Unmanned aerial vehicle cluster cooperation system and method

Abstract

The invention provides an unmanned aerial vehicle cluster cooperation system and method, and relates to the technical field of unmanned aerial vehicle control, and the method comprises the steps that a cloud server splits a global task into a plurality of sub-task chains according to a received global task instruction, forms a sub-task list, and stores the sub-task list in a block chain; each unmanned aerial vehicle obtains a sub-task list, and submits a bidding request of the corresponding sub-task chain and writes the bidding request into the block chain when the degree of adaptation between the unmanned aerial vehicle and each sub-task chain is greater than a first threshold value; the block chain performs intelligent contract decision according to each bidding request, and generates a binding relationship between each sub-task chain and the corresponding unmanned aerial vehicle; and the unmanned aerial vehicle obtains a sub-task chain needing to be executed by the unmanned aerial vehicle according to the binding relationship, plans a sub-task path based on the sub-task chain, and then moves along the sub-task path to execute each sub-task in the sub-task chain. The method has the advantages that single-point calculation pressure is effectively relieved, dynamic adaptability is high, and transparency and credibility of task allocation are guaranteed due to tampering resistance of the block chain and automatic execution of the smart contract.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle control, and particularly to an unmanned aerial vehicle cluster cooperation system and method. Background Art

[0002] With the rapid development of artificial intelligence, Internet of Things and communication technologies, unmanned aerial vehicle cluster cooperation systems have become the core technical frameworks in fields such as intelligent perception, emergency response, and logistics transportation. Unmanned aerial vehicle clusters can significantly improve task execution efficiency, expand operation coverage, and enhance system robustness through multi-aircraft collaborative operations.

[0003] However, traditional unmanned aerial vehicle cluster systems mostly adopt a centralized task planning mode, where the ground control station uniformly calculates the task allocation plan. This mode will cause the following problems when the task scale is large:

[0004] 1) Single-point computing pressure: The ground station needs to process all the computing tasks of global tasks. When the number of unmanned aerial vehicles exceeds the threshold, the computing delay increases exponentially.

[0005] 2) Poor dynamic adaptability: It cannot respond to sudden tasks or environmental mutations in real time. Task replanning needs to go through the complete process of "upload - calculation - download", and the average response delay exceeds 500 ms.

[0006] 3) Lack of trust mechanism for distributed collaboration: Some studies attempt to adopt a distributed decision-making architecture. However, since unmanned aerial vehicle nodes need to make autonomous decisions on task selection but lack a reliable collaborative verification mechanism, it is easy to cause task competition or repeated execution, and the nodes only make decisions based on local information, making it difficult to evaluate the global resource distribution, resulting in uneven distribution of computing / communication resources. Summary of the Invention

[0007] In view of the problems existing in the prior art, the present invention provides an unmanned aerial vehicle cluster cooperation system, including:

[0008] A cloud server, connected to the unmanned aerial vehicle cluster, for disassembling the global task into multiple subtask chains according to the received global task instruction and forming a subtask list to be stored in the blockchain;

[0009] The unmanned aerial vehicle cluster includes multiple unmanned aerial vehicles, and each unmanned aerial vehicle includes:

[0010] A bidding module, configured to obtain the subtask list, calculate the fitness between itself and each subtask chain, and when the fitness is greater than a first threshold, submit a bidding request for the corresponding subtask chain to be written into the blockchain;

[0011] The blockchain makes intelligent contract adjudication according to each bidding request, and generates a binding relationship between each subtask chain and the corresponding unmanned aerial vehicle;

[0012] Each of the drones further includes:

[0013] A collaborative planning module, configured to obtain the subtask chain that it needs to execute according to the binding relationship, plan a subtask path based on the subtask chain, and then move along the subtask path to execute each subtask in the subtask chain.

