A method and system for intelligent motion control of an inspection robot

Abstract

This invention provides a method and system for intelligent motion control of inspection robots, relating to the field of data processing technology. It constructs a three-dimensional topology of the target power plant, inputs inspection requirement information into an inspection planning model, outputs inspection execution information as a response execution target, collects real-time inspection data for initial identification and early warning, stores execution data for in-depth analysis and early warning, comprehensively analyzes and generates maintenance execution tasks and determines secondary inspection points, executes maintenance tasks and plans re-inspection paths, and performs re-inspection control. This invention solves the technical problems in existing technologies where inspection robot control methods are not flexible enough, lack depth of analysis in inspection planning, and have insufficient scene integration and inspection completeness, resulting in limited inspection execution effectiveness. It performs adaptive inspection planning based on inspection requirement modeling to ensure scene adaptability, and analyzes, processes, and optimizes unexpected situations during the inspection process in a timely manner, achieving flexible and precise control of regional inspections.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and specifically to a method and system for intelligent motion control of an inspection robot. Background Technology

[0002] With the development and market launch of inspection robots, they are gradually replacing manual inspections, ignoring inspection risks, and ensuring standardized and digitalized inspection execution, thus improving inspection efficiency. Currently, the control of inspection robots is mostly based on pre-configured inspection execution mechanisms for targeted execution, without sufficient consideration of inspection scenarios and influencing factors, which affects the final inspection results.

[0003] In existing technologies, the inspection control methods for inspection robots are not flexible enough, the analysis depth of inspection planning is insufficient, and the integration with the scene and the completeness of the inspection are inadequate, resulting in limited inspection execution effects. Summary of the Invention

[0004] This application provides a motion intelligent control method and system for inspection robots, which addresses the technical problems in the prior art where the inspection control methods for inspection robots are not flexible enough, the analysis depth of inspection planning is insufficient, and the integration of scenarios and the completeness of inspection are inadequate, resulting in limited inspection execution effects.

[0005] In view of the above problems, this application provides a method and system for intelligent motion control of an inspection robot.

[0006] In a first aspect, this application provides a method for intelligent motion control of an inspection robot, the method comprising:

[0007] Collect basic information about the target power plant and generate a three-dimensional topology.

[0008] An inspection planning model is built, and the inspection requirement information is input into the inspection planning model. The inspection execution information is output. The three-dimensional topology is embedded in the inspection planning model. The inspection execution information includes the inspection cycle and the inspection route. The inspection route has an inspection mode identifier.

[0009] The inspection cycle and the inspection route are used as response execution targets. Real-time inspection data is collected and a preliminary identification and warning are performed. The preliminary identification and warning are based on the directional warning criteria set by the programming program.

[0010] The real-time inspection data is stored in the database, and in-depth data analysis and early warning are performed.

[0011] Based on the initial identification warning and the in-depth analysis warning, the on-site and back-end work together to generate operation and maintenance tasks and determine secondary inspection points;

[0012] The operation and maintenance task is executed, and based on the secondary inspection points, the shortest route is used as the response target to plan the re-inspection path and perform the re-inspection control of the inspection robot.

[0013] Secondly, this application provides a motion intelligent control system for an inspection robot, the system comprising:

[0014] A structure generation module is used to collect basic information about the target power plant and generate a three-dimensional topology structure.

[0015] The inspection planning module is used to build an inspection planning model, input inspection requirement information into the inspection planning model, and output inspection execution information. The three-dimensional topology is embedded in the inspection planning model, and the inspection execution information includes inspection cycle and inspection route. The inspection route has an inspection mode identifier.

[0016] The primary identification and early warning module is used to collect real-time inspection data and perform primary identification and early warning. The primary identification and early warning is based on the directional early warning criteria set by the programming program.

[0017] The in-depth analysis and early warning module is used to perform data storage on the real-time inspection data and conduct in-depth analysis and early warning of the data.

[0018] The patrol point determination module is used to generate operation and maintenance execution tasks and determine secondary patrol points based on the primary identification warning and the in-depth analysis warning, in conjunction with the on-site and back-end systems.

[0019] The re-inspection control module is used to execute the operation and maintenance tasks. Based on the secondary inspection points, it plans the re-inspection path with the shortest route as the response target and performs re-inspection control of the inspection robot.

