Intelligent inspection control system and control method

Abstract

The invention relates to an intelligent inspection control system and method. The system comprises an environment sensing module, a situation recognition module, a task management and scheduling module, a path planning module, an execution control module and a communication linkage module. The system can collect various environmental parameters of an inspection area in real time, dynamically adjust the priority of an inspection task after identifying an abnormal state, and generate an optimal path to drive a robot to execute response operation; meanwhile, abnormal information, images, videos, sensor snapshots and other contents are uploaded to a background management platform through a communication linkage module, and sound-light alarm equipment, broadcast equipment and other equipment are linked to carry out field response. The control method has task preemption and path reconstruction capabilities, and supports distributed sensing deployment and whole-process execution state monitoring. Compared with the prior art, integrated fusion of environment situation awareness, task dynamic scheduling and response control is achieved, and the adaptive capacity, the response speed and the safety guarantee level of the system to complex scenes are improved.

Description

Technical Field

[0001] The present invention relates to the field of patrol control technology, and in particular to an intelligent patrol control system and control method, which are particularly suitable for an intelligent patrol system with multiple sensing capabilities and integrating environmental perception and dynamic task scheduling. Background Art

[0002] With the development of artificial intelligence, robotics, and the Internet of Things (IoT) technologies, inspection robots are widely used in sectors such as power generation, petrochemicals, rail transit, coal mining, and industrial parks. They replace manual labor in high-risk or repetitive inspection tasks, improving inspection efficiency and safety. Existing inspection robots typically patrol along pre-set routes or operate under manual control.

[0003] To improve the efficiency of human-machine interaction, existing technologies have proposed image recognition-assisted control solutions. For example, patent application number 202311480569.8 discloses a control method for a track-mounted inspection robot. This method identifies the position of a worker's light source in a photograph to determine the intended control command, thereby enabling control of the inspection robot's movements without a remote control. This method offers certain convenience and reduces the professional requirements for operators.

[0004] However, the above solution still has the following shortcomings: The control signal is single and limited to human input: The method relies on the operator to express control commands through specific postures and light positions. The command types are limited and the operating environment (such as lighting conditions) is demanding. Lack of environmental situational awareness: The robot cannot independently determine the current environmental status and cannot dynamically adjust its inspection strategy based on real-time working conditions (such as temperature rise, gas leaks, fire alarms, etc.); Lack of dynamic task scheduling capabilities: The robot cannot combine external sensory information to optimize paths, switch tasks, or schedule tasks, limiting the system's intelligence and ability to adapt to complex working conditions. Not suitable for unmanned or emergency response scenarios: In emergencies, robots lack autonomous decision-making and collaborative response mechanisms, and cannot achieve real-time response and task reconstruction.

[0005] Therefore, it is urgent to propose an intelligent inspection control system and method that integrates multi-source environmental perception information and has the ability to dynamically schedule tasks, so as to solve the problems of passive control methods, single response, and insufficient intelligence in existing technologies, and realize efficient and autonomous inspections in complex scenarios. Summary of the Invention

[0006] The present invention aims to solve the core technical problem of the existing inspection robot control method lacking the ability to perceive the environmental status in real time and respond autonomously.

[0007] Existing technologies mostly rely on manual instructions or fixed path control methods. They are unable to perceive key parameters of the environment such as temperature, humidity, smoke, gas, flame, etc., nor can they dynamically adjust inspection tasks or respond to emergencies based on this information. As a result, the system has a low level of intelligence and poor adaptability in complex or high-risk inspection scenarios, making it difficult to achieve truly intelligent inspection.

[0008] An intelligent inspection control system, comprising: An environmental sensing module is used to collect environmental parameter data of the inspection area in real time, wherein the environmental parameter data includes at least two of temperature, humidity, smoke concentration, combustible gas concentration, flame characteristics and image information; A situation recognition module is used to perform fusion analysis on the environmental parameter data to identify whether there is an abnormal state in the current inspection area; The task management and scheduling module is used to adjust the execution priority of the current inspection task according to preset rules when an abnormal state is identified, and generate an updated task queue; A path planning module, configured to calculate a travel path of the inspection robot according to the updated task queue; An execution control module, used to drive the inspection robot to perform corresponding task actions according to the travel path; The communication linkage module is used to upload abnormal information to the background management platform when an abnormal state is identified, and can link external devices to perform early warning response operations.

