot, the drone cluster head in the drone cluster determines its own local environmental state to the environment according to the mission data collected by the corresponding mission drones in the time slot.

2. The method according to claim 1, characterized in that The multiple drone clusters are determined according to the communication range of each drone cluster head, and the multiple mission drones in each drone cluster are used to jointly complete image acquisition of a local ground area.

3. The method according to claim 1, characterized in that The task data includes the task data size of each task drone in the time slot, the number of CPU cycles required for each task drone to process one bit of data, and the delay threshold for each task drone to process the corresponding task; The local environment state includes a set of distances between each task drone and the drone cluster head in the drone cluster head communication area, a set of task data sizes of each task drone, a set of CPU resources required for each bit of data of each task drone, a set of task delay thresholds of each task drone, a set of task queues of each task drone in this time slot, a current task queue of the drone cluster head in this time slot, and the difference between the time slot in which the current task queue has been completely processed and this time slot.

4. The method according to claim 1, characterized in that The task offloading decision includes a first decision to execute the task on the task drone, a second decision to execute the task on the drone cluster head, and a third decision to migrate the task to other drone cluster heads for execution.

5. The method according to claim 4, characterized in that Network resources include communication resources and computing resources; The UAV cluster head determines the network resource allocation result of each task UAV and the task service reward corresponding to the corresponding network resource allocation result according to the task offloading decision of each task UAV and the local environment state, including: According to the task offloading decision of each task drone and the local environment state, the communication resources of each task drone are allocated according to the frequency division multiplexing method; Prioritize the allocation of computing resources to the tasks uploaded by the corresponding task drone in the same time slot, and then reallocate computing resources to execute the tasks migrated from other drone cluster heads according to the time sequence of uploading the migration tasks, and obtain the task processing delay of the drone cluster head itself. Based on the task processing delay Determine the task service reward corresponding to the network resource allocation result.

6. The method according to claim 5, characterized in that If the task offloading decision is the first decision, the task processing delay is the sum of the computing delay of each task drone processing the task and the waiting delay of waiting for computing resources; If the task offloading decision is the second decision, the task processing delay is the sum of the calculation delay of each task drone processing the task, the waiting delay for computing resources, the delay between uploading the task from the task drone to the drone cluster head, and the return delay between the drone cluster heads after the task calculation is completed; If the task offloading decision is the third decision, then the task processing delay is the sum of the calculation delay of each task drone processing the task, the waiting delay for computing resources, the delay between uploading the task from the task drone to the drone cluster head, the backhaul delay between the drone cluster heads after the task calculation is completed, and the delay caused by task migration between drone cluster heads.

7. The method according to claim 5, characterized in that The task service reward is expressed as: Among them, r represents the task service reward, α and β represent the proportion of the two items, which are between 0 and 1. represents the derivative of the sum of service delays of all incoming tasks in time slot t, represents the task service success rate in time slot t, K t represents all the tasks of the system in time slot t, t i is the delay threshold for the i-th task UAV to process the corresponding task.

8. The method according to claim 7, characterized in that like but otherwise 9. The method according to claim 1, characterized in that The satellite control center uses the training data to train the critic network and the performer network of the corresponding drone cluster head configuration, including: The local environment states and corresponding task offloading decisions sent by each UAV cluster head received at different time slots are used as input data of the critic network, and the critic network is trained by guiding the corresponding task service rewards; The output of the critic network is used to guide the training of the performer network.

10. A task collaboration system, characterized in that: The mission coordination system includes a satellite control center and a plurality of drone clusters that detect local ground areas; Each drone cluster includes a plurality of mission drones that collect mission data from a corresponding local ground area in each time slot, and a drone cluster head that is communicatively connected to the plurality of mission drones; different drone cluster heads are communicatively connected to each other and are all connected to the satellite control center; Each drone cluster head in the multiple drone clusters and the satellite control center are respectively used to execute the corresponding method described in any one of claims 1-9.

Citation Information

Cited By

  • Unmanned cluster assisted multi-strategy network service dynamic migration method and device

    CN119653395A