[0020] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0021] This application provides a motion intelligent control method for an inspection robot. The method involves collecting basic information about a target power plant and generating a three-dimensional topology. An inspection planning model is built, and inspection requirement information is input into the model. Inspection execution information, including the inspection cycle and inspection route (with an inspection mode identifier), is output. The inspection cycle and route are used as response execution targets. Real-time inspection data is collected, and initial identification and early warning are performed. The initial identification and early warning are based on directional early warning standards set in the programming program. The real-time inspection data is then stored in a database for in-depth data analysis and early warning. Based on the initial identification and early warning... The in-depth analysis and early warning system generates maintenance execution tasks and determines secondary inspection points. These tasks are then executed, and based on the secondary inspection points, the shortest route is used as the response target for re-inspection path planning. This system enables re-inspection control of the inspection robot, addressing the technical problems in existing technologies such as insufficient flexibility in inspection robot control methods, lack of depth in inspection planning analysis, inadequate scenario integration and inspection completeness, which limit the effectiveness of inspection execution. The system models inspection requirements for adaptive inspection planning to ensure scenario adaptability. It also provides timely analysis, processing, and reasonable optimization for unexpected situations during the inspection process, achieving flexible and precise control of regional inspections. Attached Figure Description

[0022] Figure 1 This application provides a schematic flowchart of a motion intelligent control method for an inspection robot;

[0023] Figure 2 This application provides a schematic diagram of the process for adjusting the inspection route in the intelligent motion control method for an inspection robot;

[0024] Figure 3 This application provides a schematic diagram of the self-alarm analysis execution process in the motion intelligent control method for an inspection robot;

[0025] Figure 4 This application provides a schematic diagram of the motion intelligent control system structure for an inspection robot.

[0026] Figure labeling: Structure generation module 11, inspection planning module 12, primary identification and early warning module 13, in-depth analysis and early warning module 14, inspection point determination module 15, re-inspection control module 16. Detailed Implementation

[0027] This application provides a motion intelligent control method and system for inspection robots. It collects basic information about the target power plant, generates a three-dimensional topology, inputs inspection requirement information into an inspection planning model, outputs inspection execution information as a response execution target, collects real-time inspection data and performs initial identification and early warning, stores the real-time inspection data in a database for in-depth analysis and early warning, comprehensively analyzes and generates maintenance execution tasks and determines secondary inspection points, executes maintenance execution tasks and plans re-inspection paths, and performs re-inspection control. This addresses the technical problems in existing technologies where inspection robot control methods are not flexible enough, lack depth of analysis in inspection planning, and have insufficient scene integration and inspection completeness, resulting in limited inspection execution effectiveness.

[0028] Example 1

[0029] like Figure 1 As shown, this application provides a motion intelligent control method for an inspection robot, the method comprising:

[0030] Step S100: Collect basic information about the target power plant and generate a three-dimensional topology;

[0031] Specifically, with the development and market launch of inspection robots, they are gradually replacing manual inspections, ignoring inspection risks, and ensuring standardized and digitalized inspection execution, thereby improving inspection efficiency. This application provides a motion intelligent control method for inspection robots applied to power plant inspection control, enabling comprehensive inspection and management across all areas of the target power plant. Given the high inspection requirements and inherent risks in power plant inspections, using inspection robots for operation can maximize the fulfillment of inspection needs.

[0032] Specifically, the target power plant is the power plant to be inspected by the inspection robot. Basic information such as the configuration and distribution of the target power plant is collected, and a three-dimensional topology structure synchronized with the target power plant is constructed. This three-dimensional topology structure serves as the spatial visualization architecture of the target power plant, fully covering the spatial requirements of inspection and monitoring, facilitating the determination of inspection details. Preferably, the target power plant is divided into regional inspection levels, such as based on inspection risk and difficulty, dividing the target power plant into multiple sub-regions, and identifying the inspection level in the three-dimensional topology network to facilitate subsequent targeted inspection planning.

[0033] Step S200: Build an inspection planning model, input the inspection requirement information into the inspection planning model, and output the inspection execution information. The three-dimensional topology is embedded in the inspection planning model. The inspection execution information includes the inspection cycle and the inspection route. The inspection route has an inspection mode identifier.

[0034] Furthermore, the inspection route includes an inspection mode identifier, and step S200 of this application also includes:

[0035] Step S210: Obtain multi-dimensional inspection modes statistically, including continuous start inspection mode and fixed start inspection mode. The backup equipment configured for different inspection modes is different.