[0014] Preferably, the cloud server includes:

[0015] A task parsing module, configured to parse the task objectives of the global task to obtain the task area, task type, and task priority included in the global task;

[0016] A task chain generation module, connected to the task parsing module, configured to divide the task area into multiple sub-areas, and generate multiple task chains according to the task type. Each task chain includes the area coordinates of the corresponding sub-area, multiple subtasks to be executed, and the dependency relationships of each subtask;

[0017] A list generation module, connected to the task chain generation module, configured to mark the execution priority for each task chain according to the task priority, and generate a subtask list with priorities.

[0018] Preferably, at least one sensor module is mounted on each of the drones;

[0019] The bidding module includes:

[0020] An ability evaluation unit, configured to calculate the fitness between itself and each subtask chain according to its remaining power, computing load, the available status of the sensor module, and communication strength;

[0021] A request submission unit, connected to the ability evaluation unit, configured to submit a bidding request including the fitness and the computing load when the fitness is greater than the first threshold.

[0022] Preferably, the ability evaluation unit includes:

[0023] A screening subunit, configured to obtain the required sensor types of each subtask chain, and determine its available sensor types according to the available status, and screen out the subtask chains whose available sensor types can cover the required sensor types and add them to the subtask set;

[0024] An adaptation subunit, connected to the screening subunit, configured to calculate the fitness between itself and each subtask chain in the subtask set according to its remaining power, the computing load, and the communication strength.

[0025] Preferably, the communication strength includes signal strength and bandwidth utilization rate;

[0026] The calculation formula of the adaptation degree is as follows:

[0027] Adaptation degree = ∑(task weight_i * ability value_i)

[0028] where i = 1, 2, 3;

[0029] Ability value_1 = the remaining power * the preset discharge efficiency coefficient;

[0030] Ability value_2 = (1 - calculation load) * peak computing power;

[0031] Ability value_3 = the signal strength * the bandwidth utilization rate.

[0032] Preferably, when the smart contract adjudicates that multiple of the bidding requests are associated with the sub-task chain:

[0033] When the maximum difference between the adaptation degrees of the bidding requests is greater than the second threshold, establish the binding relationship between the sub-task chain and the drone with the highest adaptation degree;

[0034] When the maximum difference between the adaptation degrees of the bidding requests is not greater than the second threshold, establish the binding relationship between the sub-task chain and the drone with the lowest calculation load.

[0035] Preferably, at least one sensor module is carried on each drone;

[0036] The cooperation planning module includes:

[0037] A data acquisition unit, configured to upload the local environment data collected by its own sensor module to the cloud server;

[0038] The cloud server includes a map construction module, configured to construct a global map according to the local environment data;

[0039] The cooperation planning module further includes:

[0040] A path generation unit, configured to plan an initial task path according to the global map and the regional coordinates of the sub-region associated with the sub-task chain and upload it to the cloud server;

[0041] The cloud server further includes a conflict detection module, which is used to perform spatio-temporal conflict detection based on the initial task paths submitted by each of the drones, generate a path confirmation signal and send it to the corresponding drone when there is no spatio-temporal conflict in the initial task path, and when there is a spatio-temporal conflict in the initial task path, send the path confirmation signal to the drone executing the high-priority task and generate a replanning signal and send it to the drone executing the low-priority task;

[0042] The collaborative planning module further includes:

[0043] A task execution unit, which is used to use the initial task path as the subtask path according to the path confirmation signal, and then move along the subtask path to execute each subtask in the subtask chain.

[0044] The present invention also provides a method for collaborative operation of a drone cluster, which is applied to the above-mentioned drone cluster collaborative system. The method for collaborative operation of the drone cluster includes:

[0045] Step S1, the cloud server in the drone cluster collaborative system disassembles the global task into a plurality of subtask chains according to the received global task instruction, forms a subtask list and stores it in the blockchain;

[0046] Step S2, each drone in the drone cluster collaborative system obtains the subtask list, calculates the fitness between itself and each subtask chain, and when the fitness is greater than a first threshold, submits a bid request for the corresponding subtask chain and writes it into the blockchain;

[0047] Step S3, the blockchain in the drone cluster collaborative system makes an intelligent contract adjudication according to each bid request, and generates a binding relationship between each subtask chain and the corresponding drone;

[0048] Step S4, each drone in the drone cluster collaborative system obtains the subtask chain that itself needs to execute according to the binding relationship, plans a subtask path based on the subtask chain, and then moves along the subtask path to execute each subtask in the subtask chain.