[0009] Technical principle: This technical solution proposes an intelligent inspection control system. Its core concept is to integrate an environmental perception module, a situation recognition module, a task scheduling module, a path planning module, an execution control module, and a communication linkage module into the inspection system. By collecting a variety of environmental data, identifying on-site situation changes, and combining task scheduling and path planning algorithms, the inspection robot achieves a closed-loop control system of "perception-decision-execution-feedback." The system not only supports basic fixed-point inspection functions but also enables autonomous judgment and dynamic response in complex or unexpected environments, significantly enhancing the system's intelligence and adaptability.

[0010] Compared with the existing system that relies on manual input and fixed path inspection, this technical solution has the following significant technical effects: Possessing environmental self-perception capabilities: It can identify multi-dimensional environmental conditions such as temperature, smoke, and gas, and achieve intelligent recognition without human intervention; Dynamic task scheduling and route reconstruction capabilities: The robot proactively adjusts tasks and routes based on abnormal situations, enabling faster responses and more flexible logic. Adaptable to complex working conditions: especially suitable for application fields with environmental changes or safety risks such as power, petrochemical, coal mines, and tunnels; Supports linkage response and remote management: not only autonomous execution, but also linkage with on-site early warning systems or uploading information to remote backends; Improve system robustness and scalability: The system architecture is modularly designed, each function can be deployed on demand, and is easy to maintain and upgrade.

[0011] Furthermore, the situation recognition module includes a rule judgment submodule and a model analysis submodule, which are respectively used to make comprehensive judgments on abnormal states based on thresholds and pattern recognition.

[0012] This technical solution proposes a dual-mechanism fusion approach to environmental situation recognition, balancing response efficiency and recognition accuracy to address the limitations of a single recognition mechanism. The rule-based judgment submodule uses traditional threshold comparison methods, such as identifying abnormalities when the temperature exceeds 60°C, the smoke concentration exceeds a certain part-per-million (PPM), or the combustible gas concentration exceeds the standard. The model analysis submodule, based on historical data and scenario characteristics, constructs lightweight pattern recognition models (such as SVM, KNN, and shallow neural networks) to identify complex or gradually changing risks (such as hidden gas leaks and early fire trends). The system integrates these two types of judgments to make decisions, avoiding missed and false alarms. This fusion mechanism achieves an "explainable and learnable" situation recognition approach, retaining the traditional reliable logical judgment capabilities while introducing self-learning features for greater adaptability.

[0013] Compared with traditional situation recognition methods, this technical solution has the following advantages: 1. High recognition accuracy and low false alarm rate The dual-path fusion judgment mechanism can effectively filter out interference or critical value fluctuations, avoiding false alarms caused by failure of a single sensor.

[0014] 2. Combine speed and adaptability to complex scenarios Rule judgment is used to quickly respond to sudden out-of-limit events, while model analysis is suitable for identifying gradual, nonlinear or comprehensive risks.

[0015] 3. Support adaptive optimization Continuous performance improvement can be achieved through model updates and field data accumulation to adapt to different application scenarios and environmental conditions.

[0016] 4. Easy to expand Additional recognition models can be added as needed, such as video image anomaly recognition, audio anomaly analysis, etc., to enhance overall perception capabilities Furthermore, the task management and scheduling module is provided with a task priority queue, and the priority is dynamically adjusted according to the event level, risk level and task type.

[0017] The principle behind the above technical solution is to dynamically adjust the execution order of robot tasks through a "task priority queue" to address the problem of multi-task conflicts in complex inspection scenarios. The priority queue sorting strategy integrates three core factors: 1. Event Level: whether it is a fire alarm, gas leak, communication interruption, etc. 2. Risk Level: This includes the severity of the potential consequences of the incident and whether it involves human or equipment safety. 3. Task Type: Such as routine inspection, follow-up confirmation, video sampling, fire situation review, etc.

[0018] The system assigns weights based on the above factors, calculates the overall priority value of the task, and uses this to sort the task queue so that the robot can perform the most urgent and critical tasks first.

[0019] Compared with traditional inspection systems that use static task sequences or manual control scheduling, this technical solution brings the following improvements: 1. Dynamic response to emergencies Once a major anomaly such as fire or gas leak is identified, the system can immediately adjust the priority queue to ensure that the robot arrives at the scene as soon as possible.

[0020] 2. Improve task execution efficiency The system dynamically allocates resources based on task importance to avoid high-priority tasks being blocked by low-priority tasks and improve the timeliness of task completion.

[0021] 3. Support multi-tasking concurrency and collaborative processing In a multi-robot system, tasks can be assigned to different robots based on priority, forming a multi-terminal parallel processing mechanism.