[0036] Step S220: Based on the multi-dimensional inspection mode, and in conjunction with the inspection route, perform inspection requirement matching to determine the node matching mode;

[0037] Step S230: Identify the inspection route based on the node matching mode.

[0038] Specifically, the inspection planning model is a self-built inspection planning analysis model for the target power plant. One feasibility modeling method involves constructing a multi-layered network, including a route planning layer, an information adjustment layer, and a mode configuration layer. The hierarchical execution logic is determined for network layer training, and the three-dimensional graph structure is embedded within it for auxiliary planning analysis. The route planning layer, information adjustment layer, and mode configuration layer are hierarchically linked to generate the inspection planning model. Historical inspection records, including sample inspection requirements and sample inspection execution information, are further invoked. The sample inspection requirements are input into the inspection planning model for model validation. The model validation results are mapped and compared with the inspection execution information to verify the model's analytical accuracy. If the analytical accuracy of the constructed inspection planning model is insufficient, training samples are selectively extracted based on the historical inspection records, and the inspection planning model is trained until convergence, resulting in the completed inspection planning model.

[0039] Furthermore, the time difference, inspection requirements, and data types of inspections at various locations within the target power plant are collected and statistically organized as the inspection demand information. This inspection demand information is input into the inspection planning model, and a primary inspection route is determined based on the three-dimensional topology. This route is then transmitted to the information adjustment layer, where the inspection angles and durations for each node along the path are configured according to the inspection requirements. The primary inspection route is then refined; for example, some power plant equipment requires comprehensive, multi-angle inspection analysis. This information is further transmitted to the mode configuration layer for configuring the inspection mode for each node. Different inspection data types correspond to different inspection modes. For example, for high-risk areas, multi-dimensional data collaborative judgment is required, and the start / stop control of the inspection mode is implemented based on the inspection requirements.

[0040] Specifically, multi-dimensional inspection modes are acquired, such as video inspection mode, which is executed based on the image acquisition equipment of the assembly. This video inspection mode can be used as a continuously activated inspection mode throughout the entire inspection process. Vibration inspection mode, temperature inspection mode, etc., can be executed based on the corresponding sensing equipment of the assembly and can be used as a fixed-start inspection mode, activated when there is an inspection requirement. For different nodes along the inspection route, inspection requirements are matched, such as the operating efficiency and temperature of power plant equipment. The multi-dimensional inspection modes are traversed for matching to determine the node matching mode. The node matching mode is located and identified along the inspection route. Simultaneously, collaborative analysis is performed based on the inspection time difference at each location to determine the inspection cycle that meets the time difference. The inspection route with the inspection mode identifier and the inspection cycle are used as the inspection execution information for model output. By performing inspection execution analysis through modeling, analysis efficiency can be effectively improved, ensuring the accuracy and objectivity of the analysis results.

[0041] Step S300: Using the inspection cycle and the inspection route as the response execution target, collect real-time inspection data and perform preliminary identification and early warning, wherein the preliminary identification and early warning is determined based on the directional early warning standard set by the burning program;

[0042] Step S400: Perform data entry into the database for the real-time inspection data, and conduct in-depth data analysis and early warning.

[0043] Specifically, based on the inspection cycle, the activation time node for a single inspection is determined. The inspection route is used as the inspection execution standard to control the inspection robot to perform area inspections. The start / stop control of backup equipment is based on the inspection mode indicated by the inspection route. The backup equipment is a monitoring and acquisition device that meets the inspection requirements, such as a video acquisition device. As the inspection robot progresses, corresponding real-time inspection data is collected to determine basic judgment standards for direct anomaly detection, such as the temperature of the inspection target. An abnormal temperature threshold is added to the directional early warning standard, which is then programmed into the initial program of the inspection robot and connected to the backup equipment for data interaction and analysis. For the real-time inspection data, the directional early warning standard is called for matching and calibration. For real-time inspection data that does not meet the corresponding directional early warning standard, a preliminary identification and early warning is issued. This preliminary identification and early warning is a data warning issued by the inspection robot for real-time inspections. Furthermore, the real-time inspection data is stored in a database, and auxiliary data processing tools are used for further anomaly analysis and source tracing, such as image enhancement processing and convolutional feature recognition, to identify abnormal inspection data and perform in-depth analysis and early warning. By issuing early warnings in batches, the early warning repair time can be controlled.

