Task planning method and device for unmanned cluster

Abstract

The invention discloses a task planning method and device for an unmanned cluster. The method comprises the steps of obtaining a to-be-completed task information set and a capability index information set of the unmanned cluster; the capability index information set comprises a capability index value set of each unmanned target of an unmanned cluster; performing classification processing on the capability index information set of the unmanned cluster to obtain a plurality of sub-cluster index sets; the sub-cluster index set comprises a capability index value set of all unmanned targets contained in a sub-cluster; and performing multi-level task planning processing on the to-be-completed task information set and the sub-cluster index set to obtain a task allocation information set.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned swarm and the field of information processing, and in particular to a task planning method and device for unmanned swarm. BACKGROUND

[0002] In the field of unmanned swarm cooperative operation, the rationality of task planning directly affects the operation efficiency and task success rate. However, the existing methods have significant defects in unmanned swarm task allocation:

[0003] Coarse division of swarm capability: Traditional methods are mostly based on single indicators (such as endurance time, load capacity) to classify unmanned devices, which is difficult to reflect the comprehensive differences of devices in multi-dimensional capabilities (communication, perception, execution). For example, in the complex terrain mapping task, dividing the swarm according to endurance capacity may lead to devices with insufficient perception accuracy undertaking high-precision mapping tasks, resulting in task failure.

[0004] Low efficiency of task and capability matching: Existing task planning relies on artificial preset rules or simple weight allocation, which cannot dynamically adapt to changes in task demand. When there are multiple index constraints (such as timeliness, accuracy, energy consumption) in the task, traditional methods are prone to "overload-idle" polarization problem, i.e., part of the devices are overloaded, while other devices are idle.

[0005] High computational complexity and poor scalability: When using global search or heuristic algorithms (such as genetic algorithm, ant colony algorithm) for task allocation, the computational complexity increases exponentially with the size of the swarm, making it difficult to meet real-time requirements; at the same time, when facing dynamic joining / leaving of devices or changes in task priority, it is difficult to adjust the algorithm, and it is difficult to quickly reconstruct the task allocation strategy.

[0006] Lack of data-driven intelligent decision-making: The existing scheme does not fully utilize the capability data and task historical data of unmanned devices, making it difficult to tap the collaborative potential among devices. For example, in emergency rescue tasks, it is difficult to identify efficient collaboration combinations based on historical collaboration data of devices, resulting in low resource collaboration efficiency. SUMMARY

[0007] The present application mainly solves the problem of how to effectively plan tasks for unmanned swarm based on different task types and requirements. The present application discloses a task planning method and device for unmanned swarm.

[0008] In a first aspect, the present application discloses a task planning method for unmanned swarm, comprising:

[0009] S1, acquire a task information set to be completed and a capability index information set of the unmanned cluster; the capability index information set includes a capability index value set of each unmanned target of the unmanned cluster; the task information set to be completed includes an index demand urgency set, an index demand value set and an index loss value set of each task to be completed;

[0010] S2, classify the capability index information set of the unmanned cluster to obtain a plurality of sub-cluster index sets; the sub-cluster index set includes a capability index value set of all unmanned targets included in the sub-cluster;

[0011] S3, perform multi-level task planning processing on the task information set to be completed and the sub-cluster index set to obtain a task allocation information set.

[0012] The classification processing of the capability index information set of the unmanned cluster to obtain a plurality of sub-cluster index sets includes:

[0013] S21, calculate the number of sub-clusters of the capability index information set of the unmanned cluster to obtain the total number of sub-clusters;

[0014] S22, based on the total number of sub-clusters, classify and process the capability index information set of the unmanned cluster to obtain the sub-cluster index set.

[0015] The calculation of the number of sub-clusters of the capability index information set of the unmanned cluster to obtain the total number of sub-clusters includes:

[0016] S211, represent the capability index information set of the unmanned cluster as a cluster capability index matrix; the row vector of the cluster capability index matrix is a vector composed of a capability index value set of one unmanned target; the element of the vector is each capability index value in the capability index value set;

[0017] S212, calculate the rank value a1 of the cluster capability index matrix;

[0018] S213, multiply the cluster capability index matrix by its transpose to obtain a first matrix;

[0019] S214, calculate the eigenvalue set of the first matrix;

[0020] S215, evaluate the value distribution of the eigenvalue set to obtain a distribution evaluation value;

[0021] S216, determine the greatest common divisor of the distribution evaluation value and the rank value a1 as the total number of sub-clusters M.

