Complex component multi-robot intelligent scheduling method and system based on knowledge graph

Abstract

The invention discloses a knowledge graph-based complex component multi-robot intelligent scheduling method and system, and the method comprises the steps: obtaining a three-dimensional model of a target component, carrying out the structure partitioning of the target component according to the three-dimensional model, generating a plurality of intention kernels, taking each intention kernel as a node, calculating an edge tensor value between the nodes, and carrying out the structural partitioning of the target component according to the edge tensor value. Forming an intention knowledge graph; setting a task scheduling driving function, calculating a scheduling driving index of each intention kernel according to the intention kernel and the edge tensor value, and determining a task distribution sequence and a structural area task weight of the robot according to the scheduling driving index; and according to the task allocation sequence and the structural area task weight, generating an optimal path for the robot, and according to the robot, completing a corresponding task according to the optimal path.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of robot control, and more particularly relates to a complex component multi-robot intelligent scheduling method and system based on a knowledge graph. BACKGROUND

[0002] Currently, in the fields of aerospace, energy equipment and high-end manufacturing, large complex components exhibit the characteristics of high integration, high precision requirements and coexistence of multi-scale heterogeneous structures, which poses unprecedented challenges to the manufacturing process. Traditional manufacturing systems rely on single process path, static task division and human-machine collaborative intervention, which is not enough in high complexity scenarios. Especially in the multi-robot collaborative manufacturing scene, the existing scheduling and path planning methods face a series of core bottlenecks: first, the task decomposition granularity is rough, and the region is often divided based on the rule template, which cannot accurately capture the deep coupling characteristics of component geometry-process; second, the task allocation method is mostly a static strategy, lacking a dynamic response mechanism to the current state of the robot, execution feedback and changes in the manufacturing environment, resulting in a lack of flexibility and robustness of the scheduling result; third, the path planning mostly uses a global pre-solution method, which is difficult to quickly reconstruct the path when there are process conflicts or physical interference, which can easily cause manufacturing delays or even execution failures. In addition, the existing technology relies heavily on expert experience in process parameter selection, lacks a systematic modeling and parameter matching mechanism for component features, robot capabilities and manufacturing goals, and is difficult to form an individualized, scene-aware parameter optimization scheme, ultimately affecting manufacturing quality and efficiency. SUMMARY

[0003] To solve the above technical problems, the present application provides a complex component multi-robot intelligent scheduling method based on a knowledge graph, comprising:

[0004] Obtaining a three-dimensional model of a target component, structurally partitioning the target component according to the three-dimensional model, generating a plurality of intent kernels, taking each intent kernel as a node, and calculating the edge tensor value between the nodes to form an intent knowledge graph;

[0005] Setting a task scheduling driving function, calculating the scheduling driving index of each intent kernel according to the intent kernel and the edge tensor value, and determining the task allocation order of the robot and the structural region task weight according to the scheduling driving index;

[0006] Generating an optimal path for the robot according to the task allocation order and the structural region task weight, and completing the corresponding task according to the optimal path.

[0007] Further, generating a plurality of intent kernels comprises:

[0008] Surface pretreatment is performed on the three-dimensional model, wherein the surface pretreatment comprises mesh simplification, triangulation processing on the surface of the three-dimensional model, and boundary box cropping of invalid regions:

[0009] The maximum curvature k of each point (x, y, z) on the three-dimensional model after surface pretreatment is calculated by a principal curvature projection algorithm max and the minimum curvature k min , and a curvature field K(x, y, z) of each point (x, y, z) is constructed, K(x, y, z) = |k max -k min |;

[0010] The curvature field K(x, y, z) of each point (x, y, z) is subjected to density clustering by a DBSCAN algorithm, each cluster center is an intent kernel, and each intent kernel is assigned a corresponding region label;

[0011] The structural region corresponding to the intent kernel with the region label is spatially matched with the standard semantic region in the process dictionary, so as to confirm the task type.

[0012] Further, an edge tensor value between nodes is calculated to form an intent knowledge graph, comprising:

[0013] The Euclidean distance between two intent kernels is obtained, and two intent kernels with a Euclidean distance less than a preset distance threshold are connected as nodes;

[0014] An edge tensor value between nodes is calculated, wherein the edge tensor value comprises a geometric coupling degree between the structural regions corresponding to two intent kernels, a process continuity index, and a spatial layout coupling degree.