[0044] Furthermore, such as Figure 2As shown, step S300 of this application further includes:

[0045] Step S310-1: Receive real-time inspection requests and generate special inspection tasks;

[0046] Step S320-1: Set a time limit for the special patrol task, traverse the patrol route for matching, and determine the appropriate adjustment node based on the real-time patrol status.

[0047] Step S330-1: Insert the special inspection task at the adaptation adjustment node and determine the adjusted inspection route.

[0048] Specifically, during the inspection process based on the aforementioned inspection route, there may be temporarily received inspection tasks. These tasks are arranged according to their timeliness and the actual inspection situation. Specifically, the real-time inspection requirement is temporarily received pending information. This information is semantically analyzed and transformed to generate the special inspection task, and the required task execution time is determined. The special inspection task is then identified. Further timeliness constraints are imposed on the special inspection task, i.e., a limited completion time node. The inspection route is traversed for matching, and the inspection route node with the shortest distance to the area where the special inspection task is located is determined. The start time of the special inspection task is determined by combining the timeliness constraint and the task execution time. If it is an urgent task, comprehensive adjustments are made based on the current real-time inspection situation, and the best inspection node that meets the above inspection requirements is used as the adaptation adjustment node. The adaptation adjustment node is located in the inspection route, and the special inspection task is inserted as the adjusted inspection route. The insertion of the special inspection task is irregular and is determined based on the real-time inspection requirements.

[0049] Furthermore, step S300 of this application also includes:

[0050] Step S310-2: Based on the inspection route, perform path obstacle perception and determine target obstacle information;

[0051] Step S320-2: Perform obstacle avoidance retrieval from the built-in database on the target obstacle information to determine the safe trigger distance and obstacle avoidance method;

[0052] Step S330-2: If the obstacle avoidance method is detour obstacle avoidance, determine the initial adjustment point based on the safe trigger distance, obtain the obstacle avoidance path to locate and cover the inspection route, wherein the obstacle avoidance path includes the adjustment path, inspection speed, and inspection direction;

[0053] Step S340-2: If the obstacle avoidance method is non-detour obstacle avoidance, pause control is performed based on the safe trigger distance.

[0054] Furthermore, if the obstacle avoidance method is non-detour obstacle avoidance, and pause control is performed based on the safe trigger distance, step S340-2 of this application further includes:

[0055] Step S341-2: If the obstacle avoidance method is non-detour obstacle avoidance, generate obstacle warning information and determine whether the safe trigger distance is met;

[0056] Step S342-2: If the safe trigger distance is not met, perform an emergency stop control on the inspection robot;

[0057] Step S343-2: If the safe trigger distance is met, calculate the deceleration and perform deceleration and stop control on the inspection robot.

[0058] Specifically, during the inspection process along the designated route, obstacles inevitably exist, such as moving objects and terrain features. Timely and effective avoidance of these obstacles is necessary. Specifically, the inspection robot is equipped with infrared sensors, enabling it to perceive obstacles at predetermined distances along the inspection route in real time, determining basic information such as the size of the obstacles as the target obstacle information. The built-in database stores obstacle avoidance information within a built-in chip, including obstacle types and selected avoidance methods. This database is periodically updated through learning as the inspection robot performs obstacle avoidance. For the target obstacle information, an obstacle avoidance search is performed in the built-in database to determine the optimal distance between the inspection robot and the target obstacle under the best obstacle avoidance conditions, which is then used as the safe trigger distance. The area where the target obstacle is located is then scanned to match an appropriate obstacle avoidance method.

[0059] Furthermore, the obstacle avoidance method is determined. If the target obstacle is small and the remaining width of the path is sufficient for the inspection robot, or if there are forks in the scanned area of ​​the target obstacle that allow the inspection robot to detour, the obstacle avoidance method is determined to be detour obstacle avoidance. The inspection route location point that meets the safe trigger distance and is far from the target obstacle is further determined as the initial adjustment point. The inspection route within the safe trigger distance is adjusted by combining the target obstacle information and the area scan information, including the adjusted path, the inspection speed, and the inspection direction, as the obstacle avoidance path. The obstacle avoidance path is matched and located within the inspection route, and the original path is intercepted and the obstacle avoidance path is covered. If the safe trigger distance is not met, such as when avoiding obstacles at corners, an emergency stop and in-situ adjustment control are performed.