[0049] Preferably, the step S2 includes:

[0050] Step S21, the cloud server analyzes the task objectives of the global task to obtain the task area, task type and task priority included in the global task;

[0051] Step S22: The cloud server divides the task area into multiple sub - areas, and generates multiple task chains according to the task type. Each task chain contains the area coordinates of the corresponding sub - area, multiple sub - tasks to be executed, and the dependency relationships of each sub - task;

[0052] Step S23: The cloud server marks the execution priorities for each task chain according to the task priorities, and generates the list of sub - tasks with priorities.

[0053] Preferably, at least one sensor module is carried on each drone;

[0054] The said Step S4 includes:

[0055] Step S41: Each drone uploads the local environment data collected by its own sensor module to the cloud server;

[0056] Step S42: The cloud server constructs a global map according to each local environment data;

[0057] Step S43: Each drone plans an initial task path according to the global map and the area coordinates of the sub - area associated with the sub - task chain, and uploads it to the cloud server;

[0058] Step S44: The cloud server performs spatio - temporal conflict detection according to the initial task paths submitted by each drone. When there is no spatio - temporal conflict in the initial task path, it generates a path confirmation signal and sends it to the corresponding drone. When there is a spatio - temporal conflict in the initial task path, it sends the path confirmation signal to the drone performing the high - priority task, and generates a replanning signal and sends it to the drone performing the low - priority task;

[0059] Step S45: Each drone takes the initial task path as the sub - task path according to the path confirmation signal, and then moves along the sub - task path to execute each sub - task in the sub - task chain.

[0060] The above - mentioned technical solution has the following advantages or beneficial effects:

[0061] 1) The cloud server converts the global task into sub - task chains that can be executed in parallel, forming a distributed task pool, avoiding the ground station from centrally processing all calculations. At the same time, the drone calculates the fitness through the local bidding module, dispersing the computing pressure to the edge nodes, effectively alleviating the single - point computing pressure;

[0062] 2) After the cloud server instantaneously disassembles the global task, it directly pushes it to the blockchain. The drones can obtain the new task list in milliseconds for real-time bidding, triggering the smart contract to automatically adjust the binding relationship of subtasks, without waiting for the ground station to process, eliminating multiple upload and download steps, thus being able to effectively handle sudden tasks or environmental mutations and having strong dynamic adaptability;

[0063] 3) Use the blockchain to store the subtask list and bidding requests, and the smart contract adjudicates the binding relationship. The immutability of the blockchain and the automatic execution of the smart contract can ensure the transparency and credibility of task allocation. In addition, each drone makes decisions based on the global subtask list rather than local information, so resource allocation is more uniform, avoiding contention and duplication. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 In a preferred embodiment of the present invention, it is a schematic structural diagram of a drone swarm cooperation system.;

[0065] Figure 2 In a preferred embodiment of the present invention, it is a schematic flowchart of a drone swarm cooperation method;

[0066] Figure 3 In a preferred embodiment of the present invention, it is a schematic sub-flowchart of step S2;

[0067] Figure 4 In a preferred embodiment of the present invention, it is a schematic sub-flowchart of step S4. DETAILED DESCRIPTION OF THE INVENTION

[0068] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. The present invention is not limited to this embodiment, and other embodiments can also fall within the scope of the present invention as long as they conform to the gist of the present invention.