[0022] 4. Easy to maintain and highly configurable Scheduling rules and weight factors can be adjusted according to on-site operation and maintenance strategies, and have good engineering adaptability and strategy scalability.

[0023] Furthermore, the path planning module supports generating the shortest inspection path based on the A* algorithm or the Dijkstra algorithm, and has an obstacle avoidance function.

[0024] This technical solution is based on an optimal path planning and obstacle avoidance mechanism using heuristic or graph search algorithms. This dynamic path planning mechanism, combining classic graph search algorithms with obstacle recognition capabilities, is used in the path generation module of intelligent inspection control systems. Its core objectives are to ensure the optimal path for inspection tasks in the current environment (e.g., shortest time, lowest energy consumption); and to enable real-time obstacle detection and avoidance, ensuring the robot's feasibility and safety in complex environments. The system prioritizes constructing a path search graph based on the A* or Dijkstra algorithm. This algorithm, combined with real-time map data from the environmental perception module, generates inspection paths based on parameters such as the task starting point, target point, task weight, and obstacle location.

[0025] The technical effects of the above technical solution are: 1. Realize dynamic optimal path generation Path search is performed through the A* or Dijkstra algorithm to ensure that a reasonable path can be quickly generated in multi-task concurrency or large scenarios, saving energy consumption and execution time.

[0026] 2. Strong ability to adapt to challenging environments The static / dynamic obstacles identified by the fusion perception module ensure that the robot can execute uninterruptedly in complex scenarios, improving the system's robustness and fault tolerance.

[0027] 3. Support multi-task path reconstruction The task execution path can be dynamically updated in combination with the task scheduling module to implement advanced scheduling strategies such as asynchronous task insertion and multi-task mixed execution.

[0028] 4. Easy to integrate with existing navigation / positioning modules The output of the path planning module can be directly connected to execution layers such as SLAM, UWB, and laser navigation systems, improving system integration and implementation feasibility.

[0029] Furthermore, the communication linkage module is connected to the sound and light alarm, broadcasting equipment, monitoring system and cloud platform for on-site response and remote synchronization. Application of multi-channel linkage response and remote synchronization mechanism in intelligent inspection system.

[0030] This technical solution provides an intelligent response mechanism based on multi-terminal linkage. The communication linkage module not only uploads abnormality information to the backend platform but also triggers relevant on-site equipment (such as audible and visual alarms, emergency broadcasts, and video linkage) to execute early warning operations, thereby achieving a "local + remote" dual response closed loop. By establishing communication interfaces between the inspection system and the Industrial Internet of Things (IIoT) platform, SCADA system, and emergency broadcast platform, the inspection robot's recognition results can be instantly converted into on-site actions and management instructions, comprehensively improving the system's operational effectiveness and rapid response capabilities.

[0031] The above technical solution has the following effects: 1. Exception response is more efficient and direct: After identifying an exception, the robot can directly drive the on-site equipment to respond without waiting for manual confirmation, greatly saving handling time.

[0032] 2. Achieve a remote + local linkage closed loop: Abnormal data can be uploaded to the cloud platform synchronously to achieve remote supervision. It is suitable for applications such as large factories, tunnels, and unmanned substations.

[0033] 3. Improve system collaboration and scenario compatibility: By connecting to existing security or industrial platforms, the system can be integrated into the original safety / control system without the need for complete reconstruction, facilitating promotion and deployment.

[0034] 4. Support customized configuration for multiple abnormal scenarios: Response strategies can be set according to customer industries (such as power, petrochemical, coal mining, etc.) to meet different levels of risk management needs.

[0035] Furthermore, the various sensors in the environmental perception module are deployed in a distributed manner, and can be deployed separately on the robot body and fixed nodes at the inspection site. This technical solution adopts a distributed sensor deployment architecture, dividing the various sensors in the environmental perception module into two deployment methods: mobile nodes (robot body): dynamic perception is performed as the inspection robot moves; fixed nodes (deployed at the inspection site): long-term online sensors are deployed in specific hazardous areas and key locations to achieve fixed-point real-time monitoring. By integrating dynamic perception with fixed-point perception, the system can achieve higher-density, more timely, and lower-blind-spot environmental monitoring capabilities, while reducing the robot's sensor load, improving response range, and monitoring efficiency.

[0036] The above technical solution has the following technical effects: 1. Improve environmental monitoring coverage and frequency: Mobile robots offer patrol flexibility, while fixed nodes offer monitoring continuity. The combination of the two complements both space and time.

