Robot inspection method and system for comprehensive pipe gallery

Abstract

The application discloses a comprehensive pipe gallery robot inspection method and system, and the inspection area is divided into various inspection equipment areas according to the position relationship between the pipe gallery and the equipment, the combination of the track type robot, the foot type robot and the fixed point camera solves the problem that the existing single type of inspection robot cannot simultaneously adapt to the complex and rugged ground environment and the narrow and limited channel space in the comprehensive pipe gallery, leading to the existence of an inspection coverage blind area and the inability to realize full-area dead-angle-free monitoring, and the inspection range can be planned according to the user-selected inspection task and the various inspection equipment areas, so that the user demand and environmental factors can be matched; in addition, an inspection analysis report including environmental information and equipment information is generated according to the inspection data to support fault tracing and trend prediction.

Description

Technical Field

[0001] This invention relates to the field of robot inspection, and more particularly to a method and system for robot inspection of integrated utility tunnels. Background Technology

[0002] A nuclear power plant's integrated pipe gallery (GB gallery) is a site for laying process piping and cables to the plant's buildings and structures. Its total length can reach several kilometers (including cable and pipe sections). It houses cables and approximately 25 process system pipes, connecting about 20 BOP (Build-Operate-Place) sub-items. Having been in operation for over ten years, the related equipment is relatively old. Multiple instances of pipe ruptures and system valve leaks have occurred on-site. These problems are currently being detected through infrequent personnel inspections or by observing system pressure changes, making timely and rapid defect location and response impossible. As the unit's operating time increases and the pipes age, these problems will become particularly prominent, impacting the unit's operational safety.

[0003] The related technologies have the following technical problems: 1) Incomplete inspection coverage with blind spots: Existing single-type inspection robots cannot simultaneously adapt to the complex and rugged ground environment and narrow and restricted passage space in the integrated pipe gallery of a nuclear power plant, resulting in blind spots in inspection coverage and making it impossible to achieve full-area monitoring without blind spots. 2) It is difficult to balance inspection efficiency and accuracy: Existing inspection solutions (whether manual or single robot) lack intelligent task scheduling and hierarchical inspection capabilities, resulting in a "one-size-fits-all" inspection process that cannot dynamically adjust the level of inspection precision, leading to insufficient inspection density in key areas and wasted resources in non-critical areas. 3) Low data value and difficulty in in-depth utilization: In existing technologies, inspection data from different sources and of different types are often stored and analyzed in isolation. There is a lack of effective multi-source heterogeneous data fusion mechanisms, which makes it impossible to form a unified and comprehensive view of equipment health status. The data value has not been fully explored, making it difficult to support in-depth fault tracing and trend prediction. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to address at least one defect of the related technologies mentioned in the background: the existing single-type inspection robots cannot simultaneously adapt to the complex and rugged ground environment and narrow and restricted passage space in the integrated pipe gallery of a nuclear power plant, resulting in blind spots in the inspection and the inability to achieve full-area monitoring without blind spots. The present invention provides an integrated pipe gallery robot inspection method and system.

[0005] The technical solution adopted by this invention to solve its technical problem is: to construct a comprehensive utility tunnel robot inspection method, including the following steps: S10: Call the pre-built 3D cloud point map of the utility tunnel and the equipment ledger to obtain the positional relationship between the utility tunnel and the equipment. Divide the inspection area according to the positional relationship between the utility tunnel and the equipment. The inspection area includes the track-type robot inspection area, the legged robot inspection area and the fixed-point camera inspection area. S20: Based on the inspection task selected by the user, and in combination with the inspection areas of the track-mounted robot, the legged robot, and the fixed-point camera, the inspection range is planned for the track-mounted robot, the legged robot, and the fixed-point camera respectively. S30: Drives several track-mounted robots, legged robots, and fixed-point cameras to perform inspection tasks within the inspection range and collect inspection data in real time; S40: Generates an inspection analysis report based on inspection data, including environmental and equipment information.

[0006] Optionally, the track-mounted robot inspection area is the upper space of the corridor without any equipment obstruction or wall partitions; the legged robot inspection area is the lower space of the corridor with a wide space and a gentle slope; and the fixed-point camera inspection area is the blind spot area that the track-mounted robot and the legged robot cannot reach. The inspection content of the track-mounted robot inspection area, the legged robot inspection area, and the fixed-point camera inspection area overlaps.

[0007] Optionally, the track-mounted robot inspection area and the legged robot inspection area have several fireproof partition doors to meet fireproof partition requirements; When the track-mounted robot or legged robot moves to a preset distance in front of the fireproof partition door, it controls the fireproof partition door to execute an opening command. After the track-mounted robot or legged robot has safely passed through, it controls the fireproof partition door to execute a closing command.

[0008] Optionally, the method after S10 further includes: S11: Semantic processing is performed on the 3D cloud point map of the utility tunnel to divide the tunnel space into grid management units. Each grid management unit is assigned a risk label with an initial weight based on the equipment ledger. The inspection area is divided into at least two levels based on the risk label, including a first-level key area and a second-level routine area. Different inspection strategies are developed for track-mounted robots, legged robots and fixed-point cameras for different levels of inspection areas.

[0009] Optionally, the weight of the risk label is linked to historical inspection data and dynamically adjusted based on the real-time collected inspection data.

[0010] Optionally, during the inspection task, the battery charge status and current inspection task complexity of each track-mounted robot and each legged robot are monitored and evaluated in real time. The remaining battery life is predicted by combining the battery charge status and the current inspection task complexity to control whether to trigger the return mode.

[0011] Optionally, when any tracked or legged robot triggers a return-to-home charging, if there are other idle and fully charged backup robots within a preset range, the coordinates of the unfinished task points and the detection requirements are sent to the backup robots, driving them to continue executing the current task.

[0012] Optionally, the inspection data is preprocessed by edge computing on the robot during real-time acquisition; the inspection data is used to align and fuse with the 3D cloud point map of the utility tunnel to generate a real-time 3D view of the utility tunnel status.

[0013] Optionally, the environmental information includes at least one of the following: leak detection information, environmental anomaly information, and environmental parameter information; Equipment information includes at least one of the following: equipment readings, equipment status, equipment appearance inspection information, and equipment thermal status information.

[0014] The present invention also constructs an integrated utility tunnel robot inspection system, including: a robot control platform, a track-mounted robot, a legged robot, and a fixed-point camera; the robot control platform is connected to the track-mounted robot, the legged robot, and the fixed-point camera through at least one data node; The robot control platform is used to call the pre-built 3D cloud point map of the utility tunnel and the equipment ledger to obtain the positional relationship between the utility tunnel and the equipment. Based on the positional relationship, the inspection area is divided into the track-type robot inspection area, the legged robot inspection area and the fixed-point camera inspection area. Based on the inspection task selected by the user, the inspection areas of the track-mounted robot, the legged robot, and the fixed-point camera are combined to plan the inspection range for the track-mounted robot, the legged robot, and the fixed-point camera respectively. Drive the track-mounted robot, legged robot, and fixed-point camera to perform inspection tasks within the inspection range and collect inspection data in real time, and generate inspection analysis reports based on the inspection data; Track-mounted robots, legged robots, and fixed-point cameras are used to perform inspection tasks within the inspection area and collect and upload inspection data in real time.