[0069] In a preferred embodiment of the present invention, in view of the above problems existing in the prior art, a drone swarm cooperation system is provided, as Figure 1 shown, including:

[0070] A cloud server 100, connected to the drone swarm 200, for disassembling the global task into multiple subtask chains according to the received global task instruction and forming a subtask list to be stored in the blockchain 300;

[0071] The drone swarm 200 includes multiple drones 201, and each drone 201 includes:

[0072] A bidding module 1, for obtaining the subtask list, calculating the fitness between itself and each subtask chain, and when the fitness is greater than the first threshold, submitting a bidding request for the corresponding subtask chain to be written into the blockchain 300;

[0073] The blockchain 300 makes intelligent contract rulings based on each bidding request, generating the binding relationships between each subtask chain and the corresponding unmanned aerial vehicle 201.

[0074] Each unmanned aerial vehicle 201 further includes:

[0075] A cooperation planning module 2, configured to obtain the subtask chain that itself needs to execute according to the binding relationship, plan the subtask path based on the subtask chain, and then move along the subtask path to execute each subtask in the subtask chain.

[0076] Specifically, in this embodiment, preferably, a client is further provided, which establishes a communication connection with the cloud server 100 and is used for the user to upload a global task instruction. Or the user can upload a global task instruction through a ground station, such as "search for target A within area X and transmit data to base station Y in real time". After receiving the global task instruction, the cloud server first splits the global task included in the global task instruction into multiple subtask chains (such as area search, target tracking, etc.), and then stores the subtask list formed based on the subtask chain in the blockchain 300 for each unmanned aerial vehicle 201 to view.

[0077] Each unmanned aerial vehicle 201 preferably obtains the subtask list in batches. For example, if the subtask list contains 100 subtask chains, 10 can be obtained each time, avoiding the sudden increase in storage space occupation and affecting the computing load. Subsequently, a bidding request is submitted according to the adaptability between its own capabilities and the subtask chain. Then, the optimal executor is automatically matched through the intelligent contract, and the binding relationship between the subtask chain and the unmanned aerial vehicle as the optimal executor is established, that is, the allocation of the subtask chain is realized.

[0078] Each unmanned aerial vehicle 201 can also determine the subtask chain assigned to itself based on the above binding relationship, and then plan the subtask path. Subsequently, it moves along the subtask path to execute each subtask in the subtask chain, realizing the cooperation of the unmanned aerial vehicle group to complete the global task.

[0079] In a preferred embodiment of the present invention, the cloud server 100 includes:

[0080] A task parsing module 3, configured to parse the task objectives of the global task to obtain the task area, task type, and task priority included in the global task;

[0081] A task chain generation module 4, connected to the task parsing module 3, configured to divide the task area into multiple sub-areas, and generate multiple task chains according to the task type. Each task chain includes the area coordinates of the corresponding sub-area, multiple subtasks that need to be executed, and the dependency relationships of each subtask;

[0082] The list generation module 5 is connected to the task chain generation module 4 and is used to mark the execution priorities for each task chain according to the task priorities, and generate a sub-task list with priorities.

[0083] Specifically, in this embodiment, still taking the global task instruction "search for target A within area X and transmit data to base station Y in real time" as an example, after the cloud server 100 receives this global task instruction, it preferably extracts the keywords in the global task instruction through the natural language processing module, and obtains the following keywords: "search", "area X", "target A", "transmission", "base station Y". Furthermore, the task can be split into atomic operation units ATU:

[0084] ATU1: Rasterization segmentation of area X (spatial decomposition);

[0085] ATU2: Sub-region scanning and data collection (functional decomposition);

[0086] ATU3: Feature recognition of target A (data processing);

[0087] ATU4: Encrypted data transmission to base station Y (communication task).

[0088] For the above ATU1 spatial decomposition: Preferably, determine the segmentation granularity of area X according to the scale of the UAV cluster. For example, divide area X into multiple grids, and each grid is a sub-region. The size of each grid is, for example, 100m×100m. It can be understood that the grid size here is only an example and not a limitation on its size.

