Heterogeneous unmanned cluster task reliability modeling method based on evidence network

Abstract

According to the heterogeneous unmanned cluster task reliability modeling method based on the evidence network, uncertain information is expressed and fused by introducing the evidence network, the influence of uncertain factors in a heterogeneous unmanned cluster on a modeling result is reduced, and effective modeling of heterogeneous unmanned cluster task reliability is achieved. The method comprises the following steps: isomorphic grouping of heterogeneous unmanned clusters; constructing a heterogeneous unmanned cluster multi-layer cooperative capability framework, specifically comprising: constructing a task layer to determine task capability indexes, constructing a cooperative layer to determine formation capability indexes, and constructing a base layer to obtain base capability indexes; building an evidence network model based on a multi-layer cooperative capability framework; and evaluating the reliability of the heterogeneous unmanned cluster task based on evidence reasoning.

Description

[0001] Technical Field

[0002] This invention provides a reliability modeling method for heterogeneous unmanned cluster tasks based on evidence networks, belonging to the field of reliability engineering technology. Background Technology

[0004] Heterogeneous unmanned swarms refer to distributed, autonomous, and collaborative systems that integrate various unmanned platforms, such as drones, unmanned vehicles, and unmanned surface vessels, which differ significantly in type, function, or performance. This system has been widely applied in multiple fields, including civilian and military applications. In military scenarios, heterogeneous unmanned swarms, with their systemic collaborative advantages, have demonstrated outstanding combat effectiveness in complex missions such as joint reconnaissance and coordinated strikes, becoming a core support for enhancing mission execution capabilities. To ensure that heterogeneous unmanned swarms can stably and smoothly complete their assigned tasks and achieve expected goals in dynamic and complex environments, it is necessary to model their mission reliability.

[0005] However, traditional task reliability modeling methods mainly include analytical modeling methods based on structural logic and modeling methods based on probabilistic uncertainty. Analytical modeling methods based on structural logic typically rely on system composition relationships and fault propagation logic, using techniques such as FMEA / FMECA, fault tree analysis, and availability or risk assessment to decompose and analyze task success scenarios. However, their modeling flexibility and scalability are limited when dealing with systems with complex task structures and high degree of element coupling. Modeling methods based on probabilistic uncertainty describe system uncertainties and element relationships through probability statistics and Bayesian networks. They usually rely on sufficient historical data or reasonable prior assumptions. When sample data is insufficient or the assumptions are difficult to meet, the reliability and applicability of the model results are easily affected. Therefore, existing task reliability modeling methods still have certain limitations in heterogeneous unmanned swarm application scenarios.

[0006] To address this, the present invention proposes a heterogeneous unmanned cluster task reliability modeling method based on evidence networks. By introducing evidence networks to represent and fuse uncertain information, the method reduces the impact of uncertain factors in heterogeneous unmanned clusters on the modeling results, thereby achieving effective modeling of the reliability of heterogeneous unmanned cluster tasks. Summary of the Invention

[0008] The purpose of this invention is to provide a reliability modeling method for heterogeneous unmanned swarm tasks based on evidence networks, which overcomes the limitations of traditional reliability modeling methods in handling heterogeneity, multi-layered structures, and uncertain information.

[0009] To achieve the above objectives, the present invention provides the following technical solution:

[0010] A reliability modeling method for heterogeneous unmanned swarm tasks based on evidence networks mainly includes the following steps:

[0011] S100: Homogenized grouping of heterogeneous unmanned clusters;

[0012] S200: Constructing a multi-layered collaborative capability framework for heterogeneous unmanned clusters:

[0013] S201: Construct a task layer to define task capability indicators;

[0014] S202: Construct a coordination layer to determine formation capability metrics;

[0015] S203: Construct the base layer to obtain basic capability indicators;

[0016] S300: Building an evidence network model based on a multi-layered collaborative capability framework;

[0017] S400: Evidence-based reasoning for assessing the reliability of heterogeneous unmanned swarm missions.

