Robot inspection collaboration method and system for cable tunnel

Abstract

The invention relates to the technical field of job scheduling and resource management optimization, in particular to a robot inspection cooperation method and system for a cable tunnel. A robot inspection cooperation method for a cable tunnel comprises the following steps: S1, obtaining a key asset point of the cable tunnel, and determining an environmental load index according to the key asset point; s2, acquiring a battery life index, a sensor performance index and a body health index of the robot, and acquiring a comprehensive risk index of the robot according to the battery life index, the sensor performance index and the body health index; s3, determining a robot rotation demand according to the comprehensive risk index; and obtaining a robot inspection path, and obtaining a final inspection rotation marking path according to the rotation demand. According to the invention, by dynamically matching the environment and the robot capability, the safety cooperation path is intelligently planned, and the inspection continuity and efficiency are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of job scheduling and resource management optimization, in particular to a robot inspection coordination method and system for a cable tunnel. BACKGROUND

[0002] As the core transmission channel of urban power grids and energy networks, cable tunnels have a complex internal environment and are subjected to harsh conditions such as high temperature, high humidity, and strong electromagnetic interference for a long time. In recent years, robot inspection technology has been gradually applied to the cable tunnel operation and maintenance field. By carrying multiple types of sensors, environmental data collection and equipment state monitoring are achieved, effectively improving the automation level and coverage of inspection operations. Some systems have the initial ability to coordinate multiple robots for inspection, providing important support for intelligent tunnel operation and maintenance.

[0003] However, existing robot coordination inspection methods still have obvious shortcomings when facing the dynamic changes of the thermal field distribution, electromagnetic environment, and equipment load inside the cable tunnel. Existing technologies focus on static path planning and fixed cycle rotation strategies, lack dynamic coupling analysis of environmental load and robot risk state, and cannot adaptively adjust the predictive tasks according to real-time operation intensity and equipment risk situation. At the same time, existing systems have not established effective electrical isolation and mutual aid rescue mechanisms when dealing with safety interlocking and resource optimization problems in the process of multi-robot coordination, resulting in poor robustness and low coordination efficiency of the system under extreme working conditions.

[0004] Therefore, it is urgent to develop a multi-robot coordination inspection method that can adapt to the dynamic operating environment of a cable tunnel. By integrating environmental situation awareness and robot state evaluation, autonomous optimization of the inspection path and dynamic generation of the coordination strategy are achieved, improving the reliability, safety, and overall efficiency of robot group operations in complex environments. SUMMARY

[0005] In order to overcome the poor dynamic optimization effect of robot coordination inspection in complex environments, the present application provides a robot inspection coordination method and system for a cable tunnel.

[0006] The technical implementation scheme of the present application is: a robot inspection coordination method for a cable tunnel, comprising the following steps: S1: obtaining key asset points of the cable tunnel, and determining an environmental load index according to the key asset points; S2: obtaining battery endurance indicators, sensor performance indicators, and body health indicators of the robot, and obtaining a comprehensive risk index of the robot according to the battery endurance indicators, sensor performance indicators, and body health indicators; S3: determining a robot rotation requirement according to the comprehensive risk index; obtaining a robot inspection path, and obtaining a final inspection rotation marking path according to the rotation requirement; S4: determining an environmental mutation point according to the environmental load index, and obtaining an inspection replacement marking path according to the environmental mutation point; S5: determining a robot inspection coordination method according to the final inspection rotation marking path and the inspection replacement marking path.

[0007] Preferably, the key asset points of the cable tunnel are obtained, and the environmental load index is determined according to the key asset points, comprising: The key asset points in the cable tunnel are identified and labeled based on the engineering characteristics and historical failure data of the cable tunnel; The cable tunnel is divided into N monitoring sections according to adjacent key asset points; The operation period, operation mode and operation intensity data of the cable in each monitoring section are collected; The operation period, operation mode and operation intensity data are normalized to obtain a normalized processing result, and the environmental load index of each monitoring section is calculated according to the normalized processing result, and the environmental load index calculation formula is: Wherein, is the environmental load index, is the normalized operation period, is the normalized operation mode, is the normalized operation intensity value, , and is a weight coefficient, and the monitoring section with an environmental load index higher than a preset threshold is identified as a high-load monitoring section.