[0037] 2. Reduce system blind spots and false alarm rates: Set fixed monitoring points in key areas, do not rely on robots to cover all scenes, and significantly reduce monitoring blind spots and data gaps.

[0038] 3. Improve response timeliness and review capabilities: Fixed nodes provide early warnings for abnormal events, and robots, as mobile terminals, provide precise review and on-site intervention, achieving efficient collaboration.

[0039] 4. Reduce energy consumption and operation and maintenance pressure: Some areas are monitored through fixed nodes, reducing the number of repeated round trips by robots, improving inspection efficiency and extending operating time.

[0040] The intelligent inspection control method includes the following steps: (1) The environmental perception module on the inspection robot collects environmental parameter data of the inspection area in real time, wherein the environmental parameter data includes at least two of temperature, humidity, smoke concentration, combustible gas concentration, flame characteristics and image information; (2) The collected environmental parameter data is transmitted to the situation recognition module for multi-source fusion analysis to identify whether there is an abnormal state; (3) When an abnormal state is identified, the task management and scheduling module is triggered to dynamically adjust the current inspection task and generate an updated task queue and corresponding priority; (4) The path planning module replans the inspection path based on the updated task queue; (5) The execution control module drives the inspection robot to complete the corresponding inspection action according to the inspection path; (6) The communication linkage module uploads the abnormal status to the background platform and links the external early warning equipment to perform response operations.

[0041] The technical principle of this solution is to integrate environmental perception, intelligent identification, task scheduling, path generation, and execution feedback into a closed-loop process, achieving full automation and intelligent response during robot inspection tasks. The system no longer relies on fixed preset paths or manual remote control input, but instead makes autonomous judgments and action decisions based on real-time environmental conditions. This approach is essentially a task-triggered inspection control mechanism driven by environmental conditions.

[0042] Technical effects of this technical solution: Compared with the existing inspection mode based on "fixed path + manual control", this technical solution has the following advantages: Fully autonomous: The robot can perform dynamic task switching and path optimization without relying on remote control input; Timely response: Ability to detect and respond to abnormalities in the first place, improving processing efficiency; Closed-loop process: includes the entire process of "perception → identification → scheduling → execution → feedback", with a high degree of automatic control; Strong scalability: It can be connected to different types of sensors and platform systems, and has good engineering feasibility; Adaptable to complex environments: Applicable to tunnels, power plants, substations, chemical plants and other highly dynamic environments; Reduce human dependence and error risks: Improve the overall stability and intelligence level of the system Furthermore, in step (3), the task management and scheduling module inserts the new task into the front end of the original inspection task queue according to the abnormality level and triggers path reconstruction.

[0043] The core of this technical solution lies in the fact that when an inspection robot identifies an abnormal event, the system inserts the corresponding emergency response task into the queue of ongoing inspection tasks based on the severity of the event, instantly triggering path reconstruction. This method allows the robot to flexibly respond to emergencies without interrupting system operation, while always maintaining the scheduling principle of executing the highest-priority task first, forming an intelligent path control system with a task preemption mechanism.

[0044] The above technical solution has the following technical effects: 1. Improve emergency response efficiency: Abnormal events can immediately trigger inspection responses, greatly shortening the response time from identification to disposal, making it suitable for high-risk scenarios.

[0045] 2. Implement a task priority scheduling mechanism: The system has the ability to control task sequencing, so that inspections no longer rely solely on static paths or a single task sequence.

[0046] 3. Maintain task continuity: Unfinished original tasks can be temporarily stored and rescheduled after high-priority tasks are completed to ensure task integrity.

[0047] 4. Support large-scale scenario scheduling needs: Combined with a multi-robot system, cross-robot scheduling and optimal resource matching of tasks of different priorities can be achieved.

[0048] Furthermore, in step (5), the execution control module feeds back the execution status and location data to the background management platform in real time during the execution of the task.

[0049] The above technical solution proposes a dynamic status monitoring and information feedback mechanism during task execution. Through the execution control module, key parameters such as the current task execution status, progress, geographic location, and execution efficiency are monitored in real time and synchronously transmitted to the backend management platform, enabling the system to visualize, trace, and intervene in the entire task process. The principle is as follows: While driving the robot to perform a task, the execution control module periodically or event-drivenly collects information including execution progress, task status, current location, and abnormal event records, and uploads it to the backend via a communication link (such as 5G / WiFi / LTE / wired). This mechanism enables task process monitoring, remote operation and maintenance collaboration, execution process data tracking, and subsequent fault analysis support.