[0015] Further, the geometric coupling degree comprises:

[0016]

[0017] wherein gen(S i ,S j ) is a geometric coupling degree between the structural region corresponding to the i th intent kernel S i and the j th intent kernel S j , a 1 is a first weight of the geometric coupling degree, is an average normal vector of the structural region corresponding to the i th intent kernel S i , is an average normal vector of the structural region corresponding to the j th intent kernel S j , a 2 is a second weight of the geometric coupling degree, and d ij is a distance between the i th intent kernel S i and the j th intent kernel S jThe Euclidean distance between them, where σ is the average diameter of the structural region corresponding to all intended kernels. For the i-th intention kernel S i The corresponding boundary length of the structural region, For the j-th intention kernel S j The corresponding boundary length of the structural region, α3 is the third weight of the geometric coupling degree;

[0018] The 3D model is input into a larger model, which outputs a process continuity index, where the i-th intention kernel S... i With the j-th intention kernel S j If the relationship between them is collaborative, then the process continuity index is 1. If the i-th intention kernel S i With the j-th intention kernel S j If there is no relationship between them, the process continuity index is 0. If the i-th intention kernel S i With the j-th intention kernel S j If there is a conflicting relationship between them, the process continuity index is -1;

[0019] Calculating spatial layout coupling includes:

[0020] layout(S i ,S j )=β1·IOU bb (S i ,S j )+β2·path conflict (S i ,S j )

[0021] Among them, layout(S i S j ) is the i-th intention kernel S i With the j-th intention kernel S j The spatial layout coupling degree between the corresponding structural regions, β1 is the first weight of the process continuity index, IOUbb(S i S j ) is the i-th intention kernel S i With the j-th intention kernel S j The corresponding bounding box overlap between structural regions, β2 is the second weight of the process continuity index, path conflict (S i S j ) is the i-th intention kernel S i With the j-th intention kernel S j The risk coefficient of path conflict between corresponding structural regions.

[0022] Furthermore, the task scheduling driver functions include:

[0023]

[0024] Where, Γ(S) i ) is the i-th intention kernel S i The scheduling-driven index, α′ is the first weight of the task scheduling-driven function, deg(S i ) is the i-th intention kernel S i In the intent knowledge graph, β′ represents the second weight of the task scheduling driver function, and ||L ij || is the i-th intention kernel S i With the j-th intention kernel S j The weighted norms of geometric coupling, process continuity index, and spatial layout coupling between corresponding structural regions.

[0025] Furthermore, determining the robot's task allocation order and structural region task weights according to the scheduling drive index includes: for the i-th intention kernel S i The scheduling-driven exponent Γ(S) i Sort the robots according to their size and assign them to the appropriate robots in descending order to perform the corresponding tasks.

[0026] Furthermore, generating the optimal path for the robot includes: generating key point paths for all robots based on the structural region corresponding to the intent kernel, forming a path point set for each robot, obtaining the overlap ratio of the path point sets between robots, and if the overlap ratio is greater than a preset ratio threshold, adjusting the time window for the robot to perform the task, or inserting intermediate nodes into the path point set to adjust the path for the robot to perform the task.

[0027] This invention also proposes a knowledge graph-based intelligent scheduling system for multiple robots with complex components, comprising:

[0028] An intent knowledge graph generation module is used to obtain a three-dimensional model of a target component, perform structural partitioning on the target component based on the three-dimensional model, generate multiple intent kernels, use each intent kernel as a node, and calculate the edge tensor values ​​between nodes to form an intent knowledge graph.

[0029] The scheduling module is used to set the task scheduling driving function, calculate the scheduling driving index of each intention kernel based on the intention kernel and the edge tensor value, and determine the task allocation order and structural region task weight of the robot according to the scheduling driving index.

[0030] The path planning module is used to generate an optimal path for the robot according to the task allocation order and the task weight of the structural region, and the robot completes the corresponding task according to the optimal path.

[0031] Furthermore, generating multiple intent kernels includes:

[0032] The surface of the 3D model is preprocessed, including: simplifying the mesh and triangulating the surface of the 3D model, and clipping invalid regions using bounding boxes.