[0008] Preferably, the battery endurance index, sensor performance index and body health index of the robot are obtained, and the comprehensive risk index of the robot is obtained according to the battery endurance index, sensor performance index and body health index, comprising: The battery endurance index, sensor performance index and body health index of the robot located in the high-load monitoring section are monitored; According to the battery endurance index, the current power of the robot is obtained in real time, and the battery power drop rate of the current power in the current monitoring section is calculated, which is taken as a battery risk coefficient, and the battery risk coefficient is positively correlated with the battery attenuation risk; According to the sensor performance index, the minimum value in each sensor confidence is obtained, and a sensor risk coefficient is obtained according to the minimum value, ; wherein, is the sensor risk coefficient, For the confidence, the lower the confidence is, the higher the sensor risk is; According to the ontology health index, an ontology risk coefficient is calculated, and the calculation formula is: ontology risk coefficient=(current health index-benchmark health index) / (maximum allowed health index-benchmark health index), wherein the ontology health index includes motor temperature and vibration amplitude, and the current health index, the benchmark health index and the maximum allowed health index are measurement values of the same physical parameter; The maximum value among the battery risk coefficient, the sensor risk coefficient and the ontology risk coefficient is taken as the comprehensive risk index of the robot.

[0009] Preferably, the determination of the robot rotation demand according to the comprehensive risk index comprises: In the current high-load monitoring area, the following judgments are performed: Priority rotation judgment: the load index change rate of the environment load index is calculated; If the battery risk coefficient is greater than a preset descending rate threshold, and at the same time, the load index change rate is greater than a preset load index change rate threshold, a priority rotation demand is triggered immediately, and a high-frequency rotation scheme is started; Regular rotation judgment: if the above priority rotation judgment is not triggered, the product of the comprehensive risk index and the environment load index is calculated as a comprehensive risk load value; when any robot enters the high-load monitoring area, and the comprehensive risk load value is higher than a preset safety threshold, a regular rotation demand is triggered; According to the regular rotation demand, a robot rotation strategy is determined according to the comprehensive risk load value: if the comprehensive risk load value is higher than a first-level risk threshold, a high-frequency rotation scheme is started; if the comprehensive risk load value is higher than a second-level risk threshold but lower than the first-level risk threshold, a medium-frequency rotation scheme is started; if the comprehensive risk load value is lower than the second-level risk threshold, a low-frequency rotation scheme is started; wherein the first-level risk threshold is higher than the second-level risk threshold.

[0010] Preferably, the acquisition of the robot inspection path comprises: According to the rotation demand, a preliminary inspection rotation marked path is obtained by marking in the robot inspection path; The robot single-trip risk capacity is calculated by initializing the inspection rotation marked path, and the robot single-trip risk capacity is the product of a preset standard risk threshold and a reference path load of the robot in a standard environment, wherein the reference path load is the product of the path length covered by the robot in a standard endurance time and a standard environment load index; From the first shift point to the last shift point, multiply the comprehensive risk load value of each monitoring section by the length of the corresponding inspection shift marker path to obtain a section risk value, and accumulate the section risk values to obtain a path cumulative risk value; If the path cumulative risk value reaches the robot single-trip risk capacity, mark the corresponding shift point as a relief point, update the inspection shift marker path according to the relief point, and obtain a final inspection shift marker path.

[0011] Preferably, the determination of the environmental mutation point according to the environmental load index and the obtaining of the inspection relief marker path according to the environmental mutation point comprise: Comparing the absolute value of the difference between the environmental load indexes of adjacent monitoring sections, marking the section boundary point with an absolute value of the difference higher than a preset mutation threshold as an environmental mutation point; According to the environmental mutation point, mark in the robot inspection path to obtain an inspection relief marker path.

[0012] Preferably, the determination of the robot inspection coordination method according to the final inspection shift marker path and the inspection relief marker path comprises: Fusing the final inspection shift marker path and the inspection relief marker path to generate an individual inspection path for each robot, the individual inspection path containing all shift points and relief points; Obtaining a maximum allowed distance of maintenance complementary constraint based on the maximum moving speed of the robot and the expected response time; Based on the cable voltage level, determining a basic safety distance according to the electrical safety standard, and multiplying the safety factor determined based on the robot shell material to obtain a minimum safety distance of blind coordination constraint; Based on the individual inspection path, calculating the Euclidean distance of any two robots at any time, and ensuring that the following constraints are met through a path adjustment algorithm: For all robot pairs and all times, the Euclidean distance is less than or equal to the maximum allowed distance; For all regions in high-load monitoring sections and all times, the Euclidean distance is greater than or equal to the minimum safety distance; Output the robot path set, speed scheduling scheme and task timing arrangement that meet the above constraints to form the final robot inspection coordination method.

