A method and system for determining a grouping mode of an unmanned cluster formation

Abstract

The application provides a method and system for identifying the grouping mode of an unmanned cluster formation, and belongs to the field of automatic control. The application extracts a cluster feature description operator to calculate the local motion coordination similarity of the cluster and establish a rule base based on space-time and speed correlation. Further, the cluster formation is randomly divided into two clusters, and each cluster is disturbed in sequence. Finally, the cluster individual motion consistent order degree, i.e. the cluster similarity, is calculated, compared with the rule base, and the unmanned cluster grouping mode is identified. The application provides a feasible method for identifying the grouping mode of an unmanned cluster system, can conveniently and efficiently deduce the grouping mode of the unmanned cluster formation, and provides common mechanisms and available technical support for algorithm confrontation of diversified unmanned cluster grouping modes.

Description

Technical Field

[0001] This invention relates to the field of automatic control technology, specifically to a method and system for determining the grouping pattern of unmanned swarm formations. Background Technology

[0002] The significant advantages of unmanned swarm combat systems, such as "autonomous coordination" and "capability emergence," are inseparable from their efficient swarming and collaborative formation methods. For swarm systems, analyzing the formation mechanism of the opposing swarm through key technologies for swarming mode identification, discerning its overall formation or control method, and then dynamically interfering with it to confuse and disrupt its actions, finding weaknesses in its control methods, and thus achieving efficient swarm confrontation under specific swarm control strategies, is crucial for unmanned swarm system countermeasures research. Swarming mode reasoning refers to determining the swarming mode (mainly including centralized, decentralized, and distributed types) based on the characteristics of swarm intelligence exhibited under different external information input conditions, such as formation, movement, posture, and actions. This further enables the discovery and identification of key points and critical links in the alignment structure. However, under normal conditions, the differences in swarm activities exhibited by different swarming modes of swarm intelligence are not obvious, especially for distributed swarming modes driven by algorithms such as biological swarm intelligence optimization and multi-agent reinforcement learning, which are even more difficult to distinguish. Summary of the Invention

[0003] To address the aforementioned issues, this invention provides a method and system for identifying the swarming patterns of unmanned swarms, offering common mechanisms and usable technical support for algorithmic countermeasures against diverse unmanned swarming patterns.

[0004] To achieve the above objectives, the technical solution of the method of the present invention is as follows:

[0005] A method for determining the grouping mode of unmanned swarm formations includes the following steps:

[0006] Step 1: By analyzing the structure of different clustering methods, construct a cluster consistency evaluation index system and establish a rule base;

[0007] Step 2: During the swarm formation flight phase, the swarm is randomly divided into two clusters, and interference is carried out in sequence to disrupt the swarm formation; real-time data of the unmanned swarm formation is obtained, and the cluster orderliness measure and cluster similarity are calculated based on the unmanned swarm formation data;

[0008] Step 3: Compare the cluster orderliness measure and cluster similarity from Step 2 with the rule base to determine the cluster formation method.

[0009] Furthermore, the rule base established in step 1 is as follows:

[0010]

[0011] Where L1 represents the similarity of the unperturbed clusters during the first passive handover, and L2 represents the similarity of the unperturbed clusters during the second passive handover.

[0012] Furthermore, in step 2, during the swarm formation flight phase, the swarm is randomly divided into two clusters, and interference is carried out sequentially to disrupt the swarm formation; real-time data on the unmanned swarm formation is obtained, including the swarm size, individual positions, and speeds.

[0013] Furthermore, in step 2, the cluster ordering measure O(i) of individual i in the cluster at time t is... t The representation is as follows:

[0014]

[0015] Where N(i) represents the set of neighboring individuals of individual i, that is, the number of individuals in the cluster to which individual i belongs, ω t (i,j) is the correlation coefficient of the interaction force between individuals i and j at time t.