[0033] The maximum curvature k of each point (x, y, z) on the 3D model after surface preprocessing is calculated using the principal curvature projection algorithm. max and minimum curvature k min Construct the curvature field K(x,y,z) for each point (x,y,z), where K(x,y,z) = |k max -k min |;

[0034] For the curvature field K(x,y,z) of each point (x,y,z), density clustering is performed using the DBSCAN algorithm. Each cluster center is an intent kernel, and each intent kernel is assigned a corresponding region label.

[0035] The task type is identified by spatially matching the structural regions corresponding to the intent kernels with region labels with the standard semantic regions in the process dictionary.

[0036] Furthermore, the edge tensor values ​​between nodes are calculated to form the intent knowledge graph, including:

[0037] Obtain the Euclidean distance between any two intent kernels, and connect any two intent kernels whose Euclidean distance is less than a preset distance threshold as nodes;

[0038] The edge tensor values ​​between nodes are calculated, wherein the edge tensor values ​​include: the geometric coupling degree, the process continuity index, and the spatial layout coupling degree between the structural regions corresponding to each pair of intention kernels.

[0039] In summary, the technical solutions conceived by this invention have the following beneficial effects compared with the prior art:

[0040] The technical solution of this invention can significantly improve the intelligence and dynamism of robot task division, enhance the executability and robustness of collaborative path planning, optimize the combination of manufacturing process parameters, and ultimately achieve the goals of improving manufacturing efficiency, stabilizing manufacturing quality, and reducing scheduling and path conflict rates. It is suitable for flexible intelligent manufacturing scenarios of complex components. Attached Figure Description

[0041] Figure 1 This is a flowchart of the method in Embodiment 1 of the present invention;

[0042] Figure 2 This is a system structure diagram of Embodiment 2 of the present invention. Detailed Implementation

[0043] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0044] The method provided by this invention can be implemented in a terminal environment that may include one or more of the following components: a processor, a storage medium, and a display screen. The storage medium stores at least one instruction, which is loaded and executed by the processor to implement the method described in the following embodiments.

[0045] A processor may include one or more processing cores. The processor uses various interfaces and lines to connect various parts of the terminal, and performs various functions and processes data by running or executing instructions, programs, code sets or instruction sets stored in the storage medium, and by calling data stored in the storage medium.

[0046] Storage media can include random access memory (RAM) or read-only memory (ROM). Storage media can be used to store instructions, programs, code, code sets, or instructions.

[0047] The display screen is used to show the user interface of each application.

[0048] In addition, those skilled in the art will understand that the structure of the terminal described above does not constitute a limitation on the terminal. The terminal may include more or fewer components, or combine certain components, or have different component arrangements. For example, the terminal may also include radio frequency circuits, input units, sensors, audio circuits, power supplies, and other components, which will not be described in detail here.

[0049] Axis-Aligned Bounding Box is a simplified rectangular volume bounding body commonly used in computational geometry.

[0050] Example 1

[0051] like Figure 1 This embodiment proposes a knowledge graph-based intelligent scheduling method for multiple robots with complex components, including:

[0052] Step 101: Obtain the three-dimensional model of the target component, perform structural partitioning on the target component based on the three-dimensional model, generate multiple intent kernels, use each intent kernel as a node, calculate the edge tensor values ​​between nodes, form an intent knowledge graph, and store the nodes and edge tensor values ​​in a graph database (such as Neo4j).

[0053] Specifically, generating multiple intent kernels includes:

[0054] The surface of the 3D model (from CAD, STEP, or IGES files) is preprocessed, including: mesh simplification and triangulation of the surface of the 3D model, and clipping invalid regions using bounding boxes.

[0055] The maximum curvature k of each point (x, y, z) on the 3D model after surface preprocessing is calculated using the principal curvature projection algorithm. max and minimum curvature k min Construct the curvature field K(x, y, z) for each point (x, y, z), where K(x, y, z) = |k max -k min |;

[0056] For the curvature field K(x,y,z) of each point (x,y,z), density clustering is performed using the DBSCAN algorithm. Each cluster center is an intent kernel, and each intent kernel is assigned a corresponding region label (such as welding area, bending surface, etc.).

[0057] The task type is identified by spatially matching the structural regions corresponding to the intent kernels with region labels with the standard semantic regions in the process dictionary (such as "fillet weld" and "slope spraying area").