[0060] If no detour conditions exist for inspection, the obstacle avoidance method is determined to be non-detour obstacle avoidance. For example, if there are obstacles such as road blockages or steps, an obstacle warning message is generated, i.e., a warning response to an obstacle ahead, which requires manual adjustment. Further, it is determined whether the safe trigger distance is met. If the safe trigger distance is not met, it indicates a collision risk during normal deceleration and stopping, and an emergency stop control is executed. Since emergency stop control has a certain impact on the lifespan of the inspection robot, it is only used in special circumstances. If the safe trigger distance is met, a deceleration path is determined based on the real-time inspection position of the inspection robot and the position of the target obstacle. The deceleration is calculated based on the real-time inspection speed of the inspection robot, and deceleration and stopping control are executed on the inspection robot. Inspection is resumed after the target obstacle is cleared. Timely and effective analysis and processing of factors affecting the inspection robot's inspection are conducted to avoid affecting the normal inspection process.

[0061] Furthermore, step S300 of this application also includes:

[0062] Step S310-3: Set a charging threshold, wherein the charging threshold is determined based on the real-time distance between the inspection robot and the workstation, and is dynamically adjusted as the inspection robot is located in real time;

[0063] Step S320-3: If the real-time battery level of the inspection robot is less than or equal to the charging threshold, generate an inspection interruption command;

[0064] Step S330-3: Based on the patrol interruption command, control the patrol robot to perform automatic repatriation, wherein the patrol interruption position serves as the starting point for the subsequent patrol.

[0065] Specifically, during the inspection process of the inspection robot, strict power control is required. The robot's minimum power limit—the lowest power level that will not cause equipment damage—is determined based on the distance between the robot's real-time inspection location and the workstation. This distance is the travel distance, not a straight-line distance. The workstation serves as the robot's charging area. Based on this, the repatriation power consumption is determined. The sum of the minimum power limit and the repatriation power consumption is used as the charging threshold. This charging threshold fluctuates in real-time as the inspection robot progresses and is updated immediately to effectively prevent inspection anomalies caused by insufficient power. The charging threshold is real-time; the robot's real-time power level is compared with the charging threshold. If the real-time power level is greater than the updated charging threshold, the inspection process continues normally. If it is less than or equal to the charging threshold, it indicates insufficient power, and only the repatriation power level can be maintained, generating an inspection interruption command. Upon receiving the patrol interruption command, the patrol robot is controlled to perform automatic repatriation, returning to the workstation to complete charging. At the same time, the interruption location is used as the starting point for the next patrol. After charging is completed, the robot returns to continue patrol execution.

[0066] Step S500: Based on the primary identification warning and the in-depth analysis warning, the on-site and back-end work together to generate operation and maintenance execution tasks and determine secondary inspection points;

[0067] Step S600: Execute the operation and maintenance task, and based on the secondary inspection points, use the shortest route as the response target to plan the re-inspection path and perform the re-inspection control of the inspection robot.

[0068] Specifically, the initial identification warning is the on-site inspection warning of the inspection robot, and the in-depth analysis warning is the subsequent data anomaly analysis and source tracing warning. Based on the initial identification warning and the in-depth analysis warning, the operation and maintenance execution task is generated. The operation and maintenance execution task has a time sequence identifier, and the warning location is used as the secondary inspection point. Further, the operation and maintenance execution task is executed. After the warning is repaired, the secondary inspection point is scheduled and connected. Combining the inspection planning model, the shortest route is used as the response target to determine the re-inspection route. The re-inspection route is used as the inspection standard to control the inspection robot to perform inspection, and the inspection data is further analyzed.

[0069] Furthermore, such as Figure 3 As shown, this application also includes step S700, which includes:

[0070] Step S710: Generate a self-inspection standard list based on the normal execution conditions of the inspection robot;

[0071] Step S720: Embed the self-inspection standard list into the central control module of the inspection robot;

[0072] Step S730: Based on a predetermined time period, the inspection robot periodically performs operational self-checks and generates a self-check dataset, wherein the operational self-checks include self-checks of backup equipment and overall machine operation self-checks;

[0073] Step S740: Map and verify the self-test standard list with the self-test dataset to generate abnormal warning information;

[0074] Step S750: The inspection robot performs self-warning based on the abnormal warning information, wherein the warning execution methods are different for different components and different warning levels.