[0089] For the above ATU2-ATU3: Preferably, split them into sequential stages according to the task process to form a chain dependency, and obtain a sub-task chain: For example, it is necessary to scan the sub-region first, then perform target recognition, and finally perform data transmission.

[0090] Furthermore, store each sub-task chain in the form of a sub-task list in the sub-task list. Among them, the corresponding sub-task chain can be marked with a high priority based on the configured task urgency.

[0091] In a preferred embodiment of the present invention, at least one sensor module 202 is carried on each UAV 201;

[0092] The bidding module 1 includes:

[0093] The capability evaluation unit 11 is used to calculate the fitness between itself and each sub-task chain according to its remaining power, computing load, the available status of the sensor module, and the communication intensity;

[0094] The request submission unit 12 is connected to the capability evaluation unit 11 and is used to submit a bidding request including the fitness and the computing load when the fitness is greater than the first threshold.

[0095] In a preferred embodiment of the present invention, the capability evaluation unit 11 includes:

[0096] A screening subunit 111, configured to obtain the required sensor types of each subtask chain, determine its own available sensor types according to the available status, and screen out the subtask chains whose available sensor types can cover the required sensor types and add them to the subtask set;

[0097] An adaptation subunit 112, connected to the screening subunit 111, calculates the adaptation degree between itself and each subtask chain in the subtask set according to its remaining power, computing load, and communication intensity.

[0098] Specifically, in this embodiment, the above-mentioned sensor module includes but is not limited to radar sensors, vision sensors, and infrared sensors. If the required sensor type of a subtask chain is a radar sensor, and the current unmanned aerial vehicle itself does not carry a radar sensor, or the loaded radar sensor is unavailable, it means that the available sensor types of the unmanned aerial vehicle do not cover the required sensor types, that is, the unmanned aerial vehicle cannot execute this subtask chain and has no bidding significance, so it is directly discarded without performing the next adaptation degree calculation, saving computing power.

[0099] In a preferred embodiment of the present invention, the communication intensity includes signal strength and bandwidth utilization rate;

[0100] The calculation formula of the adaptation degree is as follows:

[0101] Adaptation degree = ∑(task weight_i * capability value_i)

[0102] Where i = 1, 2, 3;

[0103] Capability value_1 = remaining power * preset discharge efficiency coefficient;

[0104] Capability value_2 = (1 - computing load) * peak computing power;

[0105] Capability value_3 = signal strength * bandwidth utilization rate.

[0106] Specifically, in this embodiment, the above-mentioned task weight_i is preferably pre-configured and stored in a grouped manner, and each group of task weight_i is associated with different task types, such as search tasks, tracking tasks, transmission tasks, etc. Based on the different requirement weights of the capability values for different task types, they are pre-allocated for calling when calculating the adaptation degree.

[0107] In a preferred embodiment of the present invention, the smart contract adjudication is as follows when multiple bidding requests are associated with a subtask chain:

[0108] When the maximum difference between the adaptation degrees of each bidding request is greater than the second threshold, a binding relationship is established between the subtask chain and the unmanned aerial vehicle with the highest adaptation degree;

[0109] When the maximum difference between the fitness degrees of each bidding request is not greater than the second threshold, a binding relationship is established between the subtask chain and the drone with the lowest computing load.