[0015] By implementing this invention, the following beneficial effects are achieved: This invention discloses a method and system for robotic inspection of integrated utility tunnels. The invention divides the inspection area into various inspection equipment zones based on the positional relationship between the tunnel and the equipment. The combination of a track-mounted robot, a legged robot, and a fixed-point camera solves the problem that existing single-type inspection robots cannot simultaneously adapt to the complex and rugged ground environment and narrow, restricted passageways within integrated utility tunnels, resulting in blind spots and the inability to achieve comprehensive, blind-spot-free monitoring. Furthermore, the invention can plan the inspection range based on the user-selected inspection task and the various inspection equipment zones, matching user needs and environmental factors. In addition, it generates an inspection analysis report based on the inspection data, including environmental and equipment information, to support fault tracing and trend prediction. Attached Figure Description

[0016] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings: Figure 1 This invention illustrates an overall flowchart of an embodiment of the integrated utility tunnel robot inspection method. Figure 2 A track deployment plan view of an embodiment of the integrated utility tunnel robot inspection method of the present invention is shown; Figure 3 This diagram shows an overall structural diagram of an embodiment of the integrated utility tunnel robot inspection system of the present invention; Figure 4 A detailed structural diagram of an embodiment of the integrated utility tunnel robot inspection system of the present invention is shown. Detailed Implementation

[0017] To provide a clearer understanding of the technical features, objectives, and effects of the present invention, specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0018] It should be noted that the flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0019] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0020] It should be noted that "at least two" refers to at least two, which can be two, three, or any number. "At least one" can be one, two, or any number.

[0021] Currently, a large amount of work, such as equipment inspection, defect tracking, and data entry in utility tunnels, still relies on manual labor, which is time-consuming and labor-intensive. Furthermore, tasks such as temperature measurement and instrument data analysis have long time intervals, hindering real-time monitoring of the operational status of critical equipment. Because many important system cables and pipes are installed within the tunnels of integrated utility tunnels, from a safety perspective, it is necessary to strengthen inspections in this area. This allows for timely detection of problems, and through big data recording and analysis, early warnings can be issued to ensure the safe and stable operation of the system and equipment.

[0022] like Figure 1 As shown, some embodiments of the present invention disclose a method for robot inspection of integrated utility tunnels, including the following steps: S10: Call the pre-built 3D cloud point map of the utility tunnel and the equipment ledger to obtain the positional relationship between the utility tunnel and the equipment. Divide the inspection area according to the positional relationship between the utility tunnel and the equipment. The inspection area includes the track-mounted robot inspection area, the legged robot inspection area and the fixed-point camera inspection area.

[0023] Furthermore, the track-mounted robot inspection area is the upper space of the corridor without any equipment obstructions or wall partitions; the legged robot inspection area is the lower space of the corridor with a wide and gently sloping area; and the fixed-point camera inspection area is the blind spot area that the track-mounted robot and legged robot cannot reach. Optionally, the inspection content of the track-mounted robot inspection area, the legged robot inspection area, and the fixed-point camera inspection area may overlap.

[0024] Optionally, if there are a few movable devices obstructing the space above the corridor, they can be modified to meet the requirements of the track-mounted robot inspection area. If there are a few non-movable devices obstructing the space above the corridor, non-movable devices can be avoided when setting up the track to meet the requirements of the track-mounted robot inspection area.

[0025] Specifically, for example, some equipment on site (cameras, loudspeakers, brackets supporting cable trays, rusted scrap angle iron on the wall, signal antennas, protruding parts of supports under cable trays, etc.) interfere with the movement of the track-mounted robot and needs to be addressed: At the turning points of the track-mounted robot, the top pipe is lower than the adjacent cable tray, reducing the overhead space and making it difficult for the robot to pass over the cable tray. During construction, the position of this pipe should be moved upwards to be level with the cable tray. There is also a cable tray bracket on site that obstructs the track installation; a horizontal curved rail needs to be deployed to bypass this bracket.

[0026] Optionally, after on-site surveying, curved rails can be installed at locations where turning is possible in the inspection area for the track-mounted robot, enabling the robot to inspect the other side of equipment or walls. Turning is possible when the overhead space of the corridor is greater than a preset height and width for the track-mounted robot to turn. Furthermore, in addition to the rail mounting components installed at normal intervals, at least one additional rail mounting component is added at each horizontal curved rail for reinforcement and stability.

[0027] Specifically, a specific implementation example of a track deployment plan, such as Figure 2 As shown, a section of corridor is divided into inner and outer sides by a central cable trough. The outer side is more spacious with fewer obstacles, but it has a partial interruption, preventing it from running continuously. The inner side is narrower and has obstacles at the top, but it runs continuously. After on-site surveying, the space at the top of the inner side is generally suitable for deploying robot tracks, so the tracks will be deployed on the inner side of this corridor. Considering the need for external inspection as well, curved tracks will be installed at points with turning capabilities (at least 700mm from the top of the cable trough) to allow the robot to turn into the inner side, enabling inspection on both sides. Figure 2 It is known that there are no inspection devices outside the 8G section. Based on the specific location relationship between the pipe gallery and the equipment outside the 8G section, it can be set as a track-mounted robot inspection area, a legged robot inspection area, or a fixed-point camera inspection area.

[0028] Based on on-site measurements, section 8N of the corridor has turning capabilities, while section 8G, with the cable trough top 480mm from the ceiling, does not. Therefore, curved rails are only installed at the end of section 8N to allow the robot to turn outwards, enabling dual-sided inspection of section 8N. This section of track is approximately 200 meters long. At the end of section 8N, the distance from the top of the cable trough to the ceiling is relatively large, basically meeting the space requirements for the robot to pass over the cable trough. Two 90-degree curved rails are installed here to turn the track path from the inside to the outside of the corridor, enabling dual-sided inspection of this section. In addition to the rail hangers installed at normal intervals, an additional hanger is added at each horizontal curved rail for reinforcement and stability. It should be noted that the above specific data is for illustrative purposes only and can be adjusted according to actual conditions.

[0029] Optionally, a positioning device is installed at certain preset intervals along the track on which the track-mounted robot is located. Whenever the track-mounted robot passes a positioning device, its reading device reads the current position for positioning.

[0030] Specifically, the track-mounted robot's track is fixed to the top via a hanger. The track has a load-bearing capacity of 100 kg, and the hanger is 20 cm long, meaning the track is 20 cm from the top. The robot track is installed at the bottom of the uppermost support. An RFID positioning tag is installed every 20 meters along the track. Whenever the robot passes a tag, a reader installed on its body reads the tag's code, thus helping the robot locate itself. It should be noted that the above specific data is for illustrative purposes only and can be configured differently according to actual needs.