[0075] Specifically, to ensure the inspection accuracy of the inspection robot, regular self-checks are performed on its operation. The standard inspection status of the inspection robot is statistically analyzed, including path control accuracy and the control precision of backup equipment, serving as the normal execution conditions. The controllability deviations of these normal execution conditions are collected, and a mapping relationship is established between the normal execution conditions and the controllable deviations to generate a self-check standard list. This self-check standard list serves as a reference for the self-check operation of the inspection robot. The self-check standard list is embedded in the central control module of the inspection robot, which is the comprehensive control area of ​​the robot. A predetermined time period is set, i.e., the execution interval for the self-check, which can be customized by referring to the historical operation records of the inspection robot.

[0076] Based on the predetermined time period, the self-inspection of the inspected equipment and the overall machine operation self-inspection are used as the self-inspection execution direction. The inspection robot is controlled to perform operation self-inspection, and the data source of the self-inspection data is identified to generate the self-inspection dataset. Further, the self-inspection dataset is mapped to the self-inspection standard list to determine whether the self-inspection dataset meets the controllable deviation range corresponding to the self-inspection standard list. If it does not meet the standard, it indicates an operational anomaly. The abnormal data is traced to determine the warning component, warning type, and warning level, which are then used as the abnormal warning information. The inspection robot self-alarms based on the abnormal warning information. Specifically, different warning methods are configured for different warning information, such as configuring multiple warning states, such as ringing, flashing lights, etc., including diversity in color, frequency, etc., to characterize the differences in specific warnings and facilitate warning differentiation.

[0077] Example 2

[0078] Based on the same inventive concept as the intelligent motion control method for an inspection robot in the foregoing embodiments, such as Figure 4 As shown, this application provides a motion intelligent control system for an inspection robot, the system comprising:

[0079] Structure generation module 11, which is used to collect basic information of the target power plant and generate a three-dimensional topology structure;

[0080] Inspection planning module 12 is used to build an inspection planning model, input inspection requirement information into the inspection planning model, and output inspection execution information. The three-dimensional topology is embedded in the inspection planning model. The inspection execution information includes inspection cycle and inspection route. The inspection route has an inspection mode identifier.

[0081] The primary identification and early warning module 13 is used to take the inspection cycle and the inspection route as the response execution target, collect real-time inspection data and perform primary identification and early warning, wherein the primary identification and early warning is based on the directional early warning standard set by the burning program.

[0082] The deep analysis and early warning module 14 is used to perform data storage on the real-time inspection data and to perform deep analysis and early warning of the data.

[0083] The patrol point determination module 15 is used to generate operation and maintenance execution tasks in conjunction with the field and the background based on the primary identification warning and the deep analysis warning, and to determine secondary patrol points.

[0084] The re-inspection control module 16 is used to execute the operation and maintenance task. Based on the secondary inspection points, the shortest route is used as the response target to plan the re-inspection path and perform re-inspection control of the inspection robot.

[0085] Furthermore, the system also includes:

[0086] The inspection mode acquisition module is used to statistically acquire multi-dimensional inspection modes, including continuous start inspection mode and fixed start inspection mode. Different inspection modes have different backup equipment configurations.

[0087] The inspection demand matching module is used to match inspection demands based on the multi-dimensional inspection mode and the inspection route, and determine the node matching mode.

[0088] The path identification module is used to identify the inspection route based on the node matching mode.

[0089] Furthermore, the system also includes:

[0090] The task generation module is used to receive real-time inspection requests and generate special inspection tasks.

[0091] The adjustment node determination module is used to limit the timeliness of the special patrol task, traverse the patrol route for matching, and determine the appropriate adjustment node in combination with the real-time patrol status.

[0092] The inspection route adjustment module is used to insert the special inspection task at the adaptation adjustment node and determine the adjusted inspection route.

[0093] Furthermore, the system also includes:

[0094] An obstacle perception module is used to perceive path obstacles based on the inspection route and determine target obstacle information.

[0095] The obstacle avoidance retrieval module is used to perform obstacle avoidance retrieval from the built-in database of the target obstacle information to determine the safe trigger distance and obstacle avoidance method;

[0096] An obstacle avoidance path acquisition module is used to determine an initial adjustment point based on the safe trigger distance if the obstacle avoidance method is detour obstacle avoidance, and to acquire an obstacle avoidance path to locate and cover the inspection route. The obstacle avoidance path includes an adjustment path, inspection speed, and inspection direction.

