Robot adaptive operation method and device for complex cabin structure scene

Abstract

The application provides a robot adaptive operation method and device for a complex cabin structure scene, and relates to the technical field of intelligent robots, which comprises the following steps: acquiring multi-source sensor data collected by a robot from a cabin structure scene, identifying equipment objects and space units in the cabin structure scene based on the multi-source sensor data, and constructing a corresponding topology graph; performing topology correlation analysis on nodes and edges in the topology graph, adjusting the topology graph according to the analysis result, and scheduling the priority of the robot's inspection operation task according to the key level of the nodes in the adjusted operation scene model; finally, planning a path for the robot according to the scheduled operation task queue, and controlling the robot to inspect the cabin structure scene based on the planned inspection path. The application is used to solve the problem that the existing robot has insufficient inspection operation task scheduling capability and low overall inspection efficiency when inspecting a complex cabin structure scene.

Description

Technical Field

[0001] This invention relates to the field of intelligent robot technology, and specifically to a robot adaptive operation method and apparatus for complex cabin structure scenarios. Background Technology

[0002] As the number of devices in complex cabin environments continues to increase, the spatial structure becomes increasingly complex, and the functional coupling between devices continues to strengthen, embodied intelligent inspection robots, when performing inspection tasks, not only need to complete basic path passage, but also need to dynamically sort, prioritize, and coordinate multiple tasks based on the importance of devices, abnormal states, and the relationships between nodes in the scenario.

[0003] However, when inspecting complex cabin structures, existing technologies mostly focus on directly planning the robot's path based on the inspection target, primarily selecting paths based on the shortest path or obstacle avoidance strategies to execute the corresponding inspection tasks. These methods lack global scheduling of the inspection tasks themselves, and the priority of inspection tasks is related to the scene equipment itself. Existing path planning technologies emphasize obstacle avoidance and inspection efficiency without considering the scene equipment itself, resulting in insufficient task scheduling capabilities and low overall inspection efficiency for robots in complex cabin structure scenarios. This makes it difficult to meet the intelligent inspection needs of embodied intelligent inspection robots in highly complex scenarios. Summary of the Invention

[0004] In view of this, it is necessary to provide a robot adaptive operation method and device for complex cabin structure scenarios, so as to solve the problems of insufficient task scheduling capability and low overall inspection efficiency of existing robots when inspecting complex cabin structure scenarios.

[0005] To address the aforementioned problems, this invention provides an adaptive robot operation method for complex cabin structure scenarios, comprising: The robot acquires multi-source sensor data collected from the cabin structure scene, and identifies equipment objects and spatial units in the cabin structure scene based on the multi-source sensor data; Construct a topology graph corresponding to the cabin structure scene, wherein the nodes in the topology graph are the equipment objects or the space units, and the edges connecting the nodes are the spatial connectivity relationships between the nodes; A topological association analysis is performed on the nodes and edges in the topological graph, and the topological graph is adjusted according to the analysis results to obtain the robot's operation scenario model. The analysis results include the criticality level of the nodes, which is used to characterize the importance of the nodes in the cabin structure scenario. Based on the criticality level of the nodes in the work scenario model, the robot's inspection tasks are prioritized and scheduled to obtain a task queue. The robot is path-planned according to the task queue to obtain the inspection path, and the robot is controlled to inspect the cabin structure scene based on the inspection path.

[0006] In one possible implementation, the topology diagram corresponding to the construction of the cabin structure scenario includes: Calculate the union of the set of device objects corresponding to the device object and the set of spatial units corresponding to the spatial unit; Map the elements in the union set to nodes; When there is a geometric adjacency or passage constraint relationship between the nodes, an edge is established between the nodes; Based on the nodes and edges, a topology graph corresponding to the cabin structure scene is constructed.

[0007] In one possible implementation, the topological association analysis of the nodes and edges in the topological graph includes: For any two nodes in the topology graph, when the two nodes satisfy at least one association condition based on the node attribute values, it is determined that the two nodes have a topological association. The node attribute values ​​include semantic category identifiers and spatial location parameters. The association conditions include: semantic basic association, spatial proximity association, and dynamic influence association. The strength of topological association is determined based on the number of association conditions satisfied. The level of the strength of topological association is positively correlated with the number of association conditions satisfied. The strength of topological association includes: strong association, medium association, weak association, and no association. For a node in the topology graph, the criticality level of the node is determined based on the factor conditions satisfied by the node. The criticality level is positively correlated with the number of factor conditions satisfied. The criticality levels include: high criticality, medium criticality, low criticality, and non-criticality. The factor conditions include: The semantic category identifier of the node indicates that the device object corresponding to the node belongs to a preset key device or is located in a preset core area in the cabin structure scenario; The device status parameters corresponding to the node are within the preset normal parameter range; The number of nodes that are spatially adjacent to the node exceeds a preset threshold. The topological association strength between the node and its neighboring nodes is either strong or medium.

[0008] In one possible implementation, the association condition is determined in the following way: When it is determined that any one of the target conditions is met based on the semantic category identifiers of the two nodes, the association condition satisfied by the two nodes is determined as the semantic basic association. The target conditions include: the two nodes are in the same functional area in the cabin structure scene, the two nodes have a relationship of belonging between equipment objects and space units in the cabin structure scene, and the two nodes have a relationship of equipment function dependency in the cabin structure scene. The Euclidean distance between the two nodes is determined based on their spatial location parameters. When the Euclidean distance is less than a preset Euclidean distance threshold, the association condition satisfied by the two nodes is determined to be spatial proximity association. When an abnormal device state occurs in either of the two nodes, causing a change in the device state of the other node, the association condition that the two nodes satisfy is determined to be a dynamic influence association.

[0009] In one possible implementation, adjusting the topology map based on the analysis results to obtain the robot's operational scenario model includes: Two nodes with a strong topological association in the topology graph are identified. If the two nodes are in the same functional area in the cabin structure scene or have a relationship of belonging to equipment objects and space units, the two nodes are merged. Nodes in the topology graph that have more than two sub-regions are identified as nodes to be partitioned. When the standard deviation of the device state parameters corresponding to the node to be partitioned exceeds a preset heterogeneity threshold and there is no physical separation between the sub-regions, the node to be partitioned is divided into multiple homogeneous sub-nodes according to the clustering results of the device state parameters. Determine two non-adjacent nodes in the topology graph that have a strong or medium topological association. If the non-adjacent nodes have a device function dependency or a dynamic influence association in the cabin structure scenario, establish an edge between the non-adjacent nodes. When the topological association strength between two nodes in the topological graph is no association, delete the edge between the two nodes and delete the nodes in the topological graph whose criticality level is non-critical.

[0010] In one possible implementation, the robot's inspection tasks are prioritized and scheduled according to the criticality level of nodes in the work scenario model to obtain a task queue, including: The nodes are sorted according to their criticality level in the scenario model to obtain the sorting result. Nodes with the same criticality level are sorted according to the strength of their topological association. The robot is assigned inspection tasks for each node according to the sorting results. The inspection tasks of adjacent nodes with strong topological association in the sorting results are merged. The inspection tasks of adjacent nodes with the same criticality level in the sorting results are scheduled for execution according to the topological association strength. The inspection tasks, after being allocated and merged or scheduled in sequence, are added to the task scheduling queue to obtain the task queue.