[0022] The calculation expression of the value distribution evaluation is:

[0023]

[0024] wherein a2 is a distribution evaluation value, N is the number of characteristic values in the characteristic value set, and a, β, and γ are the mean, variance, and median value of all characteristic values in the characteristic value set, respectively, represents rounding up.

[0025] The classification and calculation processing of the capability index information set of the unmanned swarm based on the total number of sub-swarm is performed to obtain a sub-swarm index set, including:

[0026] S221, the capability index value set of each unmanned target in the capability index information set of the unmanned swarm is expressed as a capability vector; the element of the capability vector is each capability index value in the capability index value set;

[0027] S222, clustering analysis is performed on all capability vectors to obtain category information and capability vectors contained in each category information;

[0028] S223, the category information whose contained capability vector total number value size ranking is the first M is determined as M sub-swarm information; the unmanned target corresponding to the capability vector contained in the M category information is determined as the unmanned target contained in the M sub-swarm information;

[0029] S224, other capability vectors except the capability vectors contained in the M category information in all capability vectors are respectively subjected to attribution discrimination to obtain the category information to which the capability vector belongs, and the sub-swarm information of the unmanned target corresponding to the capability vector is determined according to the category information to which the capability vector belongs;

[0030] S225, the capability index value set of the unmanned target contained in all sub-swarm information is used to construct a sub-swarm index set; one sub-swarm index set corresponds to one sub-swarm.

[0031] The multi-level task planning processing of the to-be-completed task information set and the sub-swarm index set is performed to obtain a task allocation information set, including:

[0032] S31, the to-be-completed task information set and the sub-swarm index set are subjected to matching processing to obtain a sub-swarm matched with each to-be-completed task;

[0033] S32, task allocation is performed on the sub-swarm matched with each to-be-completed task to obtain a task allocation matrix of the sub-swarm;

[0034] S33, the task allocation information set is constructed by using the task allocation matrix of all sub-swarms.

[0035] The matching processing of the to-be-completed task information set and the sub-cluster index set obtains a sub-cluster matched with each to-be-completed task, and the matching processing comprises the following steps:

[0036] In S311, for each sub-cluster index set, the capability index value set of each unmanned target included in the sub-cluster index set is expressed as an index vector; and an element of the index vector is a capability index value in the capability index value set.

[0037] In S312, geometric center vector calculation is performed on all index vectors of the sub-cluster index set to obtain a corresponding center vector.

[0038] In S313, the index requirement urgency set of each to-be-completed task in the to-be-completed task information set is expressed as a corresponding urgency vector; and an element of the urgency vector is an urgency value of each index in the index requirement urgency set.

[0039] In S314, for each to-be-completed task, a vector distance between the urgency vector and the center vectors of all sub-cluster index sets is calculated; and a sub-cluster corresponding to a sub-cluster index set with a minimum vector distance is determined as a sub-cluster matched with the to-be-completed task.

[0040] In a second aspect of the embodiment of the present application, a task planning device of an unmanned cluster is disclosed, and the device comprises:

[0041] A memory in which an executable program code is stored;

[0042] A processor coupled with the memory;

[0043] The processor invokes the executable program code stored in the memory to execute the task planning method of the unmanned cluster.

[0044] In a third aspect of the embodiment of the present application, a computer storage medium is disclosed, and the computer storage medium stores computer instructions; when the computer instructions are invoked by a computer, the computer instructions are used to execute the task planning method of the unmanned cluster.

[0045] In a fourth aspect of the embodiment of the present application, an information data processing terminal is disclosed, and the information data processing terminal is used to implement the task planning method of the unmanned cluster.

[0046] The present application has the following advantages:

[0047] The present application significantly improves the rationality and execution efficiency of the task planning of the unmanned cluster by using the innovative sub-cluster dynamic division and multi-level task planning strategy, and has the following specific advantages:

[0048] 1. Multi-dimensional capability precision classification capability of unmanned cluster performance; by converting capability index information into matrix form (S211), combining matrix rank (a1) and eigenvalue distribution evaluation (a2) to determine the total number of sub-clusters (S212-S216), fully considering the coupling relationship of multi-dimensional capabilities of unmanned equipment in communication, perception, execution, etc. For example, the rank value a1 reflects the linear independence of the capability index, and the distribution evaluation value a2 combines the sine function with the statistics (mean α, variance β, median γ) to quantify the concentration trend and dispersion degree of the eigenvalue distribution, avoiding classification bias caused by artificial preset cluster number.