[0097] A pause control module is used to perform pause control based on the safe trigger distance if the obstacle avoidance method is non-detour obstacle avoidance.

[0098] Furthermore, the system also includes:

[0099] The distance judgment module is used to generate obstacle warning information and determine whether the safe trigger distance is met if the obstacle avoidance method is non-detour obstacle avoidance.

[0100] An emergency stop control module is used to execute an emergency stop control on the inspection robot if the safe trigger distance is not met.

[0101] A deceleration control module is used to calculate the deceleration and perform deceleration control on the inspection robot if the safe trigger distance is met.

[0102] Furthermore, the system also includes:

[0103] A threshold setting module is used to set a charging threshold, wherein the charging threshold is determined based on the real-time distance between the inspection robot and the workstation, and is dynamically adjusted according to the real-time positioning of the inspection robot;

[0104] A power level determination module is used to generate an inspection interruption command if the real-time power level of the inspection robot is less than or equal to the charging threshold.

[0105] The repatriation control module is used to control the inspection robot to perform automatic repatriation based on the inspection interruption command, wherein the inspection interruption position serves as the starting point for the subsequent inspection.

[0106] Furthermore, the system also includes:

[0107] A list generation module is used to generate a self-inspection standard list based on the normal execution conditions of the inspection robot.

[0108] A list embedding module is used to embed the self-inspection standard list into the central control module of the inspection robot;

[0109] The self-test module is used to perform self-tests periodically based on a predetermined time period, generating a self-test dataset. The self-test includes self-tests of backup equipment and self-tests of the entire machine.

[0110] The early warning information generation module is used to map and verify the self-inspection standard list and the self-inspection dataset to generate abnormal early warning information.

[0111] The warning module is used to issue self-warnings to the inspection robot based on the abnormal warning information. The warning execution methods are different for different components and different warning levels.

[0112] Through the foregoing detailed description of a motion intelligent control method for an inspection robot, those skilled in the art can clearly understand the motion intelligent control method and system for an inspection robot in this embodiment. As for the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and relevant parts can be referred to the method section description.

[0113] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for intelligent motion control of an inspection robot, characterized in that, The method includes: Collect basic information about the target power plant and generate a three-dimensional topology. The process of building an inspection planning model includes: constructing a multi-level network layer, including a route planning layer, an information adjustment layer, and a mode configuration layer; determining the hierarchical execution logic for network layer training; connecting and associating the route planning layer, the information adjustment layer, and the mode configuration layer hierarchically to generate the inspection planning model; further calling historical inspection records, including sample inspection requirements and sample inspection execution information; inputting the sample inspection requirements into the inspection planning model for model validation; mapping and verifying the model validation results with the inspection execution information to check the model's analytical accuracy; if the analytical accuracy of the constructed inspection planning model is insufficient, selectively extracting training samples based on the historical inspection records; training the inspection planning model until convergence to obtain the completed inspection planning model; inputting inspection requirement information into the inspection planning model; and outputting inspection execution information. The three-dimensional topology is embedded within the inspection planning model, and the inspection execution information includes the inspection cycle and inspection route, with the inspection route carrying an inspection mode identifier. The inspection cycle and the inspection route are used as response execution targets. Real-time inspection data is collected and a preliminary identification and warning are performed. The preliminary identification and warning are based on the directional warning criteria set by the programming program. The real-time inspection data is stored in the database, and in-depth data analysis and early warning are performed. Based on the initial identification and early warning and the in-depth analysis and early warning, the on-site and back-end work together to generate operation and maintenance tasks and determine secondary inspection points; The operation and maintenance task is executed, and based on the secondary inspection points, the shortest route is used as the response target to plan the re-inspection path and perform the re-inspection control of the inspection robot. The inspection route includes an inspection mode identifier, and the method includes: Statistical data is collected to obtain multi-dimensional inspection modes, including continuous start inspection mode and fixed start inspection mode. The backup equipment configured for different inspection modes is different. Based on the multi-dimensional inspection mode, and combined with the inspection route, inspection requirements are matched to determine the node matching mode. The inspection route is identified based on the node matching pattern.