[0011] In one possible implementation, when the robot, controlled by the inspection path, performs an inspection in a cabin structure scenario, the method further includes: If an abnormal node with abnormal equipment status parameters is detected in the work scenario model, the criticality level of the abnormal node is re-determined. When it is determined that the criticality level of the abnormal node has increased, the inspection task of the abnormal node is scheduled to the head of the task queue. When it is determined that the criticality level of the abnormal node has decreased, the inspection task of the abnormal node is scheduled to the end of the task queue or the inspection task of the abnormal node is deleted from the task queue. When an abnormal node with abnormal equipment status parameters is detected in one of the two nodes of the merged inspection task in the operation scenario model, the criticality level of the abnormal node is re-determined. The execution priority of the inspection tasks of the abnormal nodes is evaluated based on the criticality level, and the merged inspection tasks are split or the execution priority of the inspection tasks of the abnormal nodes is re-scheduled based on the evaluation results.

[0012] The present invention also provides a robot adaptive operation device for complex cabin structure scenarios, comprising: The data acquisition module is used to acquire multi-source sensor data collected by the robot from the cabin structure scene, and to identify equipment objects and space units in the cabin structure scene based on the multi-source sensor data; The topology construction module is used to construct a topology graph corresponding to the cabin structure scene, wherein the nodes in the topology graph are the equipment objects or the spatial units, and the edges connecting the nodes are the spatial connectivity relationships between the nodes; The topology analysis module is used to perform topological association analysis on the nodes and edges in the topology graph, and adjust the topology graph according to the analysis results to obtain the robot's operation scenario model. The analysis results include the criticality level of the nodes, which is used to characterize the importance of the nodes in the cabin structure scenario. The task scheduling module is used to prioritize the robot's inspection tasks based on the criticality level of the nodes in the operation scenario model. The path planning module is used to plan the robot's path according to the task queue, obtain the inspection path, and control the robot to inspect in the cabin structure scenario based on the inspection path.

[0013] The present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a program; the processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps of the above-described robot adaptive operation method for complex cabin structure scenarios.

[0014] The present invention also provides a robot adaptive operation system for complex cabin structure scenarios, including: sensors, communication devices, and the aforementioned electronic devices; the sensors and the electronic devices are both mounted on the robot; the sensors are communicatively connected to the electronic devices through the communication devices, and are used to collect multi-source sensor data in cabin structure scenarios; the communication devices are used to store the multi-source sensor data in the memory of the electronic devices.

[0015] The beneficial effects of the above implementation method are as follows: The robot adaptive operation method and device for complex cabin structure scenarios provided by this invention first digitizes the equipment objects and spatial units in the cabin structure scenario into nodes, and then establishes edges between nodes based on spatial connectivity. Through a topology graph, the complex cabin environment is transformed into a computable model with spatial semantics and structural relationships, thus solving the problem of complex environmental structures that are difficult to model effectively in existing technologies. Based on the topology graph, this embodiment of the invention analyzes the topological relationships between equipment objects and spatial units through topological association analysis, thereby effectively mining the existence of semantic and spatial topological relationships and identifying the importance of equipment, which is measured by a criticality level. In this case, the topological relationships between equipment objects are adjusted based on their importance, and scene modeling is performed, facilitating task scheduling and path planning. To ensure the scheduling capability of inspection tasks, inspection task scheduling is performed first based on the importance of equipment objects in the scene model, rather than directly planning paths. This approach establishes path planning on top of inspection task scheduling, enabling robot inspections to prioritize the inspection of more important equipment objects and the topological relationships between them, thereby improving the overall inspection efficiency of cabin structure scenarios. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating the robot adaptive operation method for complex cabin structure scenarios provided by this invention; Figure 2 This is a structural schematic diagram of the robot adaptive operation device for complex cabin structure scenarios provided by the present invention. Figure 3 A schematic diagram of an embodiment of the electronic device provided by the present invention. Detailed Implementation

[0018] 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 a part of the embodiments of the present invention, and not all of them. 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.

[0019] In the description of the embodiments of this application, unless otherwise stated, "a plurality of" means two or more.

[0020] In this embodiment of the invention, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, apparatus, product or device that includes a series of steps or modules is not necessarily limited to those steps or modules that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such process, method, product or device.

[0021] The naming or numbering of steps in the embodiments of the present invention does not mean that the steps in the method flow must be executed in the time / logical order indicated by the naming or numbering. The execution order of the named or numbered process steps can be changed according to the technical purpose to be achieved, as long as the same or similar technical effect can be achieved.

[0022] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0023] The robot adaptive operation method for complex cabin structure scenarios provided in this invention can be applied to automated robot inspection operations in certain complex cabin structure scenarios. The executing entity can be various servers or terminals installed inside the robot, or servers, terminals, or remote devices that communicate remotely with the robot. Complex cabin structure scenarios specifically include complex operation scenarios with spatial enclosure, dense equipment, and functional coupling, such as ship cabins, aviation equipment sections, underground pipe corridors, and enclosed industrial equipment areas. When the robot performs inspection in a complex cabin structure scenario, it collects and uploads multi-source sensor data through sensors installed inside the robot or external sensors. Then, it calls the robot adaptive operation method for complex cabin structure scenarios provided in this invention to sequentially construct a topology map corresponding to the cabin structure scenario, prioritize and schedule inspection tasks, plan paths, and ultimately control the robot to perform inspections in the cabin structure scenario, completing the inspection task.

[0024] The following describes in detail the robot adaptive operation method for complex cabin structure scenarios provided by this invention.

[0025] Figure 1 This is a flowchart illustrating the robot adaptive operation method for complex cabin structure scenarios provided by the present invention, as shown below. Figure 1 As shown, the robot adaptive operation method for complex cabin structure scenarios can be implemented through the following steps 101 to 105, which are explained in detail below.

[0026] Step 101: Acquire multi-source sensor data collected by the robot from the cabin structure scene, and identify equipment objects and space units in the cabin structure scene based on the multi-source sensor data.

[0027] Multi-source sensor data is collected from complex cabin environments by deploying multiple types of sensors on an embodied intelligent inspection robot, forming a unified data input set for subsequent topology modeling. The multi-source sensor data D is denoted as:

[0028] in, This represents the constructed multi-source data set. Represents a set of visual image data. Represents a depth dataset. Represents a set of laser point cloud data. Represents the set of inertial measurement data. Represents a set of environmental state data. Indicates the first Sensor-like devices at all times The collected data set Indicates the first Sensor-like devices at all times The first piece of data collected, Indicates the first Sensor-like devices at all times The first collection One data point, This indicates the total number of data points collected in the dataset.

[0029] By performing structured analysis on multi-source sensor data, basic units that can characterize the composition of the cabin space can be obtained, including space units and equipment objects.