[0050] The above technical solution has the following technical effects: 1. Realize visual monitoring of the entire task process: the background can grasp the robot task execution status, location progress and abnormal situations at any time.

[0051] 2. Improve operation and maintenance efficiency and scheduling capabilities: When a task anomaly or timeout is identified, the manager can remotely intervene or arrange for other robots to take over the task.

[0052] 3. Support task data archiving and analysis: The system can record key indicators such as execution efficiency, common anomalies, and repeated failure paths to assist in system optimization.

[0053] 4. Enhance system security and robustness: The state synchronization mechanism can effectively prevent the risk of the robot "losing connection" or malfunctioning without anyone noticing.

[0054] Furthermore, in step (6), the communication linkage module pushes the on-site photos, videos or sensor data snapshots to the backend while uploading the abnormal information. This technical solution forms a structured + unstructured fusion information upload mechanism by synchronously attaching visual and sensor data snapshots during the abnormal event reporting process, so that the backend personnel or AI system can assist in judging the authenticity, severity and priority treatment method when receiving abnormal alarms, thereby improving the quality of alarms and reducing the false alarm rate. Specifically, when the system identifies an environmental anomaly (such as smoke, excessive gas, high temperature, etc.), the communication linkage module not only uploads the alarm type, location and time, but also automatically captures photos, video clips or sensor data segments on the scene, and packages them together and uploads them to the backend management platform.

[0055] The above technical solution has the following technical effects: 1. Abnormal warning information is more intuitive and reliable: By synchronously uploading "perceptual evidence" such as images and videos, back-end staff can intuitively see the abnormal scene and improve the accuracy of judgment.

[0056] 2. Reduce false alarm rates and response delays: False alarms or suspected alarms can be confirmed remotely by the operator, eliminating the need for repeated robot verification, improving efficiency and reducing disruptions.

[0057] 3. Provide data support for post-event tracing and AI training: Abnormal images and sensor records retained by the system can be used for post-event analysis and model training to improve the intelligence level of the recognition system.

[0058] 4. It is conducive to collaborative decision-making on multiple platforms: through standardized upload formats (such as JSON+attachment file stream), it can be connected with industry supervision platforms such as public security, fire protection, and electricity to achieve cross-platform linkage response. DETAILED DESCRIPTION

[0059] The following is further described in detail through specific implementation methods: Intelligent inspection control system, including the following core modules: Environmental perception module: includes temperature and humidity sensors, smoke sensors, combustible gas detectors, infrared flame identifiers, cameras, etc.; used to collect key environmental parameters in the inspection area in real time.

[0060] Situation recognition module: This module integrates and processes various data from the environmental perception module; identifies whether there are environmental anomalies, such as high temperature, leakage, fire, etc.; it can include an abnormal event library and threshold judgment model, or use a shallow learning model to improve recognition accuracy.

[0061] Task management and scheduling module: adjusts tasks in real time based on inspection plans and identification results; has the capabilities of priority sorting, task queue jumping, and emergency response scheduling for abnormal tasks.

[0062] Path planning module: recalculates the optimal inspection path after the task changes; can combine constraint objectives such as obstacle avoidance, optimal energy consumption, and shortest time.

[0063] Execution control module: controls the robot chassis drive, visual alignment, positioning adjustment, robotic arm movement, etc.; performs inspection actions based on path planning results and task instructions.

[0064] Communication linkage module: upload key status to the background platform; link security system, fire protection system or emergency broadcast system; support remote control or task distribution.

[0065] Backend management platform (optional): used for device registration, task configuration, policy management, data analysis and remote control.

[0066] Implementation path of this technical solution: 1. System deployment: Integrate the above modules onto a wheeled or tracked mobile inspection robot platform. Sensors are configured on the robot body or deployed at key locations on site based on the characteristics of the inspection area (forming a fixed-mobile fusion perception network).

[0067] 2. Inspection process: After startup, the system begins inspection along the default path; it senses environmental data in real time; if an anomaly is identified (such as temperature exceeding the limit or smoke alarm), the abnormal task is immediately inserted through the scheduling module; the path planning module is called to regenerate the task path; the control module drives the robot to turn to the abnormal point for close inspection; the on-site images and sensor data are uploaded to the background; if necessary, the early warning device can be triggered or a push notification can be sent to the control center; after the anomaly is resolved, the original inspection task is resumed or subsequent tasks continue to be scheduled according to priority.