[0031] Furthermore, when setting the track for the tracked robot, a predetermined safe distance should be maintained between it and the pipes, equipment, and walls already installed above the utility tunnel. The robot should avoid or slightly detour around the pipes and equipment already installed above the utility tunnel to prevent mutual interference.

[0032] Optionally, the inspection areas of the track-mounted robot, the legged robot, and the fixed-point camera may overlap. This allows for collaborative inspection by another robot / fixed-point camera when a robot breaks down during subsequent inspection tasks.

[0033] Optionally, the track-mounted robot inspection area and the legged robot inspection area have several fireproof partition doors to meet fireproof partition requirements; When the track-mounted robot or legged robot moves to a preset distance in front of the fireproof partition door, it controls the fireproof partition door to execute an opening command. After the track-mounted robot or legged robot has safely passed through, it controls the fireproof partition door to execute a closing command.

[0034] Furthermore, one or two fireproof partition doors can be installed at a uniform location in the utility tunnel (i.e., there may be fireproof partition doors at the top or only at the bottom). Below each fireproof partition door is a fireproof partition wall or a similar structure. The fireproof partition doors open automatically when tracked or legged robots pass through, and can also be opened manually when necessary (even during a power outage). The door opener is controlled by AC220V power supply, has an IP54 protection rating, and operates in ambient temperatures ranging from -50°C to 50°C.

[0035] Specifically, the fireproof partition door not only serves as a fireproof partition for the corridor, but also requires sufficient space above it for installing an electrically controlled door to allow for the passage of a track-mounted robot. For example, based on on-site measurements, it is recommended to leave a 630mm long space later. The doorway is 600mm high. To meet on-site protection requirements, the modified door will be made of fire-resistant steel. It will be a double-leaf door, driven by a motor. Door-opening and closing sensor tags will be installed on tracks at a certain distance on both sides of the door, linked to the door's movement. When a robot passes a sensor tag, the tag sends a passage signal to the door control mechanism, which then opens the door. After the robot has safely passed, the sensor tag signal is disconnected, and the mechanism closes the door. It should be noted that the above specific data is for illustrative purposes only and can be adjusted according to actual conditions.

[0036] A fire door control box is installed on the ground next to each robotic electric fire door, providing AC power and control circuit for the fire door and sensor tags. The control box is equipped with local control buttons, allowing staff to manually control the opening and closing of the fire door locally.

[0037] Optionally, the method after S10 further includes: S11: Semantic processing is performed on the 3D cloud point map of the utility tunnel to divide the tunnel space into grid management units. Each grid management unit is assigned a risk label with an initial weight based on the equipment ledger. The inspection area is divided into at least two levels based on the risk label, including a first-level key area and a second-level routine area. Different inspection strategies are developed for track-mounted robots, legged robots and fixed-point cameras for different levels of inspection areas.

[0038] Specifically, before the inspection task begins, a static baseline based on a semantic map is constructed. The robot control platform performs semantic processing on the 3D cloud point map of the utility tunnel to divide the tunnel space into grid-based management units. The equipment ledger assigns a risk label with initial weight to each grid-based management unit. For example, areas containing main steam pipes, high-pressure cable joints, or old valves are marked as "Level 1 Key Areas"; ordinary corridor connection sections are marked as "Level 2 Regular Areas." Based on these labels, different inspection strategies are developed for track-mounted robots, legged robots, and fixed-point cameras for different levels of inspection areas. Preset passage speeds and basic sampling frequencies are set for different areas, forming a static baseline layer for hierarchical inspection.

[0039] Optionally, the weight of the risk label is linked to historical inspection data and dynamically adjusted based on real-time collected inspection data. Furthermore, the weight of the risk label is also dynamically adjusted based on several inspection instructions issued by the user.

[0040] Specifically, during the inspection task, the multi-source driven dynamic weight update system receives data from each sensor in real time through edge computing nodes and dynamically adjusts the area weights. The specific process includes: historical data association: the system retrieves historical data from the past N inspections to analyze the degradation trend of specific equipment. If an instrument reading shows a slow drift, or the equipment temperature shows an upward trend, even if it has not reached the alarm threshold, the system will automatically increase the "attention weight" of that area.

[0041] Each inspection device (track-mounted robot, legged robot, and fixed-point camera) has environmental perception and feedback capabilities. When the front-end sensors of the inspection device (such as gas sensors and obstacle avoidance radar) detect environmental anomalies (such as trace amounts of smoke, water accumulation, or abnormal vibration), the background immediately triggers the "dynamic upgrade strategy" to instantly upgrade the area to the highest priority "special grade inspection zone".

[0042] Each inspection device also has the ability to interrupt task instructions. If the control center receives a specific manual inspection instruction (such as a valve that needs to be checked), the system will set the coordinate point and its surrounding radius as a temporary core area.

[0043] Furthermore, different inspection strategies include: performing adaptive adjustments for different levels of inspection areas from the dimensions of movement, vision, and data processing.

[0044] Specifically, in the adaptive execution of heterogeneous robots, after receiving control commands based on updated weights, the motion control module and the perception and acquisition module synchronously perform adaptive adjustments. The specific implementation details are as follows: In the movement dimension, the system reconstructs the robot's kinematic parameters. For track-based robots, before entering key areas, the system smoothly reduces the drive motor speed via a PID controller, decreasing the movement speed from the usual 1.5 m / s to 0.2 m / s. Simultaneously, the gimbal attitude is adjusted, locking its pitch angle to vertically downward or a specific angle to eliminate image motion blur caused by high-speed movement. For legged robots, the system adjusts the gait planning algorithm, reducing stride length and increasing support phase time to ensure absolute stability of the robot body in stationary or slightly moving states, providing a physical benchmark for high-precision measurements. In the visual dimension, the perception and acquisition parameters are significantly improved. The visible light camera's exposure strategy is switched from "automatic global metering" to "fixed-point center metering," and a super-resolution mode is activated. The infrared thermal imager's temperature measurement range is automatically narrowed to the target device's expected temperature range to improve thermal sensitivity (e.g., from ±2℃ to ±0.5℃), and a continuous frame overlay filtering algorithm is activated to remove noise. In terms of data processing, the AI ​​inference model on the robot's edge has switched from a lightweight detection network to a high-precision segmentation network to perform pixel-level semantic segmentation of the target device, ensuring the ability to identify minute surface cracks, oil leaks, or slight deflections of instrument pointers. It should be noted that the specific data mentioned above is for illustrative purposes only and can be set to other values ​​depending on the actual situation.