[0030] Specifically, for multi-source sensor data, alignment processing is performed under a unified spatial coordinate system to form fused data. Spatial segmentation and target recognition processing are then performed based on this fused data. Spatial segmentation is used to divide the scene into regions, dividing the continuous cabin space into several spatial units with physical boundaries or functional attributes, forming a set of regions R. Then:

[0031] in, These represent the 1st, 2nd, 3rd, ..., mth spatial units, respectively, where m represents the total number of spatial units.

[0032] Target recognition is used to extract device objects with clear functional attributes, forming a set O of device objects, then:

[0033] in, These represent the 1st, 2nd, 3rd, ..., nth device objects, where n represents the total number of device objects.

[0034] Step 102: Construct the topology diagram corresponding to the cabin structure scene.

[0035] When constructing the topology graph, spatial units and equipment objects in the cabin structure scene are treated as nodes in the graph structure. The edge connections between nodes can be determined based on spatial connectivity or navigability. Since the cabin topology graph structure can be constructed based on the determined edge connections between nodes, a complete topology graph can be obtained.

[0036] In one possible implementation, the topology diagram corresponding to the cabin structure scene can be constructed in the following way, which is explained in detail below.

[0037] Calculate the union of the set of device objects corresponding to the device object and the set of spatial units corresponding to the spatial unit; Map the elements in the union to nodes; When there are geometric adjacency or passage constraints between nodes, an edge is established between the nodes; Based on nodes and edges, a topology graph corresponding to the cabin structure scene is constructed.

[0038] Specifically, firstly, based on the constructed set of spatial units and the set of device objects, the set of spatial semantic units is determined, then:

[0039] in, Represents a definite set of spatial semantic units. Represents the set of spatial units constructed. Represents the collection of device objects being constructed; To ensure the consistency and computability of spatial semantic units in subsequent graph structure construction, a unified structured description can be assigned to any spatial semantic unit, then:

[0040] in, Represents spatial location parameters (represents the spatial coordinate information of semantic units). Indicates the geometric extent parameter (indicates the spatial extent). Indicates semantic category identifier (indicates device or space type). The first element in the spatial semantic unit set U represents the... A spatial semantic unit.

[0041] When constructing a topological graph, the determined set of spatial semantic units is mapped to the set of nodes in the graph structure, then:

[0042] in, Represents a node in a topological graph, and is related to a spatial semantic unit. One-to-one correspondence allows nodes to inherit their spatial and semantic attributes. This represents the set of nodes in the constructed topology graph.

[0043] Based on this, for any two nodes (i.e., spatial semantic units) , Based on spatial distribution relationships and environmental constraints, a connection determination function is defined, then:

[0044] in, This indicates that the connection condition is met. This indicates that the connection condition is not met. Meeting the connection condition specifically means that a spatial connectivity relationship exists, which includes geometric adjacency relationships. Relationship with traffic constraints , A value of 1 indicates that the two nodes are geometrically adjacent, while a value of 0 indicates that they are not geometrically adjacent. Similarly, based on the defined connection conditions, the edges between nodes are constructed as follows: When the connection conditions between nodes satisfy the formula If the condition is met, it means that there is a spatial connection between the two nodes, and an edge can be established to connect them. Conversely, if the condition is not met, it means that there is no spatial connection between the two nodes, and an edge cannot be established to connect them.

[0045] Based on the defined edge connections, constructing the edge set in the graph structure yields:

[0046] in, Represents the first in the topological graph The node and the first Edges between nodes This represents the set of edges in the constructed topological graph. Furthermore, attribute information can be assigned to each edge to construct an edge attribute set, resulting in:

[0047] in, Indicates the first The node and the first The set of edge attributes between nodes, where edge attributes can include distance, difficulty of passage, and security risk.

[0048] Finally, based on the determined edge connections between nodes, a topology graph corresponding to the cabin structure scene is constructed, denoted as G, and represented as:

[0049] in, , representing the set of node attributes With edge attribute set , This represents the set of edges in the constructed topological graph. This represents the set of nodes in the constructed topology graph.

[0050] In this embodiment of the invention, the equipment objects and spatial units in the cabin structure scene are digitized into nodes, and then edges between nodes are established according to spatial connectivity. Through the topology graph, the complex cabin environment is transformed into a computable model with spatial semantics and structural relationships, thereby solving the problem of complex environmental structures that are difficult to model effectively in the prior art, and establishing a model foundation for subsequent robot inspection task scheduling and path planning.

[0051] Step 103: Perform topological association analysis on the nodes and edges in the topological graph, and adjust the topological graph according to the analysis results to obtain the robot's operation scenario model.

[0052] Here, topological association analysis is performed on the nodes and edges in the topological graph. On the one hand, the topological structure between nodes is analyzed based on node attributes and edges to determine whether topological associations exist, and this relationship is measured by the defined topological association strength. On the other hand, for each node itself, its importance in the cabin structure scenario is identified based on its attribute values, and this importance is measured by the defined criticality level.

[0053] In one possible implementation, topological association analysis of nodes and edges in a topological graph can be performed in the following way, which is explained in detail below.

[0054] For any two nodes in the topology graph, if the two nodes satisfy at least one association condition based on the node attribute values, it is determined that the two nodes have a topological association. The strength of topological association is determined based on the number of association conditions satisfied. For a node in the topology graph, the criticality level of the node is determined based on the factor conditions that the node satisfies.

[0055] Specifically, the first aspect of topological association analysis is determining the strength of the topological association between two nodes. For any two nodes in the topological graph, their node attribute values ​​can be extracted, including semantic category identifiers and spatial location parameters. When performing topological analysis on two nodes, three factors are mainly examined: semantic category association, spatial distance relationship, and the influence of dynamic state. Then, a logical combination and judgment are made through a rule system to determine the strength of the topological association.

[0056] These three factors are abstracted into three association conditions in this embodiment of the invention: semantic basic association, spatial proximity association, and dynamic influence association. Therefore, the topological analysis process here is to determine whether two nodes satisfy these three association conditions. When it is determined that two nodes satisfy at least one association condition based on the node attribute values, it is determined that the two nodes have a topological association.

[0057] In one possible implementation, the association condition is determined in the following way, which is explained in detail below.

[0058] When the semantic category identifiers of two nodes determine that either target condition is met, the association condition satisfied by the two nodes is determined to be the semantic basic association. The Euclidean distance between the two nodes is determined based on their spatial location parameters. When the Euclidean distance is less than a preset Euclidean distance threshold, the association condition satisfied by the two nodes is determined to be spatial proximity association. When an abnormal device state occurs in either of the two nodes, causing a change in the device state of the other node, the association condition that the two nodes satisfy is determined to be a dynamic influence association.

[0059] Here, the first association condition is the semantic basic association. The basis for judging this association condition is to determine whether the two nodes meet any one of the target conditions. The target conditions include: the two nodes are in the same functional area in the cabin structure scene; the two nodes have a relationship of belonging to equipment objects and space units in the cabin structure scene; and the two nodes have a relationship of equipment function dependency in the cabin structure scene.