[0068] The situation recognition module consists of two parts: a rule judgment submodule: which sets threshold conditions for multiple sensors (such as temperature > 60°C, CH4 > 800ppm) and uses Boolean logic to determine whether an abnormality flag is triggered; a model analysis submodule: which takes as input the time series or feature combination of multi-source sensor data and outputs a probability or label result, such as "normal", "suspicious", or "abnormal"; data fusion processing method (example): if the rule judgment is abnormal and the model analysis also predicts an abnormality → immediately confirm and report; if the rule judgment is normal, but the model analysis confidence exceeds the threshold (such as > 80%) → mark as "warning" status; if the rule judgment is abnormal but the model analysis is normal → can be set to "manual confirmation" or "delayed verification"; model training strategy (optional): use a pre-trained model for initial deployment and perform incremental learning based on field data; the model can be updated regularly to adapt to environmental changes (such as humidity in different seasons affecting fire recognition accuracy).

[0069] The task management and scheduling module is provided with a task priority queue, and the priority is dynamically adjusted according to the event level, risk level and task type.

[0070] 1. Data structure setting: Each inspection task has the following fields in the task management system

[0071] 2. Priority calculation method (example formula), dynamic priority score: Priority_Score = W1 × Event_Level_Score + W2 × Risk_Level + W3 ×Task_Type_Score Among them: W1 / W2 / W3 are configurable weight parameters; Event_Level_Score can be set to: High = 3, Medium = 2, Low = 1; Task_Type_Score can be set to: urgent task = 3, review task = 2, regular task = 1.

[0072] 3. Scheduling logic and process: abnormal event trigger / inspection task initialization → task information is written into the task pool → the task scheduling module calculates the priority of each task → sorts and generates a task execution queue → the control module schedules the inspection robot to execute according to the queue.

[0073] When a new high-priority task enters the queue, the scheduling module will interrupt the current low-priority task (if preemption is allowed) and readjust the task order; after the task is executed, the status is updated synchronously and the queue is automatically advanced.

[0074] The path planning module supports generating the shortest inspection path based on the A* algorithm or the Dijkstra algorithm, and has an obstacle avoidance function.

[0075] The implementation path of this technical solution is: 1. Fusion of map modeling and environmental perception The system uses grid maps, topological maps, or mixed grid maps to model inspection areas; it imports obstacle information from sensors such as cameras, radars, infrared, and ultrasonic waves in real time; and it marks areas where obstacles are located as "inaccessible" nodes. 2. Path search algorithm selection and execution A* algorithm features: A graph search algorithm based on heuristic functions, suitable for fast path planning on medium and large maps; cost function f(n) = g(n) + h(n), where g(n) is the actual cost from the starting point to the current node, and h(n) is the estimated cost from the current node to the target point; it can prioritize approaching the target area to improve planning efficiency.

[0076] Features of Dijkstra algorithm: used to obtain the shortest path between any two points in the entire graph; no heuristic function is required, ensuring the optimal solution for the shortest path, but the computational complexity is slightly higher.

[0077] 3. Obstacle avoidance mechanism During the path planning process, obstacle data is called up in real time. If new obstacles or dynamic obstacles (such as people or mobile equipment) appear during the robot's movement, the system will automatically trigger path reconstruction. "Soft obstacle" settings can be introduced (such as increasing the detour weight in dangerous areas, but not completely closing them).

[0078] 4. Path Reconstruction Trigger Strategy: Active trigger: receiving a new mission target point; Passive trigger: the path is blocked by a new obstacle; Threshold trigger: replanning when the obstacle approach distance is less than the warning distance (e.g., 50 cm).

[0079] The communication linkage module is connected to the sound and light alarm, broadcasting equipment, monitoring system and cloud platform for on-site response and remote synchronization.

[0080] 1. Module docking and interface definition The communication linkage module connects to the following platforms / devices through multiple communication protocols (such as TCP / IP, MQTT, Modbus, RESTful API, etc.):

[0081] 2. Linked response process logic Abnormal identification → Communication linkage module calls the docking protocol → Execute the following actions in parallel: ─ Sound and light alarm triggers local prompt ─ The PA system plays preset voice ─The monitoring system switches the abnormal point image to the main screen ─The cloud platform writes alarm event records and synchronizes them to the remote central control 3. Response Level Control Strategy (Optional Extension) Set different linkage response actions corresponding to different abnormal levels, for example: Level I anomaly (such as smoke) → prompt sound + cloud platform upload; Level II abnormality (such as gas leakage) → alarm + broadcast + video linkage; Level III abnormality (such as fire) → Trigger sound and light + broadcast + upload data + push to the fire platform or on-duty mobile phone.