[0045] This hierarchical inspection strategy, employing an adaptive resource allocation mechanism with dynamic weights, breaks away from the traditional "homogeneous" inspection work pattern, establishing a multi-level dynamic response mechanism based on equipment health assessment and regional risk rating. This mechanism utilizes a robot control platform as a central brain to calculate the "inspection weight" of each area within the utility tunnel in real time. Based on changes in these weights, it dynamically schedules the kinematic parameters of heterogeneous robots (legged robots and tracked robots) and the parameters collected by the sensing system. From a control theory perspective, this is a typical closed-loop feedback control system. The system's input consists of a static risk map of the entire utility tunnel and real-time status data. After processing by a fusion algorithm, it outputs differentiated control commands.

[0046] This invention achieves full coverage of the integrated utility tunnel in nuclear power plants while concentrating limited computing and energy resources in high-risk and high-value areas, greatly improving the timeliness and accuracy of hazard detection. This non-uniform, data-driven inspection strategy significantly enhances the system's intelligence level and is a key technological path to achieving unmanned operation and maintenance of nuclear power plants.

[0047] S20: Based on the inspection task selected by the user, and combining the inspection areas of the track-mounted robot, the legged robot, and the fixed-point camera, the inspection ranges are planned for the track-mounted robot, the legged robot, and the fixed-point camera respectively. For example, based on the inspection task of a certain area selected by the user, the inspection area to which that area belongs can be analyzed, and the inspection ranges of the track-mounted robot, the legged robot, and the fixed-point camera can be planned in a nearby and efficient manner.

[0048] S30: Drives several track-mounted robots, legged robots, and fixed-point cameras to perform inspection tasks within the inspection area and collect inspection data in real time. Optionally, the inspection functions include automatic inspection, manual remote-controlled inspection, and inspection classification.

[0049] Specifically, the automatic inspection function enables it to autonomously complete inspection tasks, exhibiting a high degree of automation and intelligence, thus replacing manual inspection. The automatic inspection modes mainly include: timed inspection, routine inspection, special inspection, single inspection, and manually controlled inspection, with various modes supporting switching between each other. The inspection equipment can complete tasks according to pre-set inspection modes, performing inspection activities such as meter readings, power equipment temperature measurement, and environmental monitoring according to established rules.

[0050] In addition to automated inspection, the inspection equipment can also be remotely controlled in real time via manual remote control. This application mode is suitable for maintenance personnel and management units that need to lock and monitor the status of certain types of equipment. Especially when the robot detects abnormal equipment or environmental conditions during autonomous inspection and alarms the maintenance personnel, the maintenance personnel can immediately use the manual remote control interface or remote control handle in the management backend to control the robot to quickly reach the location of the abnormal equipment, promptly check the abnormal equipment and verify the alarm information, so as to quickly formulate a response strategy.

[0051] When inspecting key areas, the robot operates at its minimum speed via backend control, allowing various instruments to collect detailed data. For non-key areas, it can operate at maximum speed. This targeted approach to inspection improves efficiency and emergency management capabilities.

[0052] In one specific embodiment, when the robot performs inspection tasks within its designated inspection area, it utilizes LiDAR SLAM and visual odometry for real-time localization. Upon encountering obstacles (such as temporary stockpiles), it automatically detours or stops using a local path planning algorithm. The robot performs tiered adaptive inspections. For regular areas, it moves at a standard speed, performing a macroscopic "scan" to collect basic environmental data. For key areas, when the robot enters a pre-defined "key area" or when an anomaly is detected in real-time by the backend, it automatically switches modes, reducing speed, adjusting the gimbal angle, and increasing the sensor sampling rate for microscopic "staring" and refined inspection. Real-time acquisition of multi-source inspection data includes visible light video, infrared thermal imaging, point cloud data, and environmental gas concentration data; simultaneously, a fixed-point infrared camera continuously transmits temperature flow data from critical equipment.

[0053] Optionally, after receiving instructions, the track-mounted or legged robot performs a local self-check (battery level, motor status, sensor communication, and storage space). Once confirmed to be correct, it sends self-check information (such as "Ready") back to the backend and detaches from the charging station. As the main body of the inspection system, the robot carries the primary detection equipment to collect various data at the front end. To ensure the robot's daily operation, it performs a self-check before starting any inspection. This self-check includes the infrared thermal imager, high-definition camera, motors, gimbal, internal storage, and various sensors. If any component malfunctions are detected, an abnormal status indication is given, and the system malfunction information is uploaded to the robot control platform. This facilitates timely fault detection by maintenance personnel, reducing processing time and improving troubleshooting efficiency.

[0054] Specifically, during the inspection process, the robot automatically issues early warnings for abnormal data using technologies such as numerical analysis, threshold comparison, trend analysis, and database analysis. Warning information is presented through both interface alarms and audible / visual alarms to promptly alert maintenance personnel. When backend analysis data shows anomalies, a list of alarm information appears on the backend monitoring software interface, which maintenance personnel can scroll through by interacting with the system. When system alarms are present, the system alerts maintenance personnel through audio and flashing lights from the inspection robot.

[0055] Optionally, during the inspection task, the battery charge status and current inspection task complexity of each track-mounted robot and each legged robot are monitored and evaluated in real time. The remaining battery life is predicted by combining the battery charge status and the current inspection task complexity to control whether to trigger the return mode.

[0056] Specifically, during the inspection process, the battery management system (BMS) monitors the battery pack voltage, current, temperature, and internal resistance in real time. Combining the ampere-hour integration method and the open-circuit voltage method, the system accurately calculates the battery's state of charge (SOC). Dynamic power consumption modeling is performed on the robot's internal battery. The system establishes a dynamic power consumption model based on the robot's current load conditions (such as whether it is climbing a slope, whether high-power sensors are activated, and its movement speed). It predicts the remaining runtime (RTE) not only based on the current battery level but also on the complexity of the current task path, preventing unexpected power outages due to sudden high-energy-consuming actions.

[0057] Furthermore, the remaining driving time is predicted by combining the battery charge status and the complexity of the current inspection task to control whether to trigger the return mode. This includes setting multi-level power thresholds and triggering different levels of response mechanisms based on the remaining power to achieve a graded alarm and response strategy.

[0058] Specifically, a three-level warning system can be configured. The first-level alarm (low battery warning) is triggered when the State of Charge (SOC) is less than 30% (adjustable based on operating conditions). The response strategy is for the backend collaborative control platform to send a low battery alert to the operator. The robot automatically assesses the current task progress, prioritizes completing the inspection of key nodes, postpones inspections of non-critical areas, and plans the optimal path to the nearest charging station. The second-level alarm (return command) is triggered when the SOC is less than 20%. The response strategy is for the system to forcibly interrupt the current routine inspection task and enter "autonomous return mode." The robot automatically shuts down unnecessary loads (such as some lights and non-critical sensors), reduces its operating speed, and returns to the charging station along the planned path in the lowest power consumption mode. The third-level alarm (emergency protection) is triggered when the SOC is less than 5% or the voltage suddenly drops to the cutoff voltage. The response strategy is for the robot to immediately stop all movement and enter "sleep protection state." The system cuts off the motor drive power, retaining only the communication module and control core in low-power standby mode, and sends a final distress signal containing precise coordinates to the backend, awaiting manual assistance.