[0060] Here, functional areas can be divided according to the functions of different areas in a specific cabin structure scenario, such as the power compartment area, control compartment area, and energy management area. If the equipment objects corresponding to two nodes are in the same power compartment area or the same control compartment area, then the target condition can be determined to be met. The affiliation relationship between equipment objects and space units determines whether an affiliation exists between them. For example, if a piece of equipment is fixedly installed in a certain area, then the equipment has an affiliation relationship with that area. If it is determined that two nodes have this affiliation relationship, then the target condition can be determined to be met. Equipment functional dependency determines whether a functional dependency exists between two equipment objects, such as power supply equipment and power consumption equipment, or control device and controlled device. If it is determined that two nodes have this functional dependency, then the target condition can be determined to be met.

[0061] In the specific judgment, if two nodes satisfy at least one target condition, it can be determined that there is a semantic basis association between the two nodes. Conversely, if none of the three target conditions are satisfied, it is determined that there is no semantic basis association between the two nodes.

[0062] The second association condition is spatial proximity association, which is determined by the spatial distance between two nodes. The Euclidean distance between the two nodes is calculated based on the spatial location parameters in their attribute values, and is used to measure the spatial proximity. If the Euclidean distance is less than a preset Euclidean distance threshold, the two nodes are determined to be spatially proximity in the cabin structure scene, and the association condition they satisfy is spatial proximity association. Conversely, if the Euclidean distance is not less than the preset Euclidean distance threshold, the two nodes are determined not to be spatially proximity association.

[0063] The third association condition is dynamic influence association. This condition is determined by whether a node's abnormal state will propagate to another node. Specifically, if either node experiences an abnormal state (i.e., its state parameters exceed the preset normal operating range, such as a temperature exceeding 40 degrees Celsius), the functional dependencies or physical connections between the two nodes will cause a change in the state of the other node. For example, if power supply device A experiences an abnormal state, it will cause a change in the state of power consumer device B. If this propagation effect is determined to exist between the two nodes, it indicates a dynamic influence association, thus confirming the existence of such an association.

[0064] In this embodiment of the invention, three association conditions are designed in terms of semantic association, spatial distance, and dynamic influence to realize the analysis of topological association between nodes, providing a reliable theoretical basis for subsequent adjustment and simplification of topological association relationships, inspection task scheduling, and path planning.

[0065] If any two nodes in a topological graph satisfy at least one association condition, a topological association can be determined. To measure the degree of this association, this invention defines a topological association strength, which is determined by the level of the topological association strength. The level of topological association strength is positively correlated with the number of association conditions satisfied. Topological association strength includes: strong association, medium association, weak association, and no association.

[0066] Specifically, when two nodes simultaneously satisfy all three association conditions, the topological association strength between the two nodes is determined to be strong; when two nodes simultaneously satisfy only any two association conditions, the topological association strength between the two nodes is determined to be medium; when two nodes satisfy only any one association condition, the topological association strength between the two nodes is determined to be weak; and when two nodes simultaneously do not satisfy any one association condition, the topological association strength between the two nodes is determined to be no association.

[0067] Another aspect of topology correlation analysis is determining the criticality level of each node. For any node in the topology graph, the main focus is on whether the node meets certain factor conditions, and then logical combinations are used to determine the criticality level. There are four factor conditions, and the topology analysis process involves determining whether each node meets these four factor conditions. The factor conditions include: the semantic category identifier of the node indicates that the equipment object corresponding to the node belongs to a preset critical equipment or is located in a preset core area in the cabin structure scenario; the status parameters of the equipment corresponding to the node are within a preset normal parameter range; the number of nodes with spatial proximity to the node exceeds a preset threshold; and the topology correlation strength between the node and its neighboring nodes is strong or medium. These will be explained one by one below.

[0068] The first factor condition is whether the node is a pre-defined key device or belongs to a core area. For example, it could be a pre-defined key device such as a power generation device or controller, or a pre-defined core area such as an energy management area. If so, this factor condition is met; otherwise, it is not. The second factor condition is whether the node's equipment status is abnormal. For example, whether equipment status parameters such as temperature, pressure, humidity, voltage, amplitude, and gas concentration are within the normal operating range. If so, this factor condition is met; otherwise, it is not. The third factor condition is whether the node has many connections with other nodes, i.e., whether the number of spatially adjacent nodes exceeds a pre-defined threshold. If so, this factor condition is met; otherwise, it is not. The fourth factor condition is whether the node has a strong or moderate topological association with its neighboring nodes. If so, this factor condition is met; otherwise, it is not.

[0069] After the factor conditions are determined for each node, the number of factor conditions that are met can be counted. The criticality level is positively correlated with the number of factor conditions that are met; therefore, the criticality level of a node can be determined based on the number of factor conditions that are met. Criticality levels include: high criticality, medium criticality, low criticality, and non-criticality.

[0070] Specifically, when a node satisfies all four factor conditions, its criticality level is determined to be high criticality; when a node satisfies only any three factor conditions, its criticality level is determined to be medium criticality; when two nodes satisfy only any two association conditions, their criticality level is determined to be low criticality; and when two nodes satisfy only one association condition or none of the four factor conditions are met, their criticality level is determined to be non-critical.

[0071] Finally, through the above topology association analysis, the analysis results are output, including the criticality level of the nodes, which characterizes the importance of the nodes in the cabin structure scenario. In addition, the analysis results also include a ranking matrix corresponding to the topology association strength between nodes, and the equipment status parameters of all nodes, which serve as the basis for subsequent topology map adjustments.

[0072] This invention, through a topological correlation analysis mechanism based on semantic association, spatial proximity, and dynamic state influence, achieves effective reasoning regarding implicit relationships and dynamic influences between nodes. Furthermore, by setting factor conditions to measure the importance of individual nodes, it addresses the problem in existing technologies that rely solely on geometric relationships and lack semantic and dynamic correlation analysis when analyzing topological graphs.

[0073] Furthermore, the topology diagram is adjusted based on the analysis results to obtain the robot's operational scenario model.

[0074] Here, adjusting the topology graph specifically involves adjusting the set of nodes, the set of edges, and updating the set of attributes, including retaining, merging, splitting, or removing nodes. The basis for adjusting the topology graph comes from the analysis results of topological association analysis, namely the criticality level of nodes and the strength of topological associations between nodes.

[0075] In one possible implementation, the topology is adjusted based on the analysis results to obtain the robot's operational scenario model. This can be achieved in the following ways, which are explained in detail below.

[0076] If two nodes in the topology graph are strongly associated, and they are located in the same functional area or have a relationship of belonging to equipment objects and space units in the cabin structure scene, then the two nodes are merged. Nodes in the topology graph with more than two sub-regions are identified as nodes to be partitioned. When the standard deviation of the device state parameters corresponding to the node to be partitioned exceeds the preset heterogeneity threshold and there is no physical separation between the sub-regions, the node to be partitioned is divided into multiple homogeneous sub-nodes according to the clustering results of the device state parameters. Determine two non-adjacent nodes in the topology graph that have a strong or medium topological association. If the non-adjacent nodes have equipment function dependencies or dynamic influence associations in the cabin structure scenario, establish an edge between the non-adjacent nodes. When the topological association strength between two nodes in the topological graph is no association, delete the edge between the two nodes and delete the nodes in the topological graph that have a non-criticality level.