[0110] In a preferred embodiment of the present invention, at least one sensor module is mounted on each drone;

[0111] The cooperation planning module 2 includes:

[0112] The data acquisition unit 21 is configured to upload the local environment data collected by its own sensor module to the cloud server;

[0113] The cloud server 100 includes a map construction module 6 for constructing a global map according to the local environment data;

[0114] The cooperation planning module 2 further includes:

[0115] The path generation unit 22 is configured to plan an initial task path according to the global map and the regional coordinates of the sub-region associated with the subtask chain and upload it to the cloud server;

[0116] The cloud server 100 further includes a conflict detection module 7 for performing spatio-temporal conflict detection according to the initial task paths submitted by each drone, generating a path confirmation signal and sending it to the corresponding drone when there is no spatio-temporal conflict in the initial task path, and when there is a spatio-temporal conflict in the initial task path, sending the path confirmation signal to the drone executing the high-priority task and generating a replanning signal and sending it to the drone executing the low-priority task;

[0117] The cooperation planning module 2 further includes:

[0118] The task execution unit 23 is configured to use the initial task path as the subtask path according to the path confirmation signal, and then move along the subtask path to execute each subtask in the subtask chain.

[0119] Specifically, in this embodiment, taking the sensor module including a millimeter-wave radar (generating 3D point cloud), a high-dynamic vision camera (extracting ORB-SLAM feature points), and an ultrasonic array (detecting obstacles within 10 meters) as an example, the map construction module 6 preferably aligns the point clouds of multiple drones through the ICP algorithm, combines GPS / visual SLAM to generate a global coordinate system, and maps the ultrasonic obstacle data into occupancy markers of vector grids (accuracy 0.1 meter).

[0120] The map construction module 6 then generates building outlines (vector polygons) based on the millimeter-wave point cloud, marks the passable areas (Bézier curve paths) based on the SLAM feature points, and annotates temporary obstacles (lightweight grid indexes) based on the ultrasonic data to construct a global map.

[0121] Subsequently, each drone can view the global map and call the deep reinforcement learning model (DRL) to plan an initial task path based on the global map and upload it to the cloud server. The cloud server detects path intersections and time overlaps based on the initial task paths uploaded by each drone (e.g., two drones will meet at coordinates (x,y) at t = 120s). If there are such path intersections between at least two initial task paths, it is considered that there is a spatio-temporal conflict. At this time, the generated replanning signal preferably includes the conflict coordinates and time so that the drone can avoid the conflict coordinates and time points when replanning the sub-task path.

[0122] The present invention also provides a method for drone swarm cooperation, which is applied to the above-mentioned drone swarm cooperation system, such as Figure 2 shown, the drone swarm cooperation method includes:

[0123] Step S1, the cloud server in the drone swarm cooperation system disassembles the global task into multiple sub-task chains according to the received global task instruction and forms a sub-task list to be stored in the blockchain;

[0124] Step S2, each drone in the drone swarm cooperation system obtains the sub-task list, calculates the fitness between itself and each sub-task chain, and when the fitness is greater than the first threshold, submits a bidding request for the corresponding sub-task chain to be written into the blockchain;

[0125] Step S3, the blockchain in the drone swarm cooperation system makes an intelligent contract adjudication according to each bidding request and generates a binding relationship between each sub-task chain and the corresponding drone;

[0126] Step S4, each drone in the drone swarm cooperation system obtains the sub-task chain that it needs to execute according to the binding relationship, plans the sub-task path based on the sub-task chain, and then moves along the sub-task path to execute each sub-task in the sub-task chain.

[0127] In a preferred embodiment of the present invention, as Figure 3 shown, step S2 includes:

[0128] Step S21, the cloud server analyzes the task objectives of the global task to obtain the task area, task type, and task priority included in the global task;

[0129] Step S22, the cloud server divides the task area into multiple sub-areas, generates multiple task chains according to the task type, and each task chain includes the area coordinates of the corresponding sub-area, multiple sub-tasks to be executed, and the dependency relationship of each sub-task;

[0130] Step S23, the cloud server marks the execution priority for each task chain according to the task priority to generate a sub-task list with priorities.