[0059] During the return journey, SLAM technology is used for real-time localization and navigation. If the original path is blocked, the system replans a backup path with the shortest distance and lowest energy consumption based on the global map. For track-mounted robots, they can accurately and autonomously dock, using RFID tags or visual markers on the track to identify the location of the charging station, control the motor to decelerate, and precisely stop at the charging contact point, completing the physical connection using the pantograph. For legged robots, visual recognition of the QR code or AprilTag marker on the charging station is used, combined with odometer calibration, to adjust the robot's posture so that the robot's electrodes are precisely aligned with the charging station's electrodes. During charging, the BMS monitors the charging current and voltage, employing a constant current-constant voltage charging strategy to prevent overcharging. Once fully charged, the system automatically cuts off power and wakes the robot, reporting its robot status (e.g., "readiness") to the system.

[0060] In some embodiments, the inspection content of the track-mounted robot, the legged robot, and the fixed-point camera overlaps. Optionally, when any track-mounted or legged robot triggers a return-to-home charging, if there are other idle and fully charged backup robots within a preset range, the coordinates of the unfinished task points and the inspection requirements are sent to the backup robots, driving them to continue executing the current task. Preferably, based on the importance classification of the area, at least one backup robot is set up in the primary key area to ensure the real-time performance of the current task. If the type of backup robot meeting the conditions within the preset range includes the type of the original task, the nearest backup robot of the same type is selected to continue executing the current task. Furthermore, the track-mounted and legged robots have breakpoint data protection and breakpoint resume mechanisms.

[0061] Specifically, when a robot is forced to return due to insufficient power, the robot control platform automatically assesses the currently unfinished inspection tasks. If there are other idle robots with sufficient power within the preset range (e.g., if a legged robot returns, it will be replaced by a legged / rail-based robot within the preset range that meets the conditions, with priority given to the legged robot; or if a rail-based robot returns, it will be replaced by a rail-based / legged robot within the preset range that meets the conditions, with priority given to the rail-based robot), the system will send the coordinates of the unfinished task points and inspection requirements to the backup robot to achieve collaborative operation. If the currently executed task has been completed by the backup robot, the returning robot can perform other tasks according to the instructions sent by the staff; if the originally executed task has not yet been completed, the returning robot will hand over to the backup robot after charging and begin performing the task.

[0062] For breakpoint data protection, before the battery is about to run out, the robot immediately writes the currently collected inspection data, log files, and pose status to non-volatile memory (such as local SSD or Flash) to prevent data loss due to sudden power outages. For breakpoint resume inspection, once the robot has finished charging and reconnected to the network, the system supports the "breakpoint resume inspection" function, allowing the robot to automatically resume inspection from the last interrupted task point, avoiding repetitive work.

[0063] In some embodiments, the legged robot optionally features an emergency physical safety mechanical braking and locking mechanism. Specifically, once the battery is detected to be completely depleted, the legged robot immediately activates the electronic brakes or mechanical locking mechanism (particularly for the joint motors of the legged robot and the wheel motors of the overhead rail robot) to prevent the robot from slipping or falling due to loss of power on the utility tunnel ramp. Easily detectable design: The robot's shell is equipped with high-brightness flashing LEDs or an audible and visual alarm that operates intermittently in low-battery standby mode, allowing maintenance personnel to quickly locate and retrieve the robot in dimly lit utility tunnel environments.

[0064] Through the above comprehensive management measures, the power shortage during robot operation can be effectively mitigated, ensuring the continuity of nuclear power plant inspection operations, data security, and the physical safety of the equipment itself.

[0065] Optionally, the inspection data is preprocessed by edge computing on the robot during real-time acquisition; the inspection data is used to align and fuse with the 3D cloud point map of the utility tunnel to generate a real-time 3D view of the utility tunnel status.

[0066] Specifically, the AI ​​chip on the robot's edge performs preliminary processing on the raw data (such as target detection and OCR reading), removes invalid data, and extracts feature vectors to obtain inspection data. All data (from the track-mounted robot, legged robot, and fixed-point camera) is transmitted in real time to the data platform of the robot control system. The system uses a unified spatiotemporal reference to align and fuse data from different perspectives and times in a digital twin model, generating a real-time 3D view of the utility tunnel status.

[0067] Optionally, if the fusion analysis detects an anomaly in the detection environment or equipment (such as pipeline leaks or excessively high temperatures), an alarm (such as audible and visual alarms or SMS alarms) is triggered, and the location of the anomaly is displayed. Furthermore, if a situation is found that a particular robot cannot handle, another robot near the anomaly location can be dispatched to conduct a collaborative verification.

[0068] Specifically, leak detection utilizes a dual-view gimbal mounted on an inspection robot to acquire real-time video images of the tunnel interior. A high-precision leak detection algorithm intelligently analyzes the video data to identify its content. For specific environments, existing deep learning object detection methods can be used to detect designated objects. By leveraging surveillance footage from specific inspection points, leaks within the tunnel can be detected immediately.

[0069] The inspection equipment, equipped with visible light sensors, collects images of the monitored area. Online annotation tools are used to label these images, selecting areas for identification. This process trains an AI model to build a database, which is then integrated into the intelligent inspection system. Once the AI ​​model is established, the inspection robot compares the collected images with those in the AI ​​database. If an anomaly is detected at a marked location, the robot annotates the image and issues an alarm, allowing staff to address the issue promptly. Simultaneously, a deep learning-based target detection scheme detects anomalies within the inspection area. Alarm thresholds are set based on the pixel area of ​​the target to meet specific on-site needs.

[0070] Specifically, infrared thermal imaging uses the robot's infrared camera to perform a comprehensive scanning temperature acquisition of equipment within the corridor, effectively preventing the problem of equipment being missed during personnel inspections. Equipped with a temperature-measuring thermal imaging module, it can automatically detect the temperature of objects inside the tunnel and supports setting different temperature over-temperature alarm thresholds for different sensitive targets. When abnormal temperatures are detected in facilities, equipment, or other objects inside the tunnel, the thermal imaging camera and high-definition pan-tilt unit will automatically take pictures of the target as evidence and mark the current tunnel location information. At the same time, the background system will trigger an over-temperature alarm and notify the on-duty personnel for confirmation.

[0071] The robot uses its onboard infrared thermal imager to collect temperature data from electrical equipment, cables, and other objects. It then analyzes the data to diagnose heat-causing equipment faults and thermal defects, triggering corresponding alarms. Infrared temperature measurement functions include general infrared measurement, precise temperature measurement, remote temperature measurement, and temperature warning. Upon completion of the inspection task, a task report is automatically generated, providing effective diagnosis of equipment temperature. When the inspection robot moves near equipment, it can detect overheating caused by aging, malfunctions, etc., and can also detect high temperatures early, preventing damage to equipment from excessive heat and providing timely warnings to reduce or avoid equipment accidents. Fixed-point cameras, deployed in locations difficult for the robot to reach, also possess the same infrared thermal imaging capabilities.