[0077] First, for node merging, two nodes with strong topological association can be retrieved in the topology graph. If these two strongly associated nodes are in the same functional area or have a relationship of belonging to equipment objects and space units in the cabin structure scenario, such as being in the same energy management area or the equipment being fixed in a certain area, then these two nodes can be merged to facilitate the simultaneous allocation of inspection tasks.

[0078] It should be noted that after the nodes are merged, the attribute values ​​of the two nodes are also merged accordingly. The merged node inherits the spatial location parameters, geometric range parameters, and semantic category identifiers of the original two nodes.

[0079] For node partitioning, nodes with more than two sub-regions in a spatial region can be searched in the topology graph and identified as nodes to be partitioned. Then, it is further determined whether the standard deviation of the device state parameters corresponding to the node to be partitioned exceeds a preset heterogeneity threshold, and whether there are physical obstructions or splits in these sub-regions. For example, if a power supply device is located in three different rooms, and its voltage standard deviation exceeds the preset heterogeneity threshold, and the three rooms are interconnected by doors without physical barriers, then the node of this power supply device can be partitioned into three homogeneous sub-nodes. That is, this power supply device is abstracted into three sub-devices. The partitioning is based on clustering the device state parameters of the node using clustering algorithms such as K-means, and then partitioning the nodes into the corresponding number of clusters based on the clustering results. In subsequent inspection task allocation, each of these partitioned nodes is independently assigned an inspection task.

[0080] Adding edges involves retrieving two non-adjacent nodes with strong or moderate topological association in the topology graph. If these two nodes have a functional dependency or dynamic influence relationship within the cabin structure scenario, an edge is established between them, creating spatial connectivity. This allows the robot to plan appropriate paths during inspections, and its assigned inspection tasks can be executed sequentially.

[0081] For edge and node deletion, the process involves retrieving two nodes with no topological association and nodes with a non-criticality level. If the topological association is no, there's no need to establish a spatial connection between these two nodes for path planning; therefore, the edge between them is removed to avoid planning unnecessary redundant paths and simplify the path. Non-critical nodes indicate that unimportant equipment objects do not require inspection, and therefore no inspection tasks or path planning are needed; these nodes are deleted.

[0082] After adjusting the nodes and edges of the topology graph based on the analysis results, each newly determined node and edge is traversed sequentially, updating the node and edge attribute values. For example, for a newly added edge, its distance parameter is calculated based on the spatial location parameters of the two nodes on that edge, its passage difficulty parameter is determined based on the physical passage conditions between the two nodes, and its safety risk parameter is determined based on the state parameters and dynamic influence relationships of the two nodes. After updating the edges and nodes, the topology graph can serve as the operational scenario model corresponding to the cabin structure scene, at which point the scenario modeling is complete. The operational scenario model facilitates subsequent inspection task scheduling and path planning.

[0083] In this embodiment of the invention, by adopting a technical means of dynamically adjusting the node set, edge set, and attribute set based on the results of topological correlation analysis, the technical effect of enabling the operation scenario model to adaptively update with environmental changes is achieved, thereby solving the problem of static models in the prior art that cannot adapt to environmental changes.

[0084] Step 104: Prioritize the robot's inspection tasks according to the criticality level of the nodes in the task scenario model to obtain the task queue.

[0085] Here, the task queue is the core scheduling basis for robot inspection operations. Subsequent path planning, dynamic obstacle avoidance, and priority adjustment all revolve around this task queue. The priority scheduling of inspection tasks is based on the criticality level of nodes in the work scenario model. The higher the criticality level of a node, the higher its scheduling priority, and the corresponding inspection task is executed first. Conversely, the lower the scheduling priority, the lower the corresponding inspection task is executed later.

[0086] In one possible implementation, the robot's inspection tasks are prioritized and scheduled according to the criticality level of nodes in the task scenario model to obtain a task queue. This can be achieved in the following ways, which are explained in detail below.

[0087] Nodes are sorted according to their criticality level in the scene model to obtain the sorting results. Nodes with the same criticality level are sorted according to the strength of their topological association. The robot is assigned inspection tasks for each node according to the sorting results. The inspection tasks of adjacent nodes with strong topological association in the sorting results are merged. The inspection tasks of adjacent nodes with the same criticality level in the sorting results are scheduled for execution according to the topological association strength. The inspection tasks are sequentially assigned and merged or executed in a sequential scheduling order, and then added to the task scheduling queue to obtain the task queue.

[0088] First, all nodes in the task scenario model are traversed to determine the criticality level of each node. Then, the nodes are sorted in descending order of their criticality level, with high-criticality nodes ranked first, followed by medium-criticality nodes, and then low-criticality nodes. Non-critical nodes are not involved in the inspection task scheduling and are not included in the ranking. For nodes with the same criticality level, they are further ranked according to the strength of their topological associations. For example, nodes with strong associations with other nodes are ranked first, followed by nodes with medium associations, and then nodes with weak associations. After the ranking is completed, an inspection task is obtained, and the robot is then assigned inspection tasks to each node according to this ranking. Inspection tasks include equipment status detection, environmental parameter acquisition, abnormal area confirmation, and routine inspection tasks.

[0089] Furthermore, considering that some inspection tasks can be executed simultaneously, after the task scheduling is completed, it is checked whether there are adjacent nodes with strong topological association in the sorting results. For such adjacent nodes, their corresponding inspection tasks can be merged. Since the two inspection tasks have the same execution priority, there is no need to perform additional path planning for these two nodes, which can reduce the number of times the robot moves and stops repeatedly.

[0090] In addition, it is possible to check whether there are adjacent nodes with the same criticality level in the sorting results. Because there are still adjacent nodes with the same criticality level in the sorting results, the inspection tasks of these adjacent nodes can be scheduled according to the strength of the topological association. For example, the inspection tasks of nodes with strong associations are sorted first, followed by those with medium associations, and then those with weak associations. This ensures that the inspection tasks of more important nodes are executed first.

[0091] After the inspection tasks are assigned, they will be sequentially assigned and merged or scheduled according to the order of execution. These tasks will be added to the task scheduling queue in strict accordance with the assignment or scheduling order, thus forming the task queue. Based on the first-in, first-out (FIFO) principle of the queue, the inspection tasks corresponding to the nodes with the highest scheduling priority are at the head of the queue and will be executed first, while the inspection tasks corresponding to the nodes with the lowest scheduling priority are at the tail of the queue and will be executed last.

[0092] In this embodiment of the invention, during the scheduling of inspection tasks, tasks are strictly sorted according to the topological association strength between nodes and the criticality level of nodes. This enables adaptive scheduling of inspection tasks in complex cabin environments and solves the problem of low operational efficiency caused by relying solely on path planning and lacking global task scheduling capabilities in the prior art.

[0093] Step 105: Perform path planning for the robot based on the task queue to obtain the inspection path, and control the robot to perform inspections in the cabin structure scenario based on the inspection path.