[0018] In step S100, based on the differences in platform type, functional configuration, or performance parameters of the unmanned equipment, the heterogeneous unmanned cluster HEUSS is split into several homogeneous unmanned formations, namely:

[0019]

[0020] Among them, HOUSS n This represents any isomorphic unmanned formation, where N is the number of unmanned equipment types.

[0021] In step S200, based on the completion of the heterogeneous unmanned cluster splitting, a multi-layered collaborative capability framework is constructed from top to bottom around the overall mission objective, including: a task layer, a collaborative layer, and a basic layer.

[0022] In step S201, a task layer is constructed to describe the overall task completion capability R of the heterogeneous unmanned cluster;

[0023] In step S202, based on the mission objectives of each homogeneous unmanned formation, a collaboration layer is constructed to describe the mission capabilities of each homogeneous unmanned formation and its formation capability indicators. First, the mission completion capability R of the heterogeneous unmanned cluster is decomposed into the mission capability P of each homogeneous unmanned formation:

[0024]

[0025] Subsequently, the mission capability P of each isomorphic unmanned formation is decomposed into multiple formation capabilities L:

[0026]

[0027] Among them, Ln,m HOUSS n The arbitrary formation capability index, M is the HOUSS n The number of formation capability indicators.

[0028] In step S203, a foundational layer is constructed to describe the individual equipment capability indicators and basic capability indicators that support the realization of formation capabilities. First, the formation capability indicators L of the coordination layer are decomposed into the individual equipment capability indicators S of the relevant unmanned equipment, namely:

[0029]

[0030] Among them, S n,m,i L represents n,m Any single-unit capacity index, I is L n,m The number of single-unit capability indicators. Then, the single-unit task capability S of each unit is broken down into various basic capability indicators U, namely:

[0031] ,

[0032] Among them, U n,m,i,j S represents n,m,i Any basic capability index, J is S n,m,i The number of basic capability indicators.

[0033] In step S300, an evidence network is built from the bottom up based on the constructed multi-layer collaborative capability framework. Specifically, this includes: for the base layer, mapping the basic capability index U and the single-unit capability index S to the network leaf nodes and first-layer intermediate nodes respectively; for the collaboration layer, mapping the task capability index P of the homogeneous unmanned formation and the decomposed formation capability index L to the second and third-layer intermediate nodes respectively; and for the task layer, mapping the overall task completion capability R of the heterogeneous unmanned cluster to the root node of the evidence network. The constructed evidence network is a directed acyclic graph structure, and the connections between nodes are set based on the framework logic to describe the dependencies and transmission relationships between capabilities.

[0034] Furthermore, weights are assigned to the child nodes of any parent node. and satisfy Where d represents any child node, and D is the number of related child nodes contained in the parent node. Meanwhile, the identification framework for each node in the network is defined as follows: In order to complete the construction of the evidence network.

[0035] In step S400, observational evidence is input to the leaf nodes of the network. Through the reasoning and synthesis mechanism of the evidence network, capability information is passed upwards layer by layer, and finally, the credibility evaluation result of the overall task completion capability of the heterogeneous unmanned swarm is obtained at the root node. The task reliability of the heterogeneous unmanned swarm is defined as the trust level of the "meeting requirements" state of the root node, that is:

[0036]

[0037] Here, Bel() represents the trust function obtained from evidence network reasoning.

[0038] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention introduces evidence networks into the field of reliability modeling for heterogeneous unmanned swarm tasks, constructing a multi-layered collaborative capability framework. This framework comprehensively considers the heterogeneity and multi-level structure of heterogeneous unmanned swarms, providing a new theoretical perspective for reliability modeling of heterogeneous unmanned swarm tasks. Simultaneously, it reduces the impact of uncertainties in heterogeneous unmanned swarms on the modeling results, effectively solving the limitations of traditional modeling methods in handling uncertain information, and achieving effective modeling of the reliability of heterogeneous unmanned swarm tasks. Attached Figure Description

[0040] Figure 1 A flowchart of a heterogeneous unmanned cluster task reliability modeling method based on evidence networks provided by the present invention;

[0041] Figure 2 The heterogeneous UAV swarm multi-layer collaborative capability framework established for this invention;

[0042] Figure 3 This invention is based on an evidence network constructed using a multi-layered collaborative capability framework. Detailed Implementation

[0044] The following will refer to the appendix. Figure 1 Specific embodiments of the invention are described in detail below. While specific embodiments of the invention have been discussed, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, the embodiments are provided to enable a more thorough understanding of the invention and to fully convey the information of the invention to those skilled in the art.