[0072] Due to the narrow space and long distance of the corridor, the traditional manual inspection mode is risky and the inspection frequency is insufficient. Intelligent inspection equipment, equipped with infrared thermal imagers, optical cameras, gas detectors and other devices, can effectively collect images, temperature environment and internal dripping and seepage information in the corridor. The system monitoring content requirements are shown in Table 1 below.

[0073] Table 1 System Monitoring Function Table Optionally, after the task is completed or the battery level is below a threshold, the track-mounted robot and the legged robot autonomously plan the shortest path back to the charging station to complete precise docking and charging.

[0074] S40: Generates inspection analysis reports including environmental and equipment information based on inspection data. Each inspection device automatically generates a report after completing a task. Maintenance personnel can search and view these reports on the robot control platform by time, inspection type, task name, etc. Maintenance personnel can search, view, and export various reports as needed, including inspection task reports, data reports, alarm record reports, and environmental information reports.

[0075] Optionally, the environmental information includes at least one of the following: leak detection information, environmental anomaly information, and environmental parameter information; the equipment information includes at least one of the following: equipment reading information, equipment status information, equipment appearance inspection information, and equipment thermal status information.

[0076] Furthermore, leak detection information includes gas leaks and liquid leaks; environmental anomaly information includes foreign object intrusion; and environmental parameter information includes air temperature and humidity, water level in ditches, and smoke and fire.

[0077] Specifically, for gas leaks, the system detects the concentration of harmful or flammable gases in the air using onboard gas sensors (such as CH4, H2, and O3 sensors). For liquid leaks, it visually identifies water accumulation and oil stains on the ground and under equipment. For foreign object intrusion, it detects unauthorized personnel entering the pipe gallery, abandoned tools, animal intrusion, or accumulation of construction debris. For air temperature and humidity, it monitors the air temperature and humidity within the pipe gallery to prevent excessive humidity from causing condensation or short circuits in equipment. For drainage ditch water levels, it monitors the water level in drainage ditches and the status of sump pits to prevent flooding. For smoke and fire, it uses smoke sensors and visual flame recognition algorithms to monitor fire hazards in real time.

[0078] Furthermore, for equipment readings, high-definition visible light cameras and OCR technology are used to read the values ​​of various pressure gauges, thermometers, level gauges, and flow meters, and compare them with background thresholds to determine whether they are within the normal range. For equipment status information, the open / closed status of valves and the position of handwheels are identified to confirm whether there are any erroneous operations or abnormal changes, and the indicator light colors (red / green / yellow) and switch open / closed status on the distribution cabinet and control box are checked. For equipment appearance inspection information, the pipe surfaces are checked for corrosion, rust spots, and paint peeling; the insulation layer is checked for damage; and the cable sheaths are checked for damage or aging.

[0079] Equipment thermal status information includes electrical joint status detection information, pipeline insulation failure detection information, and equipment overload detection information.

[0080] For electrical joint status detection, infrared detection is used to focus on the temperature of cable joints, busbar connections, transformer enclosures, and other key areas to identify potential overheating hazards caused by poor contact. For pipeline insulation failure detection, infrared thermal imagers are used to detect the surface temperature field of thermal pipelines, identifying heat leakage caused by insulation layer damage (manifested as localized high-temperature spots). For equipment overload detection, the bearing and casing temperatures of operating equipment such as motors and pumps are monitored to prevent overheating and burnout.

[0081] like Figure 3As shown, this invention also constructs a comprehensive utility tunnel robot inspection system, including: a robot control platform, a track-mounted robot, a legged robot, and a fixed-point camera; the robot control platform is connected to the track-mounted robot, the legged robot, and the fixed-point camera through at least one data node; further, the data node is configured with an application server, a switching device, and a storage device; the application server, switching device, and storage device are located in the IV zone server room (here, IV zone refers to an area of ​​the comprehensive utility tunnel, used only as an example), and the power supply for the devices is taken from the internal power supply of the installed cabinet (specified by the system maintenance department). The application server is used to deploy the robot control platform, adding several track-mounted robots, legged robots, and fixed-point cameras to the robot control platform (e.g., by specifying IP addresses), so as to realize the simultaneous operation of multiple inspection devices by switching the current controlled object on the same page; it can also connect to on-site card cameras and intelligent sensor devices supporting communication protocols such as Modbus / MQTT, to achieve comprehensive monitoring of the inspection scene. The switching device can serve as the center of the data node, used for data transmission and exchange between the track-mounted robot, the legged robot, the fixed-point camera, the application server, the storage device, and the robot control platform. Storage devices are used for storing and backing up various types of data.

[0082] The robot control platform is used to call a pre-built 3D cloud point map of the utility tunnel and equipment ledger to obtain the positional relationship between the utility tunnel and the equipment. Based on the positional relationship, the inspection area is divided into a track-mounted robot inspection area, a legged robot inspection area, and a fixed-point camera inspection area. According to the inspection task selected by the user, the inspection range of the track-mounted robot inspection area, the legged robot inspection area, and the fixed-point camera inspection area are planned for the track-mounted robot, the legged robot, and the fixed-point camera respectively. The platform drives the track-mounted robot, the legged robot, and the fixed-point camera to perform the inspection task according to the inspection range and collect inspection data in real time. The platform generates an inspection analysis report based on the inspection data.

[0083] Specifically, the robot control platform adopts a modular, layered design with a clear architecture and complete external interfaces. The drive layer is responsible for the robot's motion control, sensor access, and data acquisition; the business layer is responsible for robot status monitoring, task scheduling, image recognition, inspection data aggregation and analysis, alarm judgment, and other business logic processing; the application layer is an intelligent inspection centralized control platform, using a B / S architecture to achieve unified user interaction across platforms, while also providing API interfaces for data interaction with external systems. The functions of the control platform are shown in Table 2: Table 2. Function List of Robot Control Platform The application layer supports the integration of IoT devices such as inspection robots, fixed-point cameras, thermal imagers, and sensors, forming a comprehensive intelligent inspection management platform centered on inspection equipment. It provides various image and video algorithm models to achieve multi-target recognition and analysis of people, equipment, and the environment. The robot control platform is based on a B / S architecture web data browsing mode and operating platform, typically deployed on a central control room server. It supports multiple standard industrial protocol sets such as MODBUS, CAN, MQTT, HTTP, RTSP, and GB28181, enabling data interaction and linkage between external systems and the intelligent inspection system's centralized control platform.

[0084] Open your browser and access the corresponding IP address of the centralized control platform to enter the login interface of the intelligent inspection robot. Enter your account and password to log in. The platform supports multiple accounts to log in at the same time and can be managed by department, organization, user, and role. You can view the login time and account information of the relevant personnel at any time.

[0085] Specifically, the robot control platform is equipped with corresponding operation interfaces, including: home overview interface, robot inspection interface, video monitoring interface, data analysis interface, alarm center interface, equipment management interface, system user management interface, etc.