[0058] Specifically, calculating the edge tensor values ​​between nodes to form the intent knowledge graph includes:

[0059] Obtain the Euclidean distance between any two intent kernels, and connect any two intent kernels whose Euclidean distance is less than a preset distance threshold as nodes;

[0060] The edge tensor values ​​between nodes are calculated, wherein the edge tensor values ​​include: the geometric coupling degree, the process continuity index, and the spatial layout coupling degree between the structural regions corresponding to each pair of intention kernels.

[0061] Specifically, calculating the geometric coupling degree includes:

[0062]

[0063] Among them, geo(S) i S j ) is the i-th intention kernel S i With the j-th intention kernel S j The geometric coupling degree between the corresponding structural regions, where α1 is the first weight of the geometric coupling degree. For the i-th intention kernel S i The average normal vector of the corresponding structural region, For the j-th intention kernel S j The average normal vector of the corresponding structural region, α2 is the second weight of the geometric coupling, dij For the i-th intention kernel S i With the j-th intention kernel S j The Euclidean distance between them, where σ is the average diameter of the structural region corresponding to all intended kernels. For the i-th intention kernel S i The corresponding boundary length of the structural region, For the j-th intention kernel S j The corresponding boundary length of the structural region, α3 is the third weight of the geometric coupling degree;

[0064] The 3D model is input into a larger model, which outputs a process continuity index, where the i-th intention kernel S... i With the j-th intention kernel S j If the relationship between them is collaborative, then the process continuity index is 1. If the i-th intention kernel S i With the j-th intention kernel S j If there is no relationship between them, the process continuity index is 0. If the i-th intention kernel S i With the j-th intention kernel S j If there is a conflicting relationship between them, the process continuity index is -1;

[0065] Preferably, the 3D model is input into a GPT or BERT model, with the following prompts: Region A is the main structural beam, where welding is planned; Region B is the outer shell, which requires painting. Is there a process sequence coordination relationship between these two regions? If the output shows a coordination relationship, the process continuity index is 1; if there is no relationship, the process continuity index is 0; if there is a conflict, the process continuity index is -1.

[0066] Calculating spatial layout coupling includes:

[0067] layout(S i ,S j )=β1·IOU bb (S i ,S j )+β2·path conflict (S i ,S j )

[0068] Among them, layout(S i S j ) is the i-th intention kernel S i With the j-th intention kernel S j The spatial layout coupling degree between the corresponding structural regions, β1 is the first weight of the process continuity index, IOU bb (S i S j ) is the i-th intention kernel Si With the j-th intention kernel S j The corresponding bounding box overlap between structural regions, β2 is the second weight of the process continuity index, path conflict (S i S j ) is the i-th intention kernel S i With the j-th intention kernel S j The risk coefficient of path conflict between corresponding structural regions.

[0069] Preferably, the i-th intention kernel S is calculated. i With the j-th intention kernel S j The bounding box overlap (IOU) between corresponding structural regions bb (S i S j The formula for ) is:

[0070]

[0071] Among them, Vol(B) i ∩B j ) is the i-th intention kernel S i The corresponding structural region's bounding box B i With the j-th intention kernel S j The corresponding structural region's bounding box B j The volume of the overlapping portion, Vol(B) i ∪B j ) is the i-th intention kernel S i The corresponding structural region's bounding box B i With the j-th intention kernel S j The corresponding structural region's bounding box B j Union volume.

[0072] Calculate the i-th intention kernel S i With the j-th intention kernel S j The path conflict risk coefficient between corresponding structural regions conflict (S i S j The formula for ) is:

[0073]

[0074] Among them, L overlap (P i ,P j ) represents the robot's intent kernel S at the i-th intention. i The corresponding structural region execution task path P i With the robot at the j-th intention kernel S jThe corresponding structural region execution task path P j The total length of segments between path points that are below the distance threshold, L(P) i ) represents the robot's intent kernel S at the i-th intention. i The corresponding structural region execution task path P i The length of L(P) j ) represents the robot's intent kernel S at the j-th time. j The corresponding structural region execution task path P j The length.

[0075] Step 102: Set the task scheduling driving function, calculate the scheduling driving index of each intention kernel according to the intention kernel and the edge tensor value, and determine the task allocation order and structural region task weight of the robot according to the scheduling driving index.