[0094] After generating the task queue through task scheduling, in order to support the efficient execution of the task queue, it is necessary to plan the inspection path for the robot, determine the optimal inspection path for executing each inspection task sequentially from the current position node, that is, determine the inspection path from the current position node to the target task node, which is the end point of the robot's inspection. The details are explained below.

[0095] For the constructed operation scenario model, the node set, edge set, and attribute set are extracted from the model as the basic data for robot adaptive operation. The node set corresponds to the spatial semantic unit in the cabin, including spatial region nodes and equipment object nodes. Each node carries spatial position parameters, geometric range parameters, and semantic category identifiers. The edge set corresponds to the passage relationship and association relationship between units. Each edge carries distance parameters, passage difficulty parameters, and safety risk parameters. The attribute set contains a complete description of node attributes and edge attributes.

[0096] During path planning, the current location node is matched and determined in the work scenario model based on the robot's real-time positioning information, while the target task node is calibrated in the work scenario model according to the inspection task requirements. For candidate paths from the current location node to the target task node, all possible node sequences in the model are traversed. In each node sequence, there must be an edge connection between adjacent nodes, and the traversal difficulty parameter of the edge must be lower than the preset maximum traversal difficulty threshold.

[0097] For each candidate path that meets the conditions, the total travel cost of the sequence is calculated. The total travel cost is determined by comprehensively considering the distance parameters, travel difficulty parameters, and security risk parameters of all edges in the sequence. A larger edge distance results in a larger distance parameter based on the linear mapping, and vice versa. A higher travel difficulty results in a larger travel difficulty parameter based on the linear mapping, and vice versa. Similarly, a higher security risk results in a larger security risk parameter based on the linear mapping, and vice versa. The total travel cost is calculated by summing the distance parameters, travel difficulty parameters, and security risk parameters.

[0098] Next, the total travel cost of the node sequences corresponding to all candidate paths that meet the conditions is compared, and the candidate path with the node sequence that has the lowest total travel cost is selected as the robot's inspection path. At the same time, the inspection path segments corresponding to all tasks in the task queue are spliced ​​together in task order to form a complete inspection path. The node sequence and edge sequence in the inspection path are recorded as the navigation basis for the robot's actual movement.

[0099] If there is no node sequence that meets the conditions (i.e., no candidate path can be determined from the current position to the target task node), the current inspection task is marked as unfinishable, triggering the task rescheduling process, postponing or removing the unreachable inspection task, and regenerating the task queue.

[0100] Once the robot's inspection path is determined, it can be controlled to inspect the cabin structure based on that path. However, sudden anomalies may occur during the inspection process, requiring timely anomaly handling and rescheduling of the inspection task.

[0101] In one possible implementation, when the inspection path-controlled robot is inspecting in a cabin structure scenario, it also includes: If an abnormal node with abnormal equipment status parameters is detected in the work scenario model, the criticality level of the abnormal node is re-determined. When the criticality level of an abnormal node is determined to be increased, the inspection task of the abnormal node will be scheduled to the head of the task queue. When the criticality level of an abnormal node is determined to be reduced, the inspection task of the abnormal node will be scheduled to the end of the task queue or deleted from the task queue. When an abnormal node with abnormal equipment status parameters is detected in one of the two nodes of the merged inspection task in the operation scenario model, the criticality level of the abnormal node is re-determined. The execution priority of the inspection tasks of the abnormal nodes is evaluated based on the criticality level, and the merged inspection tasks are split or the execution priority of the inspection tasks of the abnormal nodes is re-scheduled based on the evaluation results.

[0102] Here, since the analysis results output during topology correlation analysis include the device status parameters of each node, these parameters can be monitored in real time when the robot is performing inspections along the inspection path. Abnormal device status parameters refer to situations where the device object corresponding to a node or the environmental state it is in exceeds the preset normal operating range (i.e., parameter range). Device status parameters include: device temperature, device vibration amplitude, current, voltage, pressure, humidity, harmful gas concentration, and device fault status codes. The specific judgment method is as follows: For any device status parameter, let its normal operating range be... When satisfied If this occurs, the status parameter is determined to be abnormal. and This indicates the lower and upper limits of the parameter range. To avoid misjudgments caused by instantaneous fluctuations, the device is only deemed to be in an abnormal state when its status parameters exceed the normal operating range for multiple consecutive sampling periods.

[0103] If the device status parameters of a node in the work scenario model are found to be outside the preset parameter range, such as a device temperature suddenly dropping below 25 degrees Celsius or a voltage exceeding 220V, then the node is determined to be an abnormal node with abnormal device status parameters.

[0104] Based on abnormal nodes, the anomaly handling mechanism here first redefines the criticality level of the abnormal node. If its criticality level is found to have increased, for example, from medium criticality to high criticality, the inspection task for the abnormal node is scheduled to the head of the task queue, meaning its inspection task is executed first. Conversely, if the criticality level of the abnormal node is determined to have decreased, for example, from medium criticality to low criticality, its inspection task is scheduled to the tail of the task queue, meaning its execution is postponed. Alternatively, the inspection task for the abnormal node can be directly deleted from the task queue, meaning its execution is unnecessary, simplifying the inspection process.

[0105] If an abnormal node appears in one of two nodes in a merged inspection task, the exception handling mechanism remains the same: first, the criticality level of the abnormal node is reassessed. Since the criticality level may change, the execution priority of the inspection task for the abnormal node needs to be evaluated. For example, if two high-criticality nodes have their inspection tasks merged with the same execution priority, but one node experiences an anomaly, causing its criticality level to drop to medium criticality, then the execution priority evaluation determines that the scheduling priority of the abnormal node has decreased. In this case, the merged inspection task is either split back into the original two inspection tasks, or the execution priority of the abnormal node's inspection task is re-scheduled, i.e., re-scheduled based on its criticality level, adjusting the execution priority of the abnormal node's inspection task.

[0106] In this embodiment of the invention, when the inspection path control robot is inspecting in a cabin structure scenario, it can achieve rapid response to abnormalities in the event of equipment malfunctions and perform real-time scheduling of inspection tasks, thereby further improving the global scheduling capability of inspection tasks and solving the problem of existing technologies that rely solely on path planning and lack global task scheduling capabilities.

[0107] Furthermore, when the inspection path-controlled robot inspects a cabin structure scenario, the node criticality and edge attribute parameters in the operation scenario model can be adaptively updated based on historical execution data during the inspection process. Specifically, the path execution results, task completion status, and anomaly handling records of the robot in each inspection operation are recorded to form a historical execution dataset. For each inspection path in the historical execution dataset, the difference between the total passage cost and the actual execution time of the inspection path is determined. If the actual execution time is consistently higher than the expected time corresponding to the theoretical passage cost, it is determined that the passage difficulty parameter on that inspection path needs to be corrected. Based on the ratio of the actual execution time to the expected time, the passage difficulty parameters of each edge on that inspection path are gradually adjusted. For each inspection task in the historical execution dataset, the actual execution frequency and anomaly triggering frequency of the task are analyzed. If the anomaly triggering frequency of a certain node is higher than a preset anomaly frequency threshold, the criticality level of that node is increased by one level, and the inspection task corresponding to that node is given priority in subsequent task scheduling.