[0045] This invention provides a method for reliability modeling of heterogeneous UAV swarm missions based on evidence networks, the flowchart of which is as follows. Figure 1 As shown, it includes:

[0046] S100: Based on the differences in platform type, functional configuration, or performance parameters of drones, the heterogeneous drone cluster HEUSS is split into several homogeneous drone clusters, namely:

[0047]

[0048] Among them, HOUSS n This represents any homogeneous drone cluster, where N is the number of drone types.

[0049] Example 1: Taking a heterogeneous drone swarm of four drones as an example for analysis, its overall mission background is to perform an area interception mission. Based on the differences in drone functions, the heterogeneous drone swarm HEUSS is decomposed into two homogeneous drone formations, namely:

[0050]

[0051] S200: Based on the completion of the heterogeneous UAV cluster splitting, a multi-layer collaborative capability framework is built from top to bottom around the overall mission objectives, including: mission layer, collaboration layer and basic layer.

[0052] S201: Construct a task layer to describe the overall task completion capability R of a heterogeneous drone swarm;

[0053] S202: Based on the mission objectives of each homogeneous UAV formation, a collaboration layer is constructed to describe the mission capabilities of each homogeneous UAV formation and its formation capability indicators. First, the mission completion capability R of the heterogeneous UAV swarm is decomposed into the mission capabilities P of each homogeneous UAV formation.

[0054]

[0055] Subsequently, the mission capability P of each isomorphic UAV formation is decomposed into multiple formation capabilities L:

[0056]

[0057] Among them, L n,m HOUSS n The arbitrary formation capability index, M is the HOUSS n The number of formation capability indicators.

[0058] S203: Construct a foundational layer to describe the individual equipment capability indicators and basic capability indicators that support the realization of formation capabilities. First, decompose each formation capability indicator L in the coordination layer into the individual equipment capability indicator S of the UAV, i.e.:

[0059]

[0060] Among them, S n,m,i L represents n,m Any single-unit capacity index, I is L n,m The number of single-unit capability indicators. Subsequently, the single-unit mission capability S of the UAV is broken down into various basic capability indicators U, namely:

[0061] ,

[0062] Among them, U n,m,i,j S represents n,m,iAny basic capability index, J is S n,m,i The number of basic capability indicators.

[0063] Continuing from the previous example, such as Figure 2 As shown, based on the completion of the heterogeneous drone cluster splitting, a multi-layered collaborative capability framework is built from top to bottom around the overall mission objective.

[0064] S300: Based on the constructed multi-layered collaborative capability framework, an evidence network is built from the bottom up. Specifically, for the base layer, the basic capability index U and the single-unit capability index S are mapped sequentially to the network leaf nodes and the first-layer intermediate nodes; for the collaboration layer, the task capability index P of the homogeneous UAV formation and the decomposed formation capability index L are mapped sequentially to the second and third-layer intermediate nodes; for the task layer, the overall task completion capability R of the heterogeneous UAV cluster is mapped to the root node of the evidence network. The constructed evidence network is a directed acyclic graph structure, and the connections between nodes are set based on the framework logic to describe the dependency and transmission relationships between capabilities.

[0065] Furthermore, weights are assigned to the child nodes of any parent node. and satisfy Where d represents any child node, and D is the number of related child nodes contained in the parent node. Meanwhile, the identification framework for each node in the network is defined as follows: In order to complete the construction of the evidence network.

[0066] Continuing from the previous example, such as Figure 3 As shown, based on the relationship between indicators at each level in the multi-layer collaborative capability framework of heterogeneous UAV clusters, the evidence network is constructed from the bottom up.