[0082] The various sensors in the environmental perception module are arranged in a distributed manner and can be deployed on the robot body and fixed nodes at the inspection site respectively.

[0083] 1. Hardware deployment method

[0084] The robot completes full-coverage inspections while moving; fixed nodes can provide high-frequency, all-weather, and robot-independent status perception supplements.

[0085] 2. Perception Data Synchronization and Fusion Mechanism All sensors are connected to the system's main bus or communication gateway (such as LoRa, NB-IoT, WiFi, Zigbee); the system collects and timestamps data from each node; the perception fusion module uses rules or AI algorithms to perform synchronous analysis of data from different sources; if the long-term data of a fixed node is abnormal, the system can dispatch a robot to review it.

[0086] 3. Linkage Strategy A fixed node triggers an anomaly → dispatches a robot to perform a confirmation task nearby; the robot detects sudden anomalies along the path → transmits the data back to the background and compares it with the fixed node data to assist in determining true or false alarms.

[0087] Implementation path of intelligent inspection control method: Step 1: Multi-source environmental perception The inspection robot continuously collects environmental data while moving, such as: temperature and humidity sensors read temperature and humidity; infrared / flame recognition modules identify flames or heat sources; combustible gas sensors monitor leaks; cameras or vision systems capture images and analyze whether there is intrusion, smoke, fire, etc.

[0088] Step 2: Situation Fusion and Anomaly Identification Input multi-source data into the situation recognition module: fuse sensor data for pattern recognition; set thresholds or use lightweight models to determine whether there are anomalies (such as temperature > 60°C, PM2.5 exceeding the standard, CO gas detected); if a certain abnormal condition is met, it is determined to be an "abnormal state".

[0089] Step 3: Task scheduling and path adjustment Suspend the currently executing task or insert it into the waiting queue; generate a new urgent task and set it to the highest priority; reorder the task queue; call the path planning module to generate the optimal path for the new task.

[0090] Step 4: Task execution and exception handling The execution control module schedules the robot's movement along the new path; upon reaching the abnormal point, it conducts on-site sampling, photography, data recording, etc.; at the same time, it triggers the communication module to upload alarm information, linking the sound and light alarm, warning lights or broadcasting equipment; the background platform can receive information and manually intervene or configure subsequent strategies.

[0091] In step (3), the task management and scheduling module inserts the new task into the front end of the original inspection task queue according to the abnormality level and triggers path reconstruction.

[0092] 1. Abnormal level determination mechanism Abnormality levels are divided into several grades (for example: Grade I is the most severe, Grade II is moderate, and Grade III is mild):

[0093] The judgment is based on: abnormal sensor values ​​(temperature > 65°C, CO concentration > 200 ppm, etc.); multi-source fusion recognition (such as video + smoke + sound combined triggering); and background manual intervention confirmation (if the manual verification mechanism is enabled).

[0094] 2. Task insertion logic The system maintains a dynamic task queue, in which task items include: task number, target location, task type (normal / abnormal), estimated execution time, priority value and other fields; after the abnormal identification is triggered, the following logic is executed.

[0095] Generate a new task → Read the exception level → Assign a priority value → Insert the task to the front of the task queue (if preemptive) or an appropriate position → Call the path planning module to reconstruct the path.

[0096] If the current robot is executing a task, the system determines whether preemption is allowed; if preemption is allowed, the current task is paused and the execution status is recorded, waiting to be resumed; the inserted task is executed first, and the original task continues to be executed after the task is completed.

[0097] 3. Path reconstruction mechanism Inserting a task will cause the target point to change. The path planning module automatically recalculates the path (A* or Dijkstra algorithm) with the new task as the target, and generates new path information to pass to the control module.

[0098] In step (5), the execution control module feeds back the execution status and location data to the background management platform in real time during the execution of the task.

[0099] 1. Example of status information collection fields The status fields collected by the execution control module include but are not limited to:

[0100] 2. Data return strategy Periodic reporting (e.g., uploading execution status every 10 seconds); event-triggered upload (e.g., task start, completion, abnormality identification, loss of connection, etc.); data is uploaded to the backend management platform database or real-time monitoring interface through the communication module; MQTT, HTTPS, WebSocket and other protocols can be used to achieve real-time or lightweight transmission.