[0049] 2. Self-adaptive clustering optimization resource allocation capability; based on the total number of sub-clusters M, perform clustering analysis (S221-S225), preferentially select the first M classes of capability vector size as the core sub-cluster, and dynamically allocate the remaining devices to the optimal sub-cluster through attribution discrimination. This strategy can effectively balance the resource load of each sub-cluster, for example, in the logistics distribution task, long-endurance and high-load devices are aggregated into a "long-distance transportation sub-cluster", and short-distance flexible devices form an "end distribution sub-cluster", improving the overall distribution efficiency.

[0050] 3. Multi-index dynamic matching capability: combined with the urgency, demand value and loss value of task index demand (S1), and multi-level matching with sub-cluster capability index (S31), to realize precise docking of task and device capability. For example, for tasks with high time efficiency requirements, preferentially match sub-clusters with strong communication capabilities and fast response speeds; for high-precision tasks, allocate sub-clusters with high perception accuracy to execute, reducing the risk of task failure. By constructing a sub-cluster task allocation matrix (S32), the priority and resource consumption of each sub-cluster in undertaking tasks are quantified, avoiding resource waste caused by traditional "one-size-fits-all" allocation. For example, using the Hungarian algorithm or auction algorithm to solve the task allocation matrix can minimize the total cost of task execution (such as energy consumption, time), improving the overall energy efficiency of the cluster.

[0051] 4. Improve computational efficiency and scalability: the sub-cluster division method based on matrix operations and statistical characteristics reduces the computational complexity from exponential to polynomial compared to traditional global search algorithms, meeting the real-time planning needs of large-scale clusters. At the same time, modular design supports dynamic addition or removal of devices or tasks. When new devices join, only the capability index matrix needs to be updated for quick reclassification without the need to retrain the model; by matching task index loss value with sub-cluster capability, task execution risk can be predicted in advance. For example, when the estimated loss of a sub-cluster in undertaking a task exceeds its capability threshold, the system automatically triggers task redistribution to avoid device overload damage or task interruption. BRIEF DESCRIPTION OF DRAWINGS

[0052] Figure 1 The flowchart of the method of the present application. DETAILED DESCRIPTION

[0053] In order to better understand the content of the present invention, an embodiment is given here.

[0054] Figure 1 4 is an implementation flow chart of the method of the present invention.

[0055] In a first aspect, an embodiment of the present invention discloses a task planning method for an unmanned swarm, comprising:

[0056] S1, obtaining a set of information on tasks to be completed and a set of capability indicator information of an unmanned cluster; the capability indicator information set includes a set of capability indicator values ​​of each unmanned target in the unmanned cluster; the set of information on tasks to be completed includes a set of indicator urgency, an indicator requirement value, and an indicator loss value for each task to be completed;

[0057] S2, classifying the capability indicator information set of the unmanned cluster to obtain a plurality of sub-cluster indicator sets; the sub-cluster indicator sets include capability indicator value sets of all unmanned targets included in the sub-cluster;

[0058] S3, performing multi-level task planning processing on the to-be-completed task information set and the sub-cluster indicator set to obtain a task allocation information set;

[0059] The capability indicator information set of the unmanned cluster is classified to obtain several sub-cluster indicator sets, including:

[0060] S21, calculating the number of subclusters for the capability indicator information set of the unmanned cluster to obtain the total number of subclusters;

[0061] S22 , based on the total number of subclusters, classify and calculate the capability indicator information set of the unmanned cluster to obtain a subcluster indicator set.

[0062] The calculating the number of subclusters on the capability indicator information set of the unmanned cluster to obtain the total number of subclusters includes:

[0063] S211: Represent the capability indicator information set of the unmanned cluster as a cluster capability indicator matrix; a row vector of the cluster capability indicator matrix is ​​a vector consisting of a capability indicator value set of an unmanned target; an element of the vector is each capability indicator value in the capability indicator value set;

[0064] S212, calculating and obtaining the rank value a1 of the cluster capability indicator matrix;

[0065] S213, multiplying the cluster capability indicator matrix and its transformed rank matrix to obtain a first matrix;

[0066] S214, calculating and obtaining a set of eigenvalues ​​of the first matrix;

[0067] S215, performing a value distribution evaluation on the feature value set to obtain a distribution evaluation value;

[0068] S216: Determine the greatest common divisor of the distribution evaluation value and the rank value a1, which is the total number M of subclusters.