[0016] Furthermore, the correlation coefficient ω of the interaction force between individuals i and j at time t. t The calculation form of (i,j) is as follows:

[0017]

[0018] Among them, F i,j Let be the interaction force between individuals i and j in the cluster, where j is the set of neighboring individuals of individual i.

[0019] Furthermore, the interaction force F between individuals i and j in the cluster i,j The calculation form is as follows:

[0020]

[0021] in, θ represents the velocity of an individual; θ represents the angle between the velocity direction of individual i and the direction of the force between individuals i and j; N(i) represents the set of neighboring individuals of individual i, that is, the number of individuals in the cluster to which individual i belongs.

[0022] Furthermore, the cluster similarity L is represented as follows:

[0023] L i =O t' -O t” ,i∈n,

[0024] Where n is the number of clusters in the cluster, which is 2; L1 represents the similarity of the undisturbed clusters during the first passive switch; L2 represents the similarity of the undisturbed clusters during the second passive switch; t' is the moment before the cluster implements the passive switch; t” is the moment when the cluster re-forms a formation after the passive switch; and O represents the measure of the orderliness of individual clusters in the cluster.

[0025] Furthermore, step 3 involves determining the clustering mode using a simulation testing system, specifically including:

[0026] Step 3.1: Initialize the unmanned swarm formation test dataset, including the initial positions of individuals. initial velocity Where i = 1, 2, ..., I, I represents the number of individuals; and the swarming method used by the unmanned cluster is randomly assigned.

[0027] Step 3.2: Randomly divide the cluster into two clusters. During the cluster's formation movement, sequentially perturb different clusters. The second perturbation occurs when the cluster re-forms after the first perturbation. The positions of individual cluster members before and after the perturbation are obtained using a simulation testing system. speed information;

[0028] Step 3.3: Based on the obtained position and velocity information of the cluster before and after the disturbance, calculate the cluster similarity and compare it with the rule base to determine the cluster formation method category.

[0029] On the other hand, the present invention provides a system for determining the grouping mode of unmanned swarm formations, comprising:

[0030] Rule base establishment module: It is used to construct a cluster consistency evaluation index system and establish a rule base by analyzing the structure of different clustering methods;

[0031] Rule base establishment module: It is used to construct a cluster consistency evaluation index system and establish a rule base by analyzing the structure of different clustering methods;

[0032] The calculation module is used to randomly divide the swarm into two clusters during the swarm formation flight phase, and then interfere with each cluster in sequence to disrupt the swarm formation; it also acquires unmanned swarm formation data in real time and calculates the cluster orderliness measure and cluster similarity based on the unmanned swarm formation data.

[0033] The discrimination module is used to compare the cluster ordering measure and cluster similarity in the calculation module with the rule base to determine the cluster formation mode.

[0034] The unmanned swarm formation grouping mode discrimination system is used to execute the steps in the above-described unmanned swarm formation grouping mode discrimination method.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] This invention proposes a novel metric for determining grouping patterns—cluster similarity—based on cluster orderliness measures. Cluster similarity, extended from cluster orderliness measures, is used to determine grouping patterns. Based on an established grouping rule base, the cluster similarity of test unmanned clusters is calculated multiple times and compared with the rule base to analyze grouping patterns. Using the method proposed in this invention, the grouping patterns of unmanned cluster formations can be inferred conveniently and efficiently, providing a common mechanism and usable technical support for algorithmic adversarial analysis of diverse unmanned cluster grouping patterns. Test results show that this invention can effectively infer the grouping patterns of unmanned clusters. Attached Figure Description

[0037] Figure 1 This is a flowchart of the unmanned cluster grouping method identification process of the present invention.

[0038] Figure 2 This is a schematic diagram of the test steps for the unmanned cluster grouping method discrimination test of the present invention.

[0039] Figure 3 This is a three-dimensional motion trajectory diagram of a cluster distributed formation under the unmanned cluster grouping mode discrimination method provided in this invention example.

[0040] Figure 4 This is a graph showing the formation position-time variation of a cluster distributed under the unmanned cluster grouping method provided in this invention example.