[0076] Specifically, the task scheduling driver functions include:

[0077]

[0078] Where, Γ(S) i ) is the i-th intention kernel S i The scheduling-driven index, α′ is the first weight of the task scheduling-driven function, deg(S i ) is the i-th intention kernel S i In the intent knowledge graph, β′ represents the second weight of the task scheduling driver function, and ||L ij || is the i-th intention kernel S i With the j-th intention kernel S j The weighted norms of geometric coupling, process continuity index, and spatial layout coupling between corresponding structural regions.

[0079] Specifically, determining the robot's task allocation order and structural region task weights according to the scheduling drive index includes: for the i-th intention kernel S i The scheduling-driven exponent Γ(S) i Sort the robots according to their size and assign them to the appropriate robots in descending order to perform the corresponding tasks.

[0080] Step 103: Generate an optimal path for the robot according to the task allocation order and the task weight of the structural region, and then have the robot complete the corresponding task according to the optimal path.

[0081] Specifically, generating the optimal path for the robot includes: generating key point paths for all robots based on the structural region corresponding to the intent kernel (using algorithms such as RRT* or A*), forming a path point set for each robot, obtaining the overlap ratio of the path point sets between robots, and if the overlap ratio is greater than a preset threshold, adjusting the time window for the robot to perform the task, or inserting intermediate nodes into the path point set to adjust the path for the robot to perform the task.

[0082] Example 2

[0083] like Figure 2 As shown, this embodiment proposes a multi-robot intelligent scheduling system for complex components based on knowledge graphs, including:

[0084] An intent knowledge graph generation module is used to obtain a three-dimensional model of a target component, perform structural partitioning on the target component based on the three-dimensional model, generate multiple intent kernels, use each intent kernel as a node, and calculate the edge tensor values ​​between nodes to form an intent knowledge graph.

[0085] Specifically, generating multiple intent kernels includes:

[0086] The surface of the 3D model is preprocessed, including: simplifying the mesh and triangulating the surface of the 3D model, and clipping invalid regions using bounding boxes.

[0087] The maximum curvature k of each point (x, y, z) on the 3D model after surface preprocessing is calculated using the principal curvature projection algorithm. max and minimum curvature k min Construct the curvature field K(x, y, z) for each point (x, y, z), where K(x, y, z) = |k max -k min |;

[0088] For the curvature field K(x,y,z) of each point (x,y,z), density clustering is performed using the DBSCAN algorithm. Each cluster center is an intent kernel, and each intent kernel is assigned a corresponding region label.

[0089] The task type is identified by spatially matching the structural regions corresponding to the intent kernels with region labels with the standard semantic regions in the process dictionary.

[0090] Specifically, calculating the edge tensor values ​​between nodes to form the intent knowledge graph includes:

[0091] Obtain the Euclidean distance between any two intent kernels, and connect any two intent kernels whose Euclidean distance is less than a preset distance threshold as nodes;

[0092] The edge tensor values ​​between nodes are calculated, wherein the edge tensor values ​​include: the geometric coupling degree, the process continuity index, and the spatial layout coupling degree between the structural regions corresponding to each pair of intention kernels.

[0093] Specifically, calculating the geometric coupling degree includes:

[0094]

[0095] Among them, geo(S) i S j ) is the i-th intention kernel S i With the j-th intention kernel S j The geometric coupling degree between the corresponding structural regions, where α1 is the first weight of the geometric coupling degree. For the i-th intention kernel S i The average normal vector of the corresponding structural region, For the j-th intention kernel S j The average normal vector of the corresponding structural region, α2 is the second weight of the geometric coupling, d ij For the i-th intention kernel S i With the j-th intention kernel S j The Euclidean distance between them, where σ is the average diameter of the structural region corresponding to all intended kernels. For the i-th intention kernel S i The corresponding boundary length of the structural region, For the j-th intention kernel S j The corresponding boundary length of the structural region, α3 is the third weight of the geometric coupling degree;

[0096] The 3D model is input into a larger model, which outputs a process continuity index, where the i-th intention kernel S... i With the j-th intention kernel S j If the relationship between them is collaborative, then the process continuity index is 1. If the i-th intention kernel S i With the j-th intention kernel S j If there is no relationship between them, the process continuity index is 0. If the i-th intention kernel S i With the j-th intention kernel S j If there is a conflicting relationship between them, the process continuity index is -1;