[0086] The homepage overview displays a global overview of the robot's path and location, robot status information, inspection task assignments, camera status, real-time visible light video, and real-time infrared temperature measurement video. It comprehensively summarizes and provides a global overview of the integrated utility tunnel robot inspection system, enabling unified management and control of the inspection equipment.

[0087] The robot inspection interface is where the robot's control system is imported into the centralized control platform for display. It enables remote control of the robot, customization of task templates, creation of inspection points, export of inspection reports, and querying of historical robot data, among other functional modules. The centralized control platform can switch between different robots and access their respective control systems to control, configure, and view the robot and its data.

[0088] The video surveillance interface is used for unified management and operation of fixed-point camera groups connected to the centralized control platform. On this interface, you can view real-time video or review recorded videos (including visible light monitoring and thermal imaging monitoring) and the switching list of each camera group.

[0089] The data analysis interface typically displays the data and analysis results collected by fixed-point intelligent monitoring equipment, such as cameras, infrared thermal imaging, and environmental monitoring. Users can view the data collected by each device and view the data analysis results (data collected by the robot is displayed in the robot inspection interface).

[0090] The alarm center interface mainly displays the data collected by the connected fixed-point intelligent monitoring equipment, analyzes the abnormal results, and queries alarm records. It supports multi-dimensional search queries based on alarm type, alarm status, and alarm time.

[0091] The equipment management interface is primarily used for equipment management. It allows users to create equipment ledgers and files to manage robots, camera groups, and other equipment. Users can add, edit, and delete equipment. For example, when adding a robot, the user must enter the robot's application location, IP address, name, type, ID, streaming method, visible light / thermal imaging address, port, login name, and password to complete the equipment information and establish a comprehensive equipment file.

[0092] The system user management interface manages the personnel and accounts logged into the system from four levels: department, organization, user, and role.

[0093] Track-mounted robots, legged robots, and fixed-point cameras are used to perform inspection tasks within the designated inspection area and collect and upload inspection data in real time. To ensure the reliability of the robot system's power supply, power is drawn from the nearest lighting distribution box on site. Each robot is equipped with at least one charging pile. Each track-mounted robot requires one 220V, 10A power supply on site; each legged robot requires one 220V, 4.9A power supply on site. Fixed-point cameras can share a single control box for power supply, which is drawn from the nearest lighting power junction box. In one specific embodiment, the entire system shares a single application server. The robots can communicate with the application server via the IV zone office network WIFI (locally installed wireless WIFI terminal). Signals from cameras and other devices are aggregated at the local control box and then connected to the nearest IV zone office network. The server is installed in the IV zone server room and connected to the nuclear power plant's IV zone office network via a network cable. It should be noted that the specific power supply data mentioned above is for illustrative purposes only and can be configured according to actual needs.

[0094] Specifically, the robot incorporates a GPU computing unit. Data collected by the robot's front end is preprocessed by an edge computing module before being uploaded, reducing system latency and improving data processing efficiency. With a voice module, the robot can be activated by voice, enabling two-way communication between the control room and the work area. This inspection equipment integrates cutting-edge technologies such as artificial intelligence, big data, edge computing, AI recognition, and 5G convergence, ensuring stable and reliable operation and enabling intelligent inspection, digital management, and unmanned operation.

[0095] The battery section features a fully enclosed metal cavity design for protection. In case of an accident, the cavity effectively isolates the risk and reduces adverse effects on the pipeline. The intelligent inspection robot has a built-in battery power detection circuit, and the lower limit of the battery power alarm can be manually set. Once the battery power is detected to be lower than the set value, the robot will automatically stop the current inspection task and issue an alarm; it will then autonomously run to the dock for charging. After autonomous charging is completed, the intelligent inspection robot can autonomously resume the inspection task.

[0096] Meanwhile, the system is equipped with comprehensive power allocation management measures, including early warning power settings, real-time power monitoring, and robust automatic charging control logic. The basic power control logic is as follows: Optimized battery energy distribution ensures that the robot is always powered to work; the robot can set the remaining battery power as emergency backup power; when the robot has sufficient power, it can complete the inspection task as planned, and automatically return to charge after the task is completed, keeping the battery at full charge.

[0097] When the robot's battery level reaches the protection limit set by the system, the robot can interrupt the task and automatically return to charging. When the robot is in low battery protection mode, if a high-priority task occurs, the robot can use the emergency backup power to perform the emergency task.

[0098] The track-mounted robot is the core component of the entire inspection system, undertaking the main functions of inspection within the corridor and on-site handling. The main components of the inspection robot include: vehicle body, drive motor, control box, infrared lidar, ultrasonic obstacle avoidance, 360° all-angle camera, infrared thermal imaging lens, LED lighting fixtures, toxic, harmful and flammable gas sensors, ionization smoke sensor, temperature / humidity probe, two-way voice intercom speaker, etc.

[0099] The inspection robot is powered by a high-capacity lithium battery pack, which is automatically charged via deployed charging stations. After completing its inspection task, the robot automatically locates a charging station to recharge, even in extreme situations such as prolonged power outages.

[0100] 1) Currently, contact charging technology is mature and will not generate abnormal heat. 2) Less prone to poor contact, resulting in higher charging efficiency; 3) Higher transmission efficiency and shorter charging time compared to wireless methods; 4) During the charging process, it only needs to supply power to the lithium battery, which causes less damage to the battery and ensures the lifespan of the lithium battery.

[0101] 5) The charging station adopts a spark-proof design. When the robot moves to the charging station, the charging contacts on the robot body are pressed by the spring contact piece. When the robot and the charging station are aligned, the charging station starts to be powered on to prevent sparks from being generated when the charging station contacts come into contact with the robot during charging.

[0102] The main components of the legged robot include: a bionic quadruped chassis, a core control board, an intelligent gimbal, a high-definition camera, an infrared camera, environmental sensors, a 3D LiDAR, an audio sensor, and an acoustic camera. The legged robot possesses a rich variety of flexible movement postures, enabling it to walk, trot, climb, and overcome obstacles. The legged robot's LiDAR (the number of which can be set according to actual needs, such as 1, 2, 3, etc., preferably 4) can effectively perceive and identify spatial information at a distance and below, ensuring safe operation in various complex terrains. The product's composite multi-functional intelligent gimbal combines image recognition, infrared thermal imaging temperature measurement, acoustic imaging, and other sensory data. Through complex judgment logic and calculation, it accurately analyzes and judges the on-site setup status, enabling unattended equipment inspection and diagnosis. The legged robot can flexibly ascend and descend stairs on certain slopes, navigating complex terrain and autonomously in extreme environments such as darkness, bright light, and even no light source.

[0103] The legged robot has two charging methods: direct plug-in charging or automatic charging via a charging station inside the charging room. The charging station's input voltage is 220V / 50Hz. The robot's power supply module uses filtering technology to filter out high-frequency pulses from the external power grid from interfering with the power supply.