[0041] Figure 5 This is a speed-time variation graph of the unmanned cluster grouping method provided in this invention example for determining the grouping mode of the cluster.

[0042] Figure 6 This is a three-dimensional motion trajectory diagram of a clustered centralized formation under the unmanned cluster grouping mode discrimination method provided in this invention example.

[0043] Figure 7 This is a graph showing the formation position-time variation of a cluster in a centralized manner under the unmanned cluster grouping method provided in this invention example.

[0044] Figure 8 This is a speed-time variation graph of a clustered centralized formation under the unmanned cluster grouping method provided in this invention example.

[0045] Figure 9 This is a three-dimensional motion trajectory diagram of a clustered, distributed formation under the unmanned cluster grouping method provided in this invention example.

[0046] Figure 10 This is a graph showing the position-time variation of a cluster in a dispersed formation under the unmanned cluster grouping method provided in this invention example.

[0047] Figure 11 This is a velocity-time variation graph of a clustered dispersed formation under the unmanned cluster grouping method provided in this invention example. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0049] Example 1

[0050] This invention proposes a novel swarm formation determination metric for rule generation and consistency assessment—cluster similarity. Cluster similarity, an extension of cluster orderliness measurement, determines the swarm formation pattern. Cluster orderliness measurement refers to the degree to which individuals within a cluster maintain smooth movement with other individuals in their neighborhood, used to detect the consistent orderliness of movement among individuals in a group. Cluster similarity quantifies the consistency of different clusters within a cluster in following the movement of other neighboring clusters. The swarm formation determination method for unmanned swarms provided by this invention, based on a given swarm formation determination metric, has been tested multiple times to assess its effectiveness in determining the swarm formation pattern of unmanned swarms. The specific process and steps are as follows: Figure 1 and Figure 2 As shown, during the swarm formation's movement, the swarm is randomly divided into two clusters, and the different clusters are perturbed sequentially (disrupting the swarm formation). Based on the real-time acquired swarm formation information, cluster similarity is calculated, and the corresponding swarm formation is inferred from the cluster similarity judgment range. Specifically, in the simulation test, an unmanned swarm dataset is initialized, the swarm is randomly divided into two clusters, and the different clusters are perturbed sequentially (disrupting the swarm formation) during the swarm's movement. Relying on the test system to acquire the swarm's position and velocity information in real time, the cluster similarity is calculated and output, thereby determining the corresponding swarm formation.

[0051] The following example, using a grouping method of six intelligent agents, illustrates the detailed implementation process of this invention.

[0052] First, define the criteria for determining the clustering method;

[0053] The force F between individuals i and j in the cluster i,j The calculation form is as follows:

[0054]

[0055] in, θ represents the velocity of an individual; θ represents the angle between the velocity direction of individual i and the direction of the force between individuals i and j; N(i) represents the set of neighboring individuals of individual i, that is, the number of individuals in the cluster to which individual i belongs.

[0056] The correlation coefficient of the interaction force between individuals i and j at time t is calculated as follows:

[0057]

[0058] The cluster ordering measure O for individual i at time t in the cluster is represented as follows:

[0059]

[0060] The cluster similarity L is represented as follows:

[0061] L i =O t' -O t” ,i∈n,

[0062] Where n is the number of clusters divided into groups, which is 2; L1 represents the similarity of the undisturbed clusters during the first passive handover; L2 represents the similarity of the undisturbed clusters during the second passive handover; t' is the moment before the passive handover is implemented; and t” is the moment when the cluster re-forms into a formation after the passive handover.

[0063] As a preferred approach, the rule base for inferring cluster formation methods based on cluster similarity is as follows:

[0064]

[0065] The cluster formation method category is determined based on the cluster similarity range given in the table above.

[0066] Then, based on the given clustering method determination scale, the clustering method recognition performance is tested multiple times; including the following specific steps:

[0067] S1, Initialize the unmanned swarm formation test dataset; initial positions of the agents. initial velocity Information such as i = 1, 2, ..., 6; and the grouping method used by the unmanned cluster is randomly assigned.