[0013] Preferably, the determination of the maximum allowed distance of maintenance complementary constraint based on the maximum moving speed of the robot and the expected response time comprises: the maximum allowed distance calculation formula is: ; wherein, is the maximum allowable distance, is the maximum moving speed, is the expected response time.

[0014] Preferably, the minimum safety distance of the blind cooperative constraint is obtained by multiplying the basic safety distance determined according to the electrical safety standard based on the cable voltage level by a safety factor determined based on the material of the robot shell, including: the minimum safety distance calculation formula is: ; wherein, is the minimum safety distance, is the safety factor, is the basic safety distance.

[0015] A robot inspection cooperative system for a cable tunnel, comprising: An environmental situation awareness module for acquiring cable tunnel engineering features and historical failure data, identifying key asset points and dividing monitoring sections, integrating operating parameter calculation to obtain an environmental load index and its change rate, and identifying high load monitoring areas; A robot state monitoring module for real-time monitoring of battery endurance indicators, sensor performance data and body health indicators of robots in high load areas, determining a comprehensive risk index through risk coefficient calculation and maximum value screening; A cooperative path planning module for priority rotation judgment based on the change rate of the environmental load index and the battery risk coefficient, and regular rotation judgment based on the comprehensive risk index and the environmental load index to calculate a comprehensive risk load value; according to the rotation requirement, combined with the calculation of the single-trip risk capacity of the robot and the cumulative risk value of the path, a final inspection rotation marking path is generated; A safety cooperative optimization module for calculating the maintenance complementary distance and the electrical safety distance between robots, ensuring that the double constraints are met through a path adjustment algorithm, and outputting a cooperative inspection path, a speed scheduling scheme and a task timing arrangement.

[0016] Beneficial effects: The application realizes the accurate matching of the inspection task and the dynamic environmental condition by collecting the operation parameters of the key asset points of the cable tunnel, dynamically calculating the environmental load index of each monitoring section, and evaluating the comprehensive tolerance capacity by fusing the robot battery, sensor and ontology health index. The grading rotation mechanism is triggered based on the real-time product of the risk index and the environmental load, and the priority judgment strategy based on the rate change is introduced to automatically generate the marked path of the fusion rotation point and the replacement point. Further, by calculating the maximum allowed distance of complementary maintenance and the minimum safety distance of electrical safety, the path adjustment algorithm is used to ensure that the dual constraints of mutual assistance and safety isolation are always met when multiple robots work collaboratively. This method effectively overcomes the shortcomings of traditional static path planning and fixed rotation strategy in dealing with complex electromagnetic and thermal environments of the cable tunnel, significantly improving the inspection continuity, operation safety and overall resource utilization efficiency of the system under extreme working conditions. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 A flow chart of the robot inspection collaboration method for the cable tunnel of the application; Figure 2 A structural diagram of the robot inspection collaboration system for the cable tunnel of the application. DETAILED DESCRIPTION

[0018] The application will be further described below in conjunction with specific embodiments, and the illustrative embodiments and descriptions of the application are used to explain the application, but not as a limitation of the application.

[0019] Embodiment 1: A robot inspection collaboration method for a cable tunnel, as shown in Figure 1 , comprising the following steps: S1: Acquiring key asset points of the cable tunnel, and determining an environmental load index according to the key asset points; Identifying and labeling the key asset points in the cable tunnel based on the engineering characteristics and historical failure data of the cable tunnel; Dividing the cable tunnel into N monitoring sections according to adjacent key asset points; Collecting the operation period, operation mode and operation intensity data of the cable in each monitoring section; Normalizing the operation period, operation mode and operation intensity data to obtain a normalization result, and calculating the environmental load index of each monitoring section according to the normalization result, the environmental load index calculation formula being: , wherein, is the environmental load index, is the normalized operation period, is the normalized operation mode, is the normalized operation intensity value, , and As a weight coefficient, the monitoring section with an environmental load index higher than a preset threshold value is identified as a high-load monitoring section.

[0020] It should be noted that the application first identifies and labels the key asset points in the cable tunnel by integrating the engineering features of the cable tunnel (design drawings, tunnel topology, asset layout) and historical failure data (failure records in operation and maintenance reports, preventive test data). These points (such as cable joints, insulation terminals, and lightning arresters) are the core equipment in the tunnel with the highest failure probability and the greatest impact on operation safety based on historical data analysis. Subsequently, the cable tunnel is divided into N monitoring sections according to the positions of adjacent key asset points. Each section takes two adjacent key asset points as the starting and ending points, forming independent monitoring units covering the entire tunnel and taking risk points as the framework, thereby