[0131] In a preferred embodiment of the present invention, at least one sensor module is carried on each unmanned aerial vehicle;

[0132] As Figure 4 shown, step S4 includes:

[0133] Step S41, each unmanned aerial vehicle uploads the local environment data collected by its own sensor module to the cloud server;

[0134] Step S42, the cloud server constructs a global map based on the local environment data;

[0135] Step S43, each unmanned aerial vehicle plans an initial task path according to the global map and the regional coordinates of the sub-region associated with the sub-task chain and uploads it to the cloud server;

[0136] Step S44, the cloud server performs spatio-temporal conflict detection on the initial task paths submitted by each unmanned aerial vehicle, and when there is no spatio-temporal conflict in the initial task path, generates a path confirmation signal and sends it to the corresponding unmanned aerial vehicle, and when there is a spatio-temporal conflict in the initial task path, sends the path confirmation signal to the unmanned aerial vehicle performing the high-priority task, and generates a replanning signal and sends it to the unmanned aerial vehicle performing the low-priority task;

[0137] Step S45, each unmanned aerial vehicle takes the initial task path as the sub-task path according to the path confirmation signal, and then moves along the sub-task path to execute each sub-task in the sub-task chain.

[0138] The above are only preferred embodiments of the present invention, and do not limit the implementation manners and protection scope of the present invention. For those skilled in the art, it should be able to realize that all equivalent replacements and obvious changes made by using the content of this specification and the drawings should be included in the protection scope of the present invention.

Claims

1. A drone swarm collaboration system, characterized in that, Including: A cloud server, connected to a drone cluster, for disassembling a global task into multiple sub-task chains according to the received global task instruction and forming a sub-task list to be stored in a blockchain; The drone cluster includes multiple drones, and each drone includes: A bidding module, for obtaining the sub-task list, calculating the fitness between itself and each sub-task chain, and when the fitness is greater than a first threshold, submitting a bidding request for the corresponding sub-task chain to be written into the blockchain; The blockchain makes an intelligent contract adjudication based on each bidding request to generate a binding relationship between each sub-task chain and the corresponding drone; Each drone further includes: A cooperation planning module, for obtaining the sub-task chain that itself needs to execute according to the binding relationship, planning a sub-task path based on the sub-task chain, and then moving along the sub-task path to execute each sub-task in the sub-task chain.

2. The drone swarm cooperation system according to claim 1, wherein The cloud server includes: A task parsing module, for parsing the task objective of the global task to obtain the task area, task type, and task priority included in the global task; A task chain generation module, connected to the task parsing module, for dividing the task area into multiple sub-areas, and generating multiple task chains according to the task type, where each task chain includes the area coordinates of the corresponding sub-area, multiple sub-tasks to be executed, and the dependency relationship of each sub-task; A list generation module, connected to the task chain generation module, for marking the execution priority for each task chain according to the task priority, and generating the sub-task list with priorities.

3. The drone swarm cooperation system according to claim 1, characterized in that, At least one sensor module is mounted on each drone; The bidding module includes: A capability evaluation unit, for calculating the fitness between itself and each sub-task chain according to its remaining power, computing load, the available state of the sensor module, and communication strength; A request submission unit, connected to the capability evaluation unit, for submitting the bidding request including the fitness and the computing load when the fitness is greater than the first threshold.

4. The drone swarm cooperation system according to claim 3, wherein The capability evaluation unit includes: A screening subunit, for obtaining the required sensor types of each sub-task chain, and judging its available sensor types according to the available state, and screening out the sub-task chains whose available sensor types can cover the required sensor types to be added to a sub-task set; An adaptation subunit, connected to the screening subunit, for calculating the fitness between itself and each sub-task chain in the sub-task set according to its remaining power, computing load, and communication strength.

5. The drone swarm cooperation system according to claim 4, wherein The communication strength includes signal strength and bandwidth utilization rate; The calculation formula of the fitness is as follows: Fitness = ∑(task weight_i * capability value_i) where i = 1, 2, 3; Capability value_1 = remaining power * preset discharge efficiency coefficient; Capability value_2 = (1 - computing load) * peak computing power; Capability value_3 = signal strength * bandwidth utilization rate.