[0097] Calculating spatial layout coupling includes:

[0098] layout(S i ,S j )=β1·IOU bb (S i ,S j )+β2·path conflict (Si ,S j )

[0099] Among them, layout(S i S j ) is the i-th intention kernel S i With the j-th intention kernel S j The spatial layout coupling degree between the corresponding structural regions, β1 is the first weight of the process continuity index, IOU bb (S i S j ) is the i-th intention kernel S i With the j-th intention kernel S j The corresponding bounding box overlap between structural regions, β2 is the second weight of the process continuity index, path conflict (S i ,S j ) is the i-th intention kernel S i With the j-th intention kernel S j The risk coefficient of path conflict between corresponding structural regions.

[0100] Preferably, the i-th intention kernel S is calculated. i With the j-th intention kernel S j The bounding box overlap (IOU) between corresponding structural regions bb (S i S j The formula for ) is:

[0101]

[0102] Among them, Vol(B) i ∩B j ) is the i-th intention kernel S i The corresponding structural region's bounding box B i With the j-th intention kernel S j The corresponding structural region's bounding box B j The volume of the overlapping portion, Vol(B) i ∪B j ) is the i-th intention kernel S i The bounding box Bi of the corresponding structural region and the j-th intention kernel S j The corresponding structural region's bounding box B j Union volume.

[0103] Calculate the i-th intention kernel S i With the j-th intention kernel S j The path conflict risk coefficient between corresponding structural regions conflict (S i Sj The formula for ) is:

[0104]

[0105] Among them, L overlap (P i P j ) represents the robot's intent kernel S at the i-th intention. i The corresponding structural region execution task path P i With the robot at the j-th intention kernel S j The corresponding structural region execution task path P j The total length of segments between path points that are below the distance threshold, L(P) i ) represents the robot's intent kernel S at the i-th intention. i The corresponding structural region execution task path P i The length of L(P) j ) represents the robot's intent kernel S at the j-th time. j The corresponding structural region execution task path P j The length.

[0106] The scheduling module is used to set the task scheduling driving function, calculate the scheduling driving index of each intention kernel based on the intention kernel and the edge tensor value, and determine the task allocation order and structural region task weight of the robot according to the scheduling driving index.

[0107] Specifically, the task scheduling driver functions include:

[0108]

[0109] Where, Γ(S) i ) is the i-th intention kernel S i The scheduling-driven index, α′ is the first weight of the task scheduling-driven function, deg(S i ) is the i-th intention kernel S i In the intent knowledge graph, β′ represents the second weight of the task scheduling driver function, and ||L ij || is the i-th intention kernel S i With the j-th intention kernel S j The weighted norms of geometric coupling, process continuity index, and spatial layout coupling between corresponding structural regions.

[0110] Specifically, determining the robot's task allocation order and structural region task weights according to the scheduling drive index includes: for the i-th intention kernel S i The scheduling-driven exponent Γ(S) i Sort the robots according to their size and assign them to the appropriate robots in descending order to perform the corresponding tasks.

[0111] The path planning module is used to generate an optimal path for the robot according to the task allocation order and the task weight of the structural region, and the robot completes the corresponding task according to the optimal path.

[0112] Specifically, generating the optimal path for the robot includes: generating key point paths for all robots based on the structural region corresponding to the intent kernel, forming a path point set for each robot, obtaining the overlap ratio of the path point sets between robots, and if the overlap ratio is greater than a preset threshold, adjusting the time window for the robot to perform the task, or inserting intermediate nodes into the path point set to adjust the path for the robot to perform the task.

[0113] Example 3

[0114] This invention also proposes a storage medium storing multiple instructions, which are used to implement the aforementioned knowledge graph-based intelligent scheduling method for multiple robots with complex components.

[0115] Optionally, in this embodiment, the storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals.

[0116] Optionally, in this embodiment, the storage medium is configured to store program code for performing the method steps of Embodiment 1.

[0117] Example 4

[0118] This invention also proposes an electronic device, including a processor and a storage medium connected to the processor. The storage medium stores multiple instructions, which can be loaded and executed by the processor to enable the processor to execute the aforementioned knowledge graph-based intelligent scheduling method for multiple robots with complex components.

[0119] Specifically, the electronic device in this embodiment can be a computer terminal, which may include one or more processors and a storage medium.