[0069] The calculation expression for the value distribution evaluation is:

[0070]

[0071] Where a2 is the distribution evaluation value, N is the number of eigenvalues ​​in the eigenvalue set, α, β and γ are the mean, variance and median values ​​of all eigenvalues ​​in the eigenvalue set respectively, Indicates rounding up.

[0072] The computational expression for evaluating the distribution of values ​​combines the statistical properties of eigenvalues ​​(the mean reflects the concentration trend of the data, the variance measures the degree of dispersion, and the median enhances robustness) with trigonometric functions and ratio operations to dynamically quantify the distribution of eigenvalues. For example, when the eigenvalue distribution is concentrated, the calculation tends to produce fewer subclusters to avoid over-division; conversely, when the distribution is discrete, the number of subclusters is increased to achieve refined classification. Rounding up ensures that the number of subclusters is an integer, adapting to actual task requirements.

[0073] By introducing the median as a calculation parameter, the impact of abnormal eigenvalues ​​(such as outliers caused by sensor noise) on the results can be effectively suppressed compared to relying solely on the mean and variance. For example, in complex electromagnetic environments, equipment capability indicator data may contain wild values. The use of the median makes the calculation results of a2 closer to the actual distribution, ensuring the accuracy of subcluster division. Balance classification granularity and computational efficiency: The number of eigenvalues ​​N is directly involved in the calculation, making the number of subclusters positively correlated with the data size: the larger the data volume (the larger N), the more the calculation results tend to increase the number of subclusters and improve the classification granularity; at the same time, avoid the surge in computational complexity caused by overly detailed division, and achieve a balance between classification accuracy and computational efficiency in large-scale cluster scenarios.

[0074] Based on the total number of subclusters, the capability indicator information set of the unmanned cluster is classified and calculated to obtain a subcluster indicator set, including:

[0075] S221, representing the capability index value set of each unmanned target in the capability index information set of the unmanned cluster as a capability vector; an element of the capability vector is each capability index value in the capability index value set;

[0076] S222, performing cluster analysis on all the capability vectors to obtain category information and capability vectors contained in each category information;

[0077] S223, determining category information whose contained capability vectors are ranked as the first M, as M subset cluster information; determining unmanned targets corresponding to the capability vectors contained in the M category information, as unmanned targets contained in the M subset cluster information;

[0078] S224, performing attribution discrimination on other capability vectors except the capability vectors contained in the M category information in all the capability vectors, respectively, to obtain the category information to which the capability vectors belong, and determining the subset cluster information of the unmanned target corresponding to the capability vectors according to the category information to which the capability vectors belong;

[0079] S225, constructing a subset cluster index set by using the capability index value set of the unmanned targets contained in all the subset cluster information. One subset cluster index set corresponds to one subset cluster.

[0080] One subset cluster index set corresponds to the capability index value set of the unmanned targets contained in one subset cluster information.

[0081] One subset cluster information, namely a subset cluster, includes a plurality of unmanned targets.

[0082] The multi-level task planning processing of the to-be-completed task information set and the subset cluster index set to obtain the task allocation information set includes:

[0083] S31, performing matching processing on the to-be-completed task information set and the subset cluster index set to obtain a subset cluster matched with each to-be-completed task.

[0084] S32, performing task allocation on the subset cluster matched with each to-be-completed task to obtain a task allocation matrix of the subset cluster.

[0085] S33, constructing a task allocation information set by using the task allocation matrix of all the subset clusters.

[0086] The matching processing of the to-be-completed task information set and the subset cluster index set to obtain a subset cluster matched with each to-be-completed task includes:

[0087] S311, for each subset cluster index set, representing the capability index value set of each unmanned target included in the subset cluster index set as an index vector; the element of the index vector is a capability index value in the capability index value set.

[0088] S312, performing geometric center vector calculation on all the index vectors of the subset cluster index set to obtain a corresponding center vector.

[0089] S313, the indicator demand urgency set of each to-be-completed task in the to-be-completed task information set is represented as a corresponding urgency vector; an element of the urgency vector is an urgency value of each indicator in the indicator demand urgency set;

[0090] S314, for each to-be-completed task, a vector distance between the urgency vector thereof and a center vector of all subset cluster indicator sets is calculated; and a subset cluster of a subset cluster indicator set corresponding to a smallest vector distance is determined as a subset cluster matched with the to-be-completed task;

[0091] The expression of the geometric center vector calculation is as follows:

[0092]

[0093] wherein a1 j is the jth element of the center vector, a ij is the jth element of the ith indicator vector of a subset cluster indicator set, and n is the total number of indicator vectors included in the subset cluster indicator set.