6. The drone swarm cooperation system according to claim 3, wherein When the smart contract makes a judgment that when multiple bidding requests are associated with the subtask chain: When the maximum difference between the fitness degrees of the bidding requests is greater than a second threshold, establish the binding relationship between the subtask chain and the drone with the highest fitness degree; When the maximum difference between the fitness degrees of the bidding requests is not greater than the second threshold, establish the binding relationship between the subtask chain and the drone with the lowest computing load.

7. The drone swarm cooperation system according to claim 1, characterized in that, At least one sensor module is carried on each drone; The collaborative planning module includes: A data acquisition unit for uploading the local environment data collected by its own sensor module to the cloud server; The cloud server includes a map construction module for constructing a global map according to the local environment data; The collaborative planning module further includes: A path generation unit for planning an initial task path according to the global map and the regional coordinates of the sub-region associated with the subtask chain and uploading it to the cloud server; The cloud server further includes a conflict detection module for performing spatio-temporal conflict detection according to the initial task paths submitted by the drones, and generating a path confirmation signal to the corresponding drone when there is no spatio-temporal conflict in the initial task path, and when there is a spatio-temporal conflict in the initial task path, sending the path confirmation signal to the drone performing the high-priority task and generating a replanning signal to the drone performing the low-priority task; The collaborative planning module further includes: A task execution unit for using the initial task path as the subtask path according to the path confirmation signal, and then moving along the subtask path to execute each subtask in the subtask chain.

8. A method for collaborative operation of an unmanned aerial vehicle cluster, characterized in that, Applied to the drone cluster cooperation system as described in any one of claims 1-7, the drone cluster cooperation method includes: Step S1, the cloud server in the drone cluster cooperation system disassembles the global task into multiple subtask chains according to the received global task instruction and forms a subtask list to be stored in the blockchain; Step S2, each drone in the drone cluster cooperation system obtains the subtask list, calculates the fitness degree between itself and each subtask chain, and when the fitness degree is greater than a first threshold, submits a bidding request for the corresponding subtask chain to be written into the blockchain; Step S3, the blockchain in the drone cluster cooperation system makes a smart contract judgment according to each bidding request, and generates a binding relationship between each subtask chain and the corresponding drone; Step S4, each drone in the drone cluster cooperation system obtains the subtask chain that it needs to execute according to the binding relationship, plans a subtask path based on the subtask chain, and then moves along the subtask path to execute each subtask in the subtask chain.

9. The method for collaborative operation of a drone swarm according to claim 8, wherein, The step S2 includes: Step S21, the cloud server analyzes the task objectives of the global task to obtain the task area, task type, and task priority included in the global task; Step S22: The cloud server divides the task area into multiple sub-areas and generates multiple task chains according to the task type. Each task chain contains the area coordinates of the corresponding sub-area, multiple sub-tasks to be executed, and the dependency relationships of each sub-task. Step S23: The cloud server marks the execution priorities for each task chain according to the task priorities, generating a list of the sub-tasks with priorities.

10. The method for collaborative operation of a drone swarm according to claim 8, wherein, At least one sensor module is carried on each drone. The step S4 includes: Step S41: Each drone uploads the local environment data collected by its own sensor module to the cloud server. Step S42: The cloud server constructs a global map based on the local environment data. Step S43: Each drone plans an initial task path according to the global map and the area coordinates of the sub-area associated with the sub-task chain and uploads it to the cloud server. Step S44: The cloud server performs spatio-temporal conflict detection on the initial task paths submitted by each drone. When there is no spatio-temporal conflict in the initial task path, it generates a path confirmation signal and sends it to the corresponding drone. When there is a spatio-temporal conflict in the initial task path, it sends the path confirmation signal to the drone executing the high-priority task and generates a replanning signal and sends it to the drone executing the low-priority task. Step S45: Each drone takes the initial task path as the sub-task path according to the path confirmation signal, and then moves along the sub-task path to execute each sub-task in the sub-task chain.

Images