[0067] S400: Observational evidence is input to the leaf nodes. Through the reasoning and synthesis mechanism of the evidence network, capability information is passed upwards layer by layer, ultimately obtaining a credibility evaluation result of the overall mission completion capability of the heterogeneous UAV swarm at the root node. The mission reliability of the heterogeneous UAV swarm is defined as the trust level of the root node's "meets requirements" state, i.e.:

[0068] Continuing from the previous example, by inputting observational evidence into the leaf nodes of the network, the confidence level Bel (meets requirements) of the heterogeneous drone swarm meeting the regional interception capability is obtained at the root node, which is 0.55. That is, the overall task completion capability of the heterogeneous drone swarm R = Bel (meets requirements) = 0.05.

[0069] Although embodiments of the present invention have been described above in conjunction with the accompanying drawings, the present invention is not limited to the specific embodiments and application fields described above. The specific embodiments described above are merely illustrative and instructive, and not restrictive. Those skilled in the art can make many other forms based on the guidance of this specification and without departing from the scope of protection of the claims of the present invention, and all of these are within the scope of protection of the present invention.

Claims

1. A reliability modeling method for heterogeneous unmanned swarm tasks based on evidence networks, characterized in that, Includes: S100: Homogeneous grouping of heterogeneous unmanned swarms; based on the differences in platform type, functional configuration, or performance parameters of unmanned equipment, the heterogeneous unmanned swarm HEUSS is split into several homogeneous unmanned formations, namely: Among them, HOUSS n Let N represent any homogeneous unmanned formation, where N is the number of unmanned equipment types; S200: Construct a multi-layered collaborative capability framework for heterogeneous unmanned clusters; Based on the completion of the heterogeneous unmanned cluster decomposition, construct a multi-layered collaborative capability framework from top to bottom around the overall mission objective, including: a task layer, a collaboration layer, and a basic layer; S201: Construct the task layer to clarify task capability indicators; The task layer is constructed to describe the overall mission completion capability R of the heterogeneous unmanned cluster; S202: Construct the collaboration layer to determine formation capability indicators; Based on the mission objectives of each homogeneous unmanned formation, construct the collaboration layer to describe the mission capabilities of each homogeneous unmanned formation and its formation capability indicators; First, decompose the heterogeneous unmanned cluster mission completion capability R into the mission capabilities P of each homogeneous unmanned formation: Subsequently, the mission capability P of each isomorphic unmanned formation is decomposed into multiple formation capabilities L: Among them, L n,m HOUSS n The arbitrary formation capability index, M is the HOUSS n The number of formation capability indicators; S203: Construct a basic layer to obtain basic capability indicators; Construct a basic layer to describe the individual equipment capability indicators and basic capability indicators that support the realization of formation capability; First, decompose each formation capability indicator L of the coordination layer into the individual equipment capability indicators S of the relevant unmanned equipment, that is: Among them, S n,m,i L represents n,m Any single-unit capacity index, I is L n,m The number of single-unit capability indicators; subsequently, the single-unit task capability S of each unit is decomposed into various basic capability indicators U, namely: Among them, U n,m,i,j S represents n,m,i Any basic capability index, J is S n,m,i The number of basic capability indicators; S300: Building an evidence network model based on a multi-layer collaborative capability framework; Based on the constructed multi-layer collaborative capability framework, an evidence network is built from bottom to top; Specifically, for the basic layer, the basic capability indicator U and the single-assembly capability indicator S are mapped to the network leaf nodes and the first-layer intermediate nodes in sequence; for the collaborative layer, the task capability indicator P of the homogeneous unmanned formation and the formation capability indicator L obtained by decomposition are mapped to the second-layer and third-layer intermediate nodes in sequence; for the task layer, the overall task completion capability R of the heterogeneous unmanned cluster is mapped to the root node of the evidence network; The constructed evidence network is a directed acyclic graph structure, and the connections between each node are set based on the framework logic to describe the dependency and transmission relationship between capabilities; Furthermore, weights are set for the child nodes of any parent node. and satisfy Where d represents any child node, and D is the number of related child nodes contained in the parent node; meanwhile, the identification framework for each node in the network is defined as: To complete the construction of the evidence network; S400: Evaluate the reliability of heterogeneous unmanned cluster tasks based on evidence reasoning.

Images