[0068] S2, the cluster is randomly divided into two clusters. During the cluster's formation movement, different clusters are perturbed sequentially. The second perturbation occurs when the cluster re-forms after the first perturbation. The positions of individual cluster members before and after the perturbation are obtained using a simulation testing system. speed information;

[0069] S3. Based on the position and velocity information of the cluster before and after the disturbance, calculate the cluster similarity and determine the grouping method category used by the unmanned cluster.

[0070] In this embodiment, Python 3.6 is used as the simulation software to test the grouping method of unmanned swarm formation.

[0071] Test dataset 1 is used to test the swarm reasoning technology for rule generation and consistency evaluation of unmanned swarms. The test data contains 100 sets of relevant data on different swarming methods of swarm formation. Among them, there are 33 test data for distributed swarming, 33 test data for centralized swarming, and 34 test data for decentralized swarming. Each test data includes the initial position uniform(0,18), initial velocity uniform(0,2), and cluster number of the disturbed agent for 6 agents. Considering the influence of terrain height, the initial position of the agent on the Z-axis is uniformly distributed uniform(0,18)+156.

[0072] Figure 3 This indicates that the cluster uses a distributed three-dimensional motion trajectory graph. Figure 4 This is a graph showing the positional changes of the distributed cluster formation across various dimensions throughout the entire testing process. Figure 5 The graph shows the velocity changes of the distributed cluster formation across various dimensions throughout the entire test. In this case, the cluster similarity output is L1 = 0.7357 and L2 = 1.0847. According to the clustering rule base, this result indicates distributed control, which is consistent with the clustering method used.

[0073] Figure 6 This represents a three-dimensional motion trajectory diagram showing the grouping method used by the cluster in a centralized manner. Figure 7 This is a graph showing the positional changes of the centralized cluster formation across various dimensions throughout the entire testing process. Figure 8 The graph shows the velocity changes of the centralized cluster formation across various dimensions throughout the entire test. In this case, the cluster similarity output is L1 = 3.6481 and L2 = 0. According to the clustering rule base, the result indicates centralized control, which is consistent with the clustering method used.

[0074] Figure 9 This indicates that the clustering method used is a distributed three-dimensional motion trajectory diagram. Figure 10This is a graph showing the positional changes of the distributed cluster formation across various dimensions throughout the entire testing process. Figure 11 This is a graph showing the velocity changes of the distributed cluster formation across various dimensions throughout the entire test. In this case, the cluster similarity output is L1=0 and L2=0. According to the clustering rule base, the result is a distributed control, consistent with the clustering method used.

[0075] Example 2

[0076] This embodiment provides a system for determining the grouping mode of unmanned swarm formations, including:

[0077] Rule base establishment module: It is used to construct a cluster consistency evaluation index system and establish a rule base by analyzing the structure of different clustering methods;

[0078] Rule base establishment module: It is used to construct a cluster consistency evaluation index system and establish a rule base by analyzing the structure of different clustering methods;

[0079] The calculation module is used to randomly divide the swarm into two clusters during the swarm formation flight phase, and then interfere with each cluster in sequence to disrupt the swarm formation; it also acquires unmanned swarm formation data in real time and calculates the cluster orderliness measure and cluster similarity based on the unmanned swarm formation data.

[0080] The discrimination module is used to compare the cluster ordering measure and cluster similarity in the calculation module with the rule base to determine the cluster formation mode.

[0081] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0082] Obviously, those skilled in the art can make various modifications and variations to the embodiments of the present invention without departing from the spirit and scope of the embodiments of the present invention. Thus, if these modifications and variations to the embodiments of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention also intends to include these modifications and variations.

[0083] All other parts not described in detail are existing technologies.