[0120] The storage medium can be used to store software programs and modules, such as the knowledge graph-based intelligent scheduling method for multiple robots with complex components in this embodiment of the invention. The processor executes the software programs and modules stored in the storage medium to perform various functional applications and data processing, thus realizing the aforementioned knowledge graph-based intelligent scheduling method for multiple robots with complex components. The storage medium may include high-speed random access storage media, and may also include non-volatile storage media, such as one or more magnetic storage systems, flash memory, or other non-volatile solid-state storage media. In some instances, the storage medium may further include storage media remotely configured relative to the processor, which can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0121] The processor can execute the method steps of Embodiment 1 by calling the information and application stored in the storage medium through the transmission system.

[0122] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0123] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0124] In the several embodiments provided by this invention, it should be understood that the disclosed technical content can be implemented in other ways. The system embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection between units or modules, and may be electrical or other forms.

[0125] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0126] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0127] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, optical disks, and other media capable of storing program code.

[0128] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A knowledge graph-based intelligent scheduling method for multiple robots with complex components, characterized in that, include: A three-dimensional model of the target component is obtained, and the target component is structurally partitioned according to the three-dimensional model to generate multiple intent kernels. Each intent kernel is used as a node, and the edge tensor values ​​between nodes are calculated to form an intent knowledge graph. Set up a task scheduling driving function, calculate the scheduling driving index of each intention kernel based on the intention kernel and the edge tensor value, and determine the task allocation order and structural region task weight of the robot according to the scheduling driving index. Based on the task allocation order and the task weight of the structural region, an optimal path is generated for the robot, and the robot completes the corresponding task according to the optimal path.

2. The intelligent scheduling method for multiple robots with complex components based on knowledge graphs as described in claim 1, characterized in that, Generating multiple intent kernels includes: The surface of the 3D model is preprocessed, including: simplifying the mesh and triangulating the surface of the 3D model, and clipping invalid regions using bounding boxes. The maximum curvature k of each point (x, y, z) on the 3D model after surface preprocessing is calculated using the principal curvature projection algorithm. max and minimum curvature k min Construct the curvature field K(x,y,z) for each point (x,y,z), where K(x,y,z) = |k max -k min |; For the curvature field K(x,y,z) of each point (x,y,z), density clustering is performed using the DBSCAN algorithm. Each cluster center is an intent kernel, and each intent kernel is assigned a corresponding region label. The task type is identified by spatially matching the structural regions corresponding to the intent kernels with region labels with the standard semantic regions in the process dictionary.

3. The intelligent scheduling method for multiple robots with complex components based on knowledge graphs as described in claim 1, characterized in that, Calculating the edge tensor values ​​between nodes to form an intent knowledge graph includes: Obtain the Euclidean distance between any two intent kernels, and connect any two intent kernels whose Euclidean distance is less than a preset distance threshold as nodes; The edge tensor values ​​between nodes are calculated, wherein the edge tensor values ​​include: the geometric coupling degree, the process continuity index, and the spatial layout coupling degree between the structural regions corresponding to each pair of intention kernels.

4. The intelligent scheduling method for multiple robots with complex components based on knowledge graphs as described in claim 3, characterized in that, Calculating the geometric coupling degree includes: Among them, geo(S) i S j ) is the i-th intention kernel S i With the j-th intention kernel S j The geometric coupling degree between the corresponding structural regions, where α1 is the first weight of the geometric coupling degree. For the i-th intention kernel S i The average normal vector of the corresponding structural region, For the j-th intention kernel S j The average normal vector of the corresponding structural region, α2 is the second weight of the geometric coupling, d ij For the i-th intention kernel S i With the j-th intention kernel S j The Euclidean distance between them, where σ is the average diameter of the structural region corresponding to all intended kernels. For the i-th intention kernel S i The corresponding boundary length of the structural region, For the j-th intention kernel S j The corresponding boundary length of the structural region, α3 is the third weight of the geometric coupling degree; The 3D model is input into a larger model, which outputs a process continuity index, where the i-th intention kernel S... i With the j-th intention kernel S j If the relationship between them is collaborative, then the process continuity index is 1. If the i-th intention kernel S i With the j-th intention kernel S j If there is no relationship between them, the process continuity index is 0. If the i-th intention kernel S i With the j-th intention kernel S j If there is a conflicting relationship between them, the process continuity index is -1; Calculating spatial layout coupling includes: layout(S i ,S j )=β1·IOU bb (S i ,S j )+β2·path conflict (S i ,S j ) Among them, layout(S i ,S j ) is the i-th intention kernel S i With the j-th intention kernel S j The spatial layout coupling degree between the corresponding structural regions, β1 is the first weight of the process continuity index, IOU bb (S i S j ) is the i-th intention kernel S i With the j-th intention kernel S j The corresponding bounding box overlap between structural regions, β2 is the second weight of the process continuity index, path conflict (S i S j ) is the i-th intention kernel S i With the j-th intention kernel S j The risk coefficient of path conflict between corresponding structural regions.