2. The method as described in claim 1, characterized in that, The methods include: Receive real-time inspection requests and generate special inspection tasks; The special patrol task is subject to a time limit, the patrol route is traversed for matching, and the appropriate adjustment node is determined based on the real-time patrol status. The special inspection task is inserted at the adaptation and adjustment node to determine the adjusted inspection route.

3. The method as described in claim 2, characterized in that, The methods include: Based on the inspection route, path obstacle perception is performed to determine target obstacle information; The target obstacle information is retrieved from the built-in database to determine the safe trigger distance and obstacle avoidance method; If the obstacle avoidance method is detour obstacle avoidance, an initial adjustment point is determined based on the safe trigger distance, and an obstacle avoidance path is obtained to locate and cover the inspection route. The obstacle avoidance path includes the adjustment path, inspection speed, and inspection direction. If the obstacle avoidance method is non-detour obstacle avoidance, pause control is performed based on the safe trigger distance.

4. The method as described in claim 3, characterized in that, If the obstacle avoidance method is non-detour obstacle avoidance, the pause control is performed based on the safe trigger distance, and the method includes: If the obstacle avoidance method is non-detour obstacle avoidance, generate obstacle warning information and determine whether the safe trigger distance is met; If the safe trigger distance is not met, an emergency stop control is executed on the inspection robot; If the safe trigger distance is met, calculate the deceleration and perform deceleration and stop control on the inspection robot.

5. The method as described in claim 1, characterized in that the method include: A charging threshold is set, wherein the charging threshold is determined based on the real-time distance between the inspection robot and the workstation, and is dynamically adjusted according to the real-time positioning of the inspection robot; If the real-time battery level of the inspection robot is less than or equal to the charging threshold, an inspection interruption command is generated. Based on the patrol interruption command, the patrol robot is controlled to perform automatic repatriation, wherein the patrol interruption position serves as the starting point for the subsequent patrol.

6. The method as described in claim 1, characterized in that the method include: Based on the normal operating conditions of the inspection robot, a self-inspection standard list is generated; The self-inspection standard list is embedded into the central control module of the inspection robot; Based on a predetermined time period, the inspection robot periodically performs operational self-checks and generates a self-check dataset. The operational self-checks include self-checks of backup equipment and overall machine operation self-checks. The self-inspection standard list and the self-inspection dataset are mapped and verified to generate abnormal warning information; The inspection robot will issue a self-warning based on the abnormal warning information, wherein the warning execution methods are different for different components and different warning levels.

7. A motion intelligent control system for an inspection robot, characterized in that, The system is used to perform the motion intelligent control method for the inspection robot according to any one of claims 1 to 6, the system comprising: A structure generation module is used to collect basic information about the target power plant and generate a three-dimensional topology structure. The inspection planning module is used to build an inspection planning model, including: building a multi-level network layer, including a route planning layer, an information adjustment layer, and a mode configuration layer; determining the hierarchical execution logic for network layer training; connecting and associating the route planning layer, the information adjustment layer, and the mode configuration layer hierarchically to generate the inspection planning model; further calling historical inspection records, including sample inspection requirements and sample inspection execution information; inputting the sample inspection requirements into the inspection planning model for model verification; mapping and verifying the model verification results with the inspection execution information to check the model's analytical accuracy; if the analytical accuracy of the built inspection planning model is not up to standard, selecting the best training samples based on the historical inspection records, training the inspection planning model until convergence, obtaining the completed inspection planning model; inputting inspection requirement information into the inspection planning model; and outputting inspection execution information. The three-dimensional topology structure is embedded in the inspection planning model, and the inspection execution information includes the inspection cycle and inspection route, with the inspection route carrying an inspection mode identifier. The primary identification and early warning module is used to collect real-time inspection data and perform primary identification and early warning. The primary identification and early warning is based on the directional early warning criteria set by the programming program. The in-depth analysis and early warning module is used to perform data storage on the real-time inspection data and conduct in-depth analysis and early warning of the data. The patrol point determination module is used to generate operation and maintenance execution tasks and determine secondary patrol points based on the primary identification warning and the in-depth analysis warning, in conjunction with the on-site and back-end systems. The re-inspection control module is used to execute the operation and maintenance tasks. Based on the secondary inspection points, it plans the re-inspection path with the shortest route as the response target and performs re-inspection control of the inspection robot.

Images