[0108] Finally, based on the dynamic optimization results, the edge attribute parameters and node criticality levels in the operation scenario model can be updated to ensure that the operation scenario model can adaptively reflect the actual operating characteristics and operation requirements of the cabin environment.

[0109] In summary, this invention digitizes equipment objects and spatial units in a cabin structure scene into nodes, then establishes edges between nodes based on spatial connectivity. Through a topology graph, it transforms a complex cabin environment into a computable model with spatial semantics and structural relationships, thus solving the problem of complex but difficult-to-model environments in existing technologies. Based on the topology graph, this invention analyzes the topological relationships between equipment objects and spatial units through topological association analysis. This effectively uncovers semantic and spatial topological relationships and identifies the importance of equipment, measuring it through a criticality level. Then, based on the importance of the equipment objects, the topological relationships between them are adjusted to perform scene modeling, facilitating task scheduling and path planning. To ensure the scheduling capability of inspection tasks, inspection tasks are scheduled first based on the importance of equipment objects within the scene model, rather than directly planning paths. This establishes path planning on top of inspection task scheduling, enabling robot inspection to prioritize more important equipment objects and tasks based on their importance and topological relationships, improving the overall inspection efficiency of the cabin structure scene.

[0110] The following describes in detail the robot adaptive operation device for complex cabin structure scenarios provided by the present invention.

[0111] like Figure 2 As shown, the robot adaptive operation device for complex cabin structure scenarios specifically includes: a data acquisition module 201, a topology construction module 202, a topology analysis module 203, a task scheduling module 204, and a path planning module 205.

[0112] Specifically, the data acquisition module 201 is used to acquire multi-source sensor data collected by the robot from the cabin structure scene, and identify equipment objects and spatial units in the cabin structure scene based on the multi-source sensor data; the topology construction module 202 is used to construct a topology graph corresponding to the cabin structure scene, wherein the nodes in the topology graph are the equipment objects or spatial units, and the edges connecting the nodes are the spatial connectivity relationships between the nodes; the topology analysis module 203 is used to perform topology association analysis on the nodes and edges in the topology graph, and adjust the topology graph according to the analysis results to obtain the robot's operation scene model, wherein the analysis results include the criticality level of the nodes, and the criticality level is used to characterize the importance of the nodes in the cabin structure scene; the task scheduling module 204 is used to prioritize the robot's inspection tasks according to the criticality level of the nodes in the operation scene model; the path planning module 205 is used to perform path planning for the robot according to the task queue to obtain the inspection path, and control the robot to inspect in the cabin structure scene based on the inspection path.

[0113] The robot adaptive operation device for complex cabin structure scenarios provided in the above embodiments can realize the technical solutions described in the above embodiments of the robot adaptive operation method for complex cabin structure scenarios. The specific implementation principles of each module or unit can be found in the corresponding content in the above embodiments of the robot adaptive operation method for complex cabin structure scenarios, and their technical effects can also be referred to each other, which will not be repeated here.

[0114] like Figure 3 As shown, the present invention also provides an electronic device 300. The electronic device 300 includes a processor 301, a memory 302, and a display 303. Figure 3 Only some components of the electronic device 300 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0115] In some embodiments, memory 302 may be an internal storage unit of electronic device 300, such as a hard disk or memory of electronic device 300. In other embodiments, memory 302 may also be an external storage device of electronic device 300, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on electronic device 300.

[0116] Furthermore, the memory 302 may include both internal storage units of the electronic device 300 and external storage devices. The memory 302 is used to store application software and various types of data installed on the electronic device 300.

[0117] In some embodiments, processor 301 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 302 or process data, such as the robot adaptive operation method for complex cabin structure scenarios in this invention.

[0118] In some embodiments, display 303 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 303 is used to display information from electronic device 300 and to display a visual user interface. Components 301-303 of electronic device 300 communicate with each other via a system bus.

[0119] In some embodiments of the present invention, when the processor 301 executes the computer program in the memory 302, the following steps can be implemented: acquiring multi-source sensor data collected by the robot from the cabin structure scene, and identifying equipment objects and spatial units in the cabin structure scene based on the multi-source sensor data; constructing a topology graph corresponding to the cabin structure scene, wherein the nodes in the topology graph are the equipment objects or the spatial units, and the edges connecting the nodes are the spatial connectivity relationships between the nodes; performing topological association analysis on the nodes and edges in the topology graph, and adjusting the topology graph according to the analysis results to obtain a robot operation scene model, wherein the analysis results include the criticality level of the nodes, and the criticality level is used to characterize the importance of the nodes in the cabin structure scene; prioritizing the robot's inspection tasks according to the criticality level of the nodes in the operation scene model to obtain an operation task queue; performing path planning on the robot according to the operation task queue to obtain an inspection path, and controlling the robot to inspect in the cabin structure scene based on the inspection path.

[0120] It should be understood that when the processor 301 executes the computer program in the memory 302, in addition to the functions described above, it can also perform other functions, as can be found in the description of the corresponding method embodiments above.

[0121] Furthermore, the embodiments of the present invention do not specifically limit the type of electronic device 300 mentioned. Electronic device 300 can be a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, laptop computer, or other portable electronic device. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic device can also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, electronic device 300 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0122] In another aspect, the present invention also provides a robot adaptive operation system for complex cabin structure scenarios, including: sensors, communication devices, and electronic devices 300 in the above embodiments; both the sensors and electronic devices 300 are mounted on the robot; the sensors are connected to the electronic devices 300 through the communication devices to collect multi-source sensor data in cabin structure scenarios, and the communication devices are used to store the multi-source sensor data in the memory 302 of the electronic devices 300.

[0123] In some embodiments, sensors mounted on the robot may include scientific instruments for capturing images, industrial cameras, depth cameras, LiDAR, long-range radar, high-end scanners, cameras, inertial navigation systems, and inertial measurement units (IMUs). Communication equipment in some embodiments may be network communication devices deployed with Internet or Ethernet network links, such as switches, routers, firewalls, gateways, bridges, repeaters, wireless access points (APs), modems, optical transceivers, and fiber optic transceivers.

[0124] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0125] The above provides a detailed description of the robot adaptive operation method and apparatus for complex cabin structure scenarios provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A robot adaptive operation method for complex cabin structure scenarios, characterized in that, include: The robot acquires multi-source sensor data collected from the cabin structure scene, and identifies equipment objects and spatial units in the cabin structure scene based on the multi-source sensor data; Construct a topology graph corresponding to the cabin structure scene, wherein the nodes in the topology graph are the equipment objects or the space units, and the edges connecting the nodes are the spatial connectivity relationships between the nodes; A topological association analysis is performed on the nodes and edges in the topological graph, and the topological graph is adjusted according to the analysis results to obtain the robot's operation scenario model. The analysis results include the criticality level of the nodes, which is used to characterize the importance of the nodes in the cabin structure scenario. Based on the criticality level of the nodes in the work scenario model, the robot's inspection tasks are prioritized and scheduled to obtain a task queue. The robot is path-planned according to the task queue to obtain the inspection path, and the robot is controlled to inspect the cabin structure scene based on the inspection path.