5. The intelligent scheduling method for multiple robots with complex components based on knowledge graphs as described in claim 4, characterized in that, Task scheduling driver functions include: Where, Γ(S) i ) is the i-th intention kernel S i The scheduling-driven index, α′ is the first weight of the task scheduling-driven function, deg(S i ) is the i-th intention kernel S i In the intent knowledge graph, β′ represents the second weight of the task scheduling driver function, and ||L ij || is the i-th intention kernel S i With the j-th intention kernel S j The weighted norms of geometric coupling, process continuity index, and spatial layout coupling between corresponding structural regions.

6. The intelligent scheduling method for multiple robots with complex components based on knowledge graphs as described in claim 5, characterized in that, Determining the robot's task allocation order and structural region task weights according to the aforementioned scheduling drive index includes: for the i-th intention kernel S i The scheduling-driven exponent Γ(S) i Sort the robots according to their size and assign them to the appropriate robots in descending order to perform the corresponding tasks.

7. The intelligent scheduling method for multiple robots with complex components based on knowledge graphs as described in claim 1, characterized in that, Generating the optimal path for the robot includes: generating key point paths for all robots based on the structural region corresponding to the intent kernel, forming a path point set for each robot, obtaining the overlap ratio of the path point sets between robots, and if the overlap ratio is greater than a preset threshold, adjusting the time window for the robot to perform the task, or inserting intermediate nodes into the path point set to adjust the path for the robot to perform the task.

8. A knowledge graph-based intelligent scheduling system for multiple robots with complex components, characterized in that, include: An intent knowledge graph generation module is used to obtain a three-dimensional model of a target component, perform structural partitioning on the target component based on the three-dimensional model, generate multiple intent kernels, use each intent kernel as a node, and calculate the edge tensor values ​​between nodes to form an intent knowledge graph. The scheduling module is used to set the task scheduling driving function, calculate the scheduling driving index of each intention kernel based on the intention kernel and the edge tensor value, and determine the task allocation order and structural region task weight of the robot according to the scheduling driving index. The path planning module is used to generate an optimal path for the robot according to the task allocation order and the task weight of the structural region, and the robot completes the corresponding task according to the optimal path.

9. The knowledge graph-based intelligent scheduling system for multiple robots with complex components as described in claim 8, characterized in that, Generating multiple intent kernels includes: The surface of the 3D model is preprocessed, including: simplifying the mesh and triangulating the surface of the 3D model, and clipping invalid regions using bounding boxes. The maximum curvature k of each point (x, y, z) on the 3D model after surface preprocessing is calculated using the principal curvature projection algorithm. max and minimum curvature k min Construct the curvature field K(x,y,z) for each point (x,y,z), where K(x,y,z) = |k max -k min |; For the curvature field K(x,y,z) of each point (x,y,z), density clustering is performed using the DBSCAN algorithm. Each cluster center is an intent kernel, and each intent kernel is assigned a corresponding region label. The task type is identified by spatially matching the structural regions corresponding to the intent kernels with region labels with the standard semantic regions in the process dictionary.

10. The knowledge graph-based intelligent scheduling system for multiple robots with complex components as described in claim 8, characterized in that, Calculating the edge tensor values ​​between nodes to form an intent knowledge graph includes: Obtain the Euclidean distance between any two intent kernels, and connect any two intent kernels whose Euclidean distance is less than a preset distance threshold as nodes; The edge tensor values ​​between nodes are calculated, wherein the edge tensor values ​​include: the geometric coupling degree, the process continuity index, and the spatial layout coupling degree between the structural regions corresponding to each pair of intention kernels.

Images