[0094] The geometric mean (root of product) is used to calculate the center vector, which can reflect the synergistic relationship and comprehensive level of each device capability indicator in the subset cluster more than the arithmetic mean. For example, in a surveying and mapping task, if a subset cluster includes a high-precision camera device (high-resolution indicator) and a long-endurance unmanned aerial vehicle (high-endurance indicator), the geometric mean can avoid the short board of other indicators (such as resolution) being covered by a single strong indicator (such as endurance), and ensure that the center vector fully represents the capability characteristics of the subset cluster.

[0095] The vector distance can be obtained by calculating the Frobenius norm of the difference vector of two vectors, or by using the Hamming distance and Jaccard coefficient to calculate.

[0096] The task allocation is performed on each subset cluster matched with the to-be-completed task to obtain a task allocation matrix of the subset cluster, including:

[0097] S321, for each subset cluster matched with the to-be-completed task, a corresponding task allocation model is constructed;

[0098] The expression of the task allocation model is as follows:

[0099]

[0100] wherein f(C) is an objective function, C is a task allocation matrix, c ik is an element of the ith row and the kth column of the matrix C, which takes a value of 0 or 1, and c ikWhen the value is 0, it indicates that the kth unmanned target of the sub-cluster does not participate in the ith step of the task to be completed, c ik When the value is 1, it indicates that the kth unmanned target of the sub-cluster participates in the ith step of the task to be completed, d kj sj represents the jth capability index value of the kth unmanned target of the sub-cluster, s i (C) and g i (C) respectively represent the loss sub-function and the benefit sub-function of the ith step of the task to be completed, N1 is the total number of steps, K is the total number of unmanned targets of the sub-cluster, M1 is the total number of capability index values, x ij fij represents the jth index demand value of the ith step of the task to be completed, f ij gij represents the jth index loss value of the ith step of the task to be completed.

[0101] S322, the task allocation model corresponding to the sub-cluster matched for each task to be completed is solved, and a task allocation matrix of the sub-cluster is obtained.

[0102] The task allocation model corresponding to the sub-cluster matched for each task to be completed can be solved by using an ant colony algorithm, a particle filtering algorithm, a wolf swarm algorithm, etc.

[0103] The index loss value and the index demand value of each step can be respectively obtained from the index loss value set and the index demand value set of each task to be completed.

[0104] The unmanned target is an unmanned device, including an unmanned vehicle, a drone, etc.

[0105] In a second aspect of the embodiment of the present application, a task planning device of an unmanned cluster is disclosed, and the device comprises:

[0106] A memory storing executable program codes;

[0107] A processor coupled with the memory;

[0108] The processor invokes the executable program codes stored in the memory to execute the task planning method of the unmanned cluster.

[0109] In a third aspect of the embodiment of the present application, a computer storage medium is disclosed, and the computer storage medium stores computer instructions, which are invoked by a computer to execute the task planning method of the unmanned cluster.

[0110] In a fourth aspect of the embodiment of the present application, an information data processing terminal is disclosed, which is used to implement the task planning method of the unmanned cluster.

[0111] The above merely illustrates the embodiments of the present application but should not be taken as limitations. Various changes and modifications can be made by those skilled in the art. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application should be included in the scope of claims of the present application.

Claims

1. A task planning method for unmanned swarms, characterized in that: include: S1, obtaining a set of information on tasks to be completed and a set of capability indicator information of an unmanned cluster; the capability indicator information set includes a set of capability indicator values ​​of each unmanned target in the unmanned cluster; the set of information on tasks to be completed includes a set of indicator urgency, an indicator requirement value, and an indicator loss value for each task to be completed; S2, classifying the capability indicator information set of the unmanned cluster to obtain a plurality of sub-cluster indicator sets; the sub-cluster indicator sets include capability indicator value sets of all unmanned targets included in the sub-cluster; S3, performing multi-level task planning processing on the to-be-completed task information set and the sub-cluster indicator set to obtain a task allocation information set.