[0104] In some embodiments, such as Figure 4 As shown, the integrated utility tunnel robot inspection system also includes a data platform and a reality twin platform. The data platform is used to receive and integrate the final data from the robot control backend; the reality twin platform is used to retrieve and display data from the data platform.

[0105] By implementing this invention, the following beneficial effects are achieved: This invention discloses a method and system for robotic inspection of integrated utility tunnels. The invention divides the inspection area into various inspection equipment zones based on the positional relationship between the tunnel and the equipment. The combination of a track-mounted robot, a legged robot, and a fixed-point camera solves the problem that existing single-type inspection robots cannot simultaneously adapt to the complex and rugged ground environment and narrow, restricted passageways within integrated utility tunnels, resulting in blind spots and the inability to achieve comprehensive, blind-spot-free monitoring. Furthermore, the invention can plan the inspection range based on the user-selected inspection task and the various inspection equipment zones, matching user needs and environmental factors. In addition, it generates an inspection analysis report based on the inspection data, including environmental and equipment information, to support fault tracing and trend prediction.

[0106] This invention establishes a comprehensive utility tunnel robot inspection method and system suitable for remote communication and autonomous patrol. It employs three monitoring devices—a track-mounted robot, a legged robot, and a fixed-point camera—as monitoring carriers. Based on an intelligent inspection system with coordinated scheduling of multiple monitoring devices, it achieves functions such as internal environment detection, video monitoring, infrared thermal imaging, and AI intelligent analysis. It can intelligently inspect the comprehensive utility tunnel area, automatically identifying anomalies in power cables, various pipelines, and the tunnel environment, issuing high-temperature warnings, and detecting leaks and seepage within pipelines. This improves the efficiency and quality of operation and maintenance work, reducing manpower and increasing efficiency.

[0107] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0108] It is understood that the above embodiments only illustrate some implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can freely combine the above embodiments or technical features without departing from the concept of the present invention, and can also make several modifications and improvements, all of which fall within the protection scope of the present invention. That is, the embodiments described "in some embodiments" can be freely combined with any of the preceding and following embodiments. Therefore, all equivalent transformations and modifications made within the scope of the claims of the present invention should be covered by the claims of the present invention.

Claims

1. A method for robotic inspection of integrated utility tunnels, characterized in that, Applied to robot control platforms, it includes the following steps: S10: Call the pre-built 3D cloud point map of the utility tunnel and the equipment ledger to obtain the positional relationship between the utility tunnel and the equipment, and divide the inspection area according to the positional relationship between the utility tunnel and the equipment. The inspection area includes the track-type robot inspection area, the legged robot inspection area and the fixed-point camera inspection area. S20: Based on the inspection task selected by the user, and in conjunction with the inspection areas of the track-mounted robot, the legged robot, and the fixed-point camera, the inspection ranges are planned for the track-mounted robot, the legged robot, and the fixed-point camera, respectively. S30: Drive several of the track-mounted robots, the legged robots, and the fixed-point cameras to perform the inspection task within the inspection range and collect inspection data in real time; S40: Generate an inspection analysis report including environmental and equipment information based on the inspection data.

2. The integrated utility tunnel robot inspection method according to claim 1, characterized in that, The track-mounted robot inspection area is the upper space of the corridor without any equipment obstructions or wall partitions; the legged robot inspection area is the lower space of the corridor with a wide space and a gentle slope; the fixed-point camera inspection area is the blind spot area that the track-mounted robot and the legged robot cannot reach. The inspection content of the track-mounted robot inspection area, the legged robot inspection area, and the fixed-point camera inspection area overlaps.

3. The integrated utility tunnel robot inspection method according to claim 1, characterized in that, The track-mounted robot inspection area and the legged robot inspection area have several fireproof partition doors to meet the fireproof partition requirements. When the track-mounted robot or the legged robot moves to a preset distance in front of the fireproof partition door, it controls the fireproof partition door to execute an opening command. After the track-mounted robot or the legged robot has safely passed through, it controls the fireproof partition door to execute a closing command.

4. The integrated utility tunnel robot inspection method according to claim 1, characterized in that, The method after S10 also includes: S11: Semantic processing is performed on the three-dimensional cloud point map of the utility tunnel to divide the tunnel space into grid management units. Each grid management unit is assigned a risk label with an initial weight based on the equipment ledger. The inspection area is divided into at least two levels according to the risk label, including a first-level key area and a second-level regular area. Different inspection strategies are formulated for the track-mounted robot, the legged robot, and the fixed-point camera for different levels of the inspection area.

5. The integrated utility tunnel robot inspection method according to claim 4, characterized in that, The weight of the risk label is associated with historical inspection data and dynamically adjusted based on the real-time collected inspection data.

6. The integrated utility tunnel robot inspection method according to claim 1, characterized in that, During the execution of the inspection task, the battery charge status and current inspection task complexity of each of the track-mounted robots and each of the legged robots are monitored and evaluated in real time. The remaining battery life is predicted by combining the battery charge status and the current inspection task complexity to control whether to trigger the return mode.

7. The integrated utility tunnel robot inspection method according to claim 6, characterized in that, When any of the track-mounted or legged robots triggers a return-to-home charging, if there are other idle and fully charged backup robots within a preset range, the coordinates of the unfinished task points and the detection requirements are sent to the backup robots, driving them to continue executing the current task.

8. The integrated utility tunnel robot inspection method according to claim 1, characterized in that, The inspection data is preprocessed by edge computing on the robot during real-time acquisition; the inspection data is used to align and fuse with the three-dimensional cloud point map of the utility tunnel to generate a real-time three-dimensional view of the utility tunnel status.

9. The integrated utility tunnel robot inspection method according to claim 1, characterized in that, The environmental information includes at least one of the following: leak detection information, environmental anomaly information, and environmental parameter information; The equipment information includes at least one of the following: equipment reading information, equipment status information, equipment appearance inspection information, and equipment thermal status information.

10. A robotic inspection system for integrated utility tunnels, characterized in that, include: The system includes a robot control platform, a track-mounted robot, a legged robot, and a fixed-point camera; the robot control platform is connected to the track-mounted robot, the legged robot, and the fixed-point camera via at least one data node. The robot control platform is used to call a pre-built 3D cloud point map of the utility tunnel and equipment ledger to obtain the positional relationship between the utility tunnel and the equipment, and divide the inspection area into a track-type robot inspection area, a legged robot inspection area and a fixed-point camera inspection area according to the positional relationship. Based on the inspection task selected by the user, the inspection areas of the track-mounted robot, the legged robot, and the fixed-point camera are combined to plan the inspection ranges for the track-mounted robot, the legged robot, and the fixed-point camera, respectively. The system drives the track-mounted robot, the legged robot, and the fixed-point camera to perform the inspection task within the inspection range and collect inspection data in real time, and generates an inspection analysis report based on the inspection data. The track-mounted robot, the legged robot, and the fixed-point camera are used to perform the inspection task within the inspection range and collect and upload inspection data in real time.

Images