[0101] 3. Backend platform reception and application scenarios The backend platform functions include: real-time monitoring of the operating status of each robot; visual display of maps + status heat maps; remote issuance of abort / adjustment commands; historical task logs can be exported / queried / compared; and execution data is used for AI training and efficiency evaluation.

[0102] In step (6), the communication linkage module uploads the abnormal information and pushes the on-site photos, videos or sensor data snapshots to the background.

[0103] 1. Exception information structure definition The uploaded data content includes the following fields:

[0104] 2. Data collection and packaging upload process Anomaly identification is triggered → control module instructions are executed to collect camera images / videos and sensor cache data → the communication module performs packaging and encoding (such as JSON+Base64, video compression format) → uploads to the designated interface of the backend platform (RESTAPI or MQTT Broker, etc.) → the backend stores and displays the anomaly and image information in a pop-up window.

[0105] Images / videos are collected by the robot's forward-facing camera or PTZ camera. Sensor data is a real-time stream segment cached by the system (with flexible configuration of forward / backward time windows). The size of uploaded content is limited by network conditions, and the "snapshot first, video delayed" strategy can be enabled.

[0106] The above description is merely an embodiment of the present invention. Common knowledge regarding the specific structure and characteristics of the solution is not described in detail herein. It should be noted that those skilled in the art may make various modifications and improvements without departing from the structure of the present invention. These modifications and improvements should also be considered within the scope of protection of the present invention and will not affect the effectiveness and practicality of the present invention. The specific embodiments and other descriptions in the specification may be used to interpret the content of the claims.

Claims

1. Intelligent inspection control system, characterized in that: include: An environmental sensing module is used to collect environmental parameter data of the inspection area in real time, wherein the environmental parameter data includes at least two of temperature, humidity, smoke concentration, combustible gas concentration, flame characteristics and image information; A situation recognition module is used to perform fusion analysis on the environmental parameter data to identify whether there is an abnormal state in the current inspection area; The task management and scheduling module is used to adjust the execution priority of the current inspection task according to preset rules when an abnormal state is identified, and generate an updated task queue; A path planning module, configured to calculate a travel path of the inspection robot according to the updated task queue; An execution control module, used to drive the inspection robot to perform corresponding task actions according to the travel path; The communication linkage module is used to upload abnormal information to the background management platform when an abnormal state is identified, and can link external devices to perform early warning response operations.

2. The system according to claim 1, wherein: The situation recognition module includes a rule judgment submodule and a model analysis submodule, which are respectively used to make comprehensive judgments on abnormal states based on thresholds and pattern recognition.

3. The system according to claim 2, characterized in that The task management and scheduling module is provided with a task priority queue, and the priority is dynamically adjusted according to the event level, risk level and task type.

4. The system according to claim 3, characterized in that The path planning module supports generating the shortest inspection path based on the A* algorithm or the Dijkstra algorithm, and has an obstacle avoidance function.

5. The system according to claim 4, characterized in that The communication linkage module is connected to the sound and light alarm, broadcasting equipment, monitoring system and cloud platform for on-site response and remote synchronization.

6. The system according to any one of claims 1 to 5, characterized in that: The various sensors in the environmental perception module are arranged in a distributed manner and can be deployed on the robot body and fixed nodes at the inspection site respectively.

7. Intelligent inspection control method, characterized in that: The steps include: (1) The environmental perception module on the inspection robot collects environmental parameter data of the inspection area in real time, wherein the environmental parameter data includes at least two of temperature, humidity, smoke concentration, combustible gas concentration, flame characteristics, and image information; (2) The collected environmental parameter data is transmitted to the situation recognition module for multi-source fusion analysis to identify whether there is an abnormal state; (3) When an abnormal state is identified, the task management and scheduling module is triggered to dynamically adjust the current inspection task and generate an updated task queue and corresponding priority; (4) The path planning module replans the inspection path based on the updated task queue; (5) The execution control module drives the inspection robot to complete the corresponding inspection action according to the inspection path; (6) The communication linkage module uploads the abnormal status to the background platform and links the external warning equipment to perform response operations.

8. The method according to claim 7, characterized in that In step (3), the task management and scheduling module inserts the new task into the front end of the original inspection task queue according to the abnormality level and triggers path reconstruction.

9. The method according to claim 8, characterized in that In step (5), the execution control module feeds back the execution status and location data to the background management platform in real time during the execution of the task.

10. The method according to any one of claims 7 to 9, characterized in that: In step (6), the communication linkage module uploads the abnormal information and pushes the on-site photos, videos or sensor data snapshots to the background.

Images