Claims

1. A method for determining the grouping mode of unmanned swarm formations, characterized in that, The steps include the following: Step 1: By analyzing the structure of different clustering methods, construct a cluster consistency evaluation index system and establish a rule base; Step 2: During the swarm formation flight phase, the swarm is randomly divided into two clusters, and interference is performed sequentially to disrupt the swarm formation; real-time swarm formation data is acquired, and based on this data, the swarm orderliness measure and cluster similarity are calculated; individuals within the swarm... exist Cluster ordering measure at time The representation is as follows: in, Represents an individual The neighborhood set of individuals, that is, the individual The number of individuals in the cluster for individual moment and The correlation coefficient of the forces between them; Cluster similarity The representation of is: in The number of clusters to be divided into is 2. This represents the similarity of clusters that were not disturbed during the first passive handover. This indicates the similarity of the clusters that were not disturbed during the second passive handover; Just before the cluster performs a passive switchover, This is the moment when the cluster re-forms into a formation after a passive switch. This represents a measure of the cluster ordering of individuals within a cluster; Step 3: Compare the cluster orderliness measure and cluster similarity from Step 2 with the rule base to determine the cluster formation method.

2. The method for determining the grouping mode of unmanned swarm formations according to claim 1, characterized in that, The rule base established in step 1 includes clustering methods and their cluster similarity, wherein the clustering methods include decentralized, distributed, and centralized. The similarity of distributed clusters is: The similarity of distributed clusters is: The similarity of centralized clusters is ; in, This represents the similarity of clusters that were not disturbed during the first passive handover. This indicates the similarity of clusters that were not disturbed during the second passive handover.

3. The method for determining the grouping mode of unmanned swarm formations according to claim 1, characterized in that, The unmanned swarm formation data in step 2 includes swarm size, individual position, and speed.

4. The method for determining the grouping mode of unmanned swarm formations according to claim 1, characterized in that, individual moment and Correlation coefficient of the forces between The calculation form is as follows: in, For individuals in the cluster and The forces between them For individuals The set of neighboring individuals is the most important individual.

5. The method for determining the grouping mode of unmanned swarm formations according to claim 4, characterized in that, Individuals in a cluster and Interaction forces The calculation form is as follows: in, Indicates the speed of movement of an individual; Represents an individual The direction of the movement speed and the individual and The angle between the directions of the forces acting on each other; Represents an individual The neighborhood set of individuals, that is, the individual The number of individuals in the cluster.

6. The method for determining the grouping mode of unmanned swarm formations according to claim 1, characterized in that, Step 3 involves determining the clustering mode using a simulation testing system, specifically including: Step 3.1: Initialize the unmanned swarm formation test dataset, including the initial positions of individuals. Initial velocity ,in , I Indicate the number of individuals; and randomly assign the swarming method used by the unmanned cluster; Step 3.2: Randomly divide the cluster into two clusters. During the cluster's formation movement, sequentially perturb different clusters. The second perturbation occurs when the cluster re-forms after the first perturbation. The positions of individual cluster members before and after the perturbation are obtained using a simulation testing system. ,speed information; Step 3.3: Based on the obtained position and velocity information of the cluster before and after the disturbance, calculate the cluster similarity and compare it with the rule base to determine the cluster formation method category.

7. A system for determining the grouping pattern of unmanned swarm formations, characterized in that, include: Rule base establishment module: It is used to construct a cluster consistency evaluation index system and establish a rule base by analyzing the structure of different clustering methods; The calculation module is used to randomly divide the swarm into two clusters during the swarm formation flight phase, and then interfere with each cluster in sequence to disrupt the swarm formation; it also acquires unmanned swarm formation data in real time and calculates the cluster orderliness measure and cluster similarity based on the unmanned swarm formation data. The discrimination module is used to compare the cluster ordering measure and cluster similarity in the calculation module with the rule base to determine the cluster formation mode. The unmanned swarm formation grouping mode discrimination system is used to perform the steps in the unmanned swarm formation grouping mode discrimination method as described in any one of claims 1-6.

Images