2. The robot adaptive operation method for complex cabin structure scenarios according to claim 1, characterized in that, The topology diagram corresponding to the constructed cabin structure scenario includes: Calculate the union of the set of device objects corresponding to the device object and the set of spatial units corresponding to the spatial unit; Map the elements in the union set to nodes; When there is a geometric adjacency or passage constraint relationship between the nodes, an edge is established between the nodes; Based on the nodes and edges, a topology graph corresponding to the cabin structure scene is constructed.

3. The robot adaptive operation method for complex cabin structure scenarios according to claim 1, characterized in that, The topological association analysis of the nodes and edges in the topological graph includes: For any two nodes in the topology graph, when the two nodes satisfy at least one association condition based on the node attribute values, it is determined that the two nodes have a topological association. The node attribute values ​​include semantic category identifiers and spatial location parameters. The association conditions include: semantic basic association, spatial proximity association, and dynamic influence association. The strength of topological association is determined based on the number of association conditions satisfied. The level of the strength of topological association is positively correlated with the number of association conditions satisfied. The strength of topological association includes: strong association, medium association, weak association, and no association. For a node in the topology graph, the criticality level of the node is determined based on the factor conditions satisfied by the node. The criticality level is positively correlated with the number of factor conditions satisfied. The criticality levels include: high criticality, medium criticality, low criticality, and non-criticality. The factor conditions include: The semantic category identifier of the node indicates that the device object corresponding to the node belongs to a preset key device or is located in a preset core area in the cabin structure scenario; The device status parameters corresponding to the node are within the preset normal parameter range; The number of nodes that are spatially adjacent to the node exceeds a preset threshold. The topological association strength between the node and its neighboring nodes is either strong or medium.

4. The robot adaptive operation method for complex cabin structure scenarios according to claim 3, characterized in that, The association conditions are determined in the following way: When it is determined that any one of the target conditions is met based on the semantic category identifiers of the two nodes, the association condition satisfied by the two nodes is determined as the semantic basic association. The target conditions include: the two nodes are in the same functional area in the cabin structure scene, the two nodes have a relationship of belonging between equipment objects and space units in the cabin structure scene, and the two nodes have a relationship of equipment function dependency in the cabin structure scene. The Euclidean distance between the two nodes is determined based on their spatial location parameters. When the Euclidean distance is less than a preset Euclidean distance threshold, the association condition satisfied by the two nodes is determined to be spatial proximity association. When an abnormal device state occurs in either of the two nodes, causing a change in the device state of the other node, the association condition that the two nodes satisfy is determined to be a dynamic influence association.

5. The robot adaptive operation method for complex cabin structure scenarios according to claim 1, characterized in that, The step of adjusting the topology map based on the analysis results to obtain the robot's operational scenario model includes: Two nodes with a strong topological association in the topology graph are identified. If the two nodes are in the same functional area in the cabin structure scene or have a relationship of belonging to equipment objects and space units, the two nodes are merged. Nodes in the topology graph that have more than two sub-regions are identified as nodes to be partitioned. When the standard deviation of the device state parameters corresponding to the node to be partitioned exceeds a preset heterogeneity threshold and there is no physical separation between the sub-regions, the node to be partitioned is divided into multiple homogeneous sub-nodes according to the clustering results of the device state parameters. Determine two non-adjacent nodes in the topology graph that have a strong or medium topological association. If the non-adjacent nodes have a device function dependency or a dynamic influence association in the cabin structure scenario, establish an edge between the non-adjacent nodes. When the topological association strength between two nodes in the topological graph is no association, delete the edge between the two nodes and delete the nodes in the topological graph whose criticality level is non-critical.

6. The robot adaptive operation method for complex cabin structure scenarios according to claim 1, characterized in that, Based on the criticality level of nodes in the aforementioned work scenario model, the robot's inspection tasks are prioritized and scheduled to obtain a task queue, including: The nodes are sorted according to their criticality level in the scenario model to obtain the sorting result. Nodes with the same criticality level are sorted according to the strength of their topological association. The robot is assigned inspection tasks for each node according to the sorting results. The inspection tasks of adjacent nodes with strong topological association in the sorting results are merged. The inspection tasks of adjacent nodes with the same criticality level in the sorting results are scheduled for execution according to the topological association strength. The inspection tasks, after being allocated and merged or scheduled in sequence, are added to the task scheduling queue to obtain the task queue.

7. The robot adaptive operation method for complex cabin structure scenarios according to claim 1, characterized in that, When the robot, controlled by the inspection path, performs an inspection in a cabin structure scenario, the method further includes: If an abnormal node with abnormal equipment status parameters is detected in the work scenario model, the criticality level of the abnormal node is re-determined. When it is determined that the criticality level of the abnormal node has increased, the inspection task of the abnormal node is scheduled to the head of the task queue. When it is determined that the criticality level of the abnormal node has decreased, the inspection task of the abnormal node is scheduled to the end of the task queue or the inspection task of the abnormal node is deleted from the task queue. When an abnormal node with abnormal equipment status parameters is detected in one of the two nodes of the merged inspection task in the operation scenario model, the criticality level of the abnormal node is re-determined. The execution priority of the inspection tasks of the abnormal nodes is evaluated based on the criticality level, and the merged inspection tasks are split or the execution priority of the inspection tasks of the abnormal nodes is re-scheduled based on the evaluation results.

8. A robot adaptive operation device for complex cabin structure scenarios, characterized in that, include: The data acquisition module is used to acquire multi-source sensor data collected by the robot from the cabin structure scene, and to identify equipment objects and space units in the cabin structure scene based on the multi-source sensor data; The topology construction module is used to construct a topology graph corresponding to the cabin structure scene, wherein the nodes in the topology graph are the equipment objects or the spatial units, and the edges connecting the nodes are the spatial connectivity relationships between the nodes; The topology analysis module is used to perform topological association analysis on the nodes and edges in the topology graph, and adjust the topology graph according to the analysis results to obtain the robot's operation scenario model. The analysis results include the criticality level of the nodes, which is used to characterize the importance of the nodes in the cabin structure scenario. The task scheduling module is used to prioritize the robot's inspection tasks based on the criticality level of the nodes in the operation scenario model. The path planning module is used to plan the robot's path according to the task queue, obtain the inspection path, and control the robot to inspect in the cabin structure scenario based on the inspection path.

9. An electronic device, characterized in that, Including memory and processor, among which, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps of the robot adaptive operation method for complex cabin structure scenarios as described in any one of claims 1 to 7.

10. A robot adaptive operation system for complex cabin structure scenarios, characterized in that, include: Sensors, communication devices, and electronic devices as described in claim 9; Both the sensors and the electronic devices are mounted on the robot. The sensor is connected to the electronic device via the communication device and is used to collect multi-source sensor data in a cabin structure scenario. The communication device is used to store the multi-source sensor data in the memory of the electronic device.

Images