2. The unmanned swarm task planning method according to claim 1, wherein: The capability indicator information set of the unmanned cluster is classified to obtain several sub-cluster indicator sets, including: S21, calculating the number of subclusters for the capability indicator information set of the unmanned cluster to obtain the total number of subclusters; S22 , based on the total number of subclusters, classify and calculate the capability indicator information set of the unmanned cluster to obtain a subcluster indicator set.

3. The unmanned swarm task planning method according to claim 1, wherein: The calculating the number of subclusters on the capability indicator information set of the unmanned cluster to obtain the total number of subclusters includes: S211: Represent the capability indicator information set of the unmanned cluster as a cluster capability indicator matrix; a row vector of the cluster capability indicator matrix is ​​a vector consisting of a capability indicator value set of an unmanned target; an element of the vector is each capability indicator value in the capability indicator value set; S212, calculating and obtaining the rank value a1 of the cluster capability indicator matrix; S213, multiplying the cluster capability indicator matrix and its transformed rank matrix to obtain a first matrix; S214, calculating and obtaining a set of eigenvalues ​​of the first matrix; S215, performing a value distribution evaluation on the feature value set to obtain a distribution evaluation value; S216: Determine the greatest common divisor of the distribution evaluation value and the rank value a1, which is the total number M of subclusters.

4. The unmanned swarm task planning method according to claim 3, wherein: The calculation expression for the value distribution evaluation is: Where a2 is the distribution evaluation value, N is the number of eigenvalues ​​in the eigenvalue set, α, β and γ are the mean, variance and median values ​​of all eigenvalues ​​in the eigenvalue set respectively, Indicates rounding up.

5. The unmanned swarm task planning method according to claim 2, wherein: Based on the total number of subclusters, the capability indicator information set of the unmanned cluster is classified and calculated to obtain a subcluster indicator set, including: S221, representing the capability index value set of each unmanned target in the capability index information set of the unmanned cluster as a capability vector; an element of the capability vector is each capability index value in the capability index value set; S222, performing cluster analysis on all capability vectors to obtain category information and capability vectors contained in each category information; S223: Determine the categories that are ranked as the top M in terms of the total number of capability vectors, as M sub-clusters; and determine the unmanned targets corresponding to the capability vectors included in the M categories, as the unmanned targets included in the M sub-clusters. S224: performing attribution discrimination on all capability vectors other than the capability vectors included in the M category information to obtain category information to which the capability vectors belong, and determining subcluster information of unmanned targets corresponding to the capability vectors based on the category information to which the capability vectors belong; S225 , constructing a subcluster indicator set using the capability indicator value set of the unmanned targets included in all subcluster information; one subcluster indicator set corresponds to one subcluster.

6. The unmanned swarm mission planning method according to claim 1, wherein: The multi-level task planning process is performed on the to-be-completed task information set and the sub-cluster indicator set to obtain a task allocation information set, including: S31, matching the to-be-completed task information set and the sub-cluster indicator set to obtain a sub-cluster matching each to-be-completed task; S32, assigning tasks to each subcluster that matches the task to be completed, and obtaining a task assignment matrix for the subcluster; S33, using the task allocation matrices of all subclusters, a task allocation information set is constructed.

7. The unmanned swarm mission planning method according to claim 1, wherein: The matching process of the to-be-completed task information set and the subcluster indicator set to obtain a subcluster matching each to-be-completed task includes: S311, for each sub-cluster indicator set, representing a capability indicator value set of each unmanned target included in the sub-cluster indicator set as an indicator vector; elements of the indicator vector are capability indicator values ​​in the capability indicator value set; S312, performing geometric center vector calculation on all indicator vectors of the sub-cluster indicator set to obtain corresponding center vectors; S313, representing the indicator requirement urgency set of each task to be completed in the set of task information to be completed as a corresponding urgency vector; the elements of the urgency vector are the urgency values ​​of each indicator in the indicator requirement urgency set; S314 , for each task to be completed, calculate the vector distance between its urgency vector and the central vector of all sub-cluster indicator sets; determine the sub-cluster of the sub-cluster indicator set corresponding to the minimum vector distance as the sub-cluster matching the task to be completed.

8. A mission planning device for unmanned swarms, characterized in that: The device comprises: a memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the task planning method for the unmanned swarm according to any one of claims 1 to 7.

9. A computer storable medium, characterized in that The computer storable medium stores computer instructions, and when the computer instructions are called by a computer, they are used to execute the unmanned swarm mission planning method according to any one of claims 1 to 7.

10. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the unmanned swarm mission planning method according to any one of claims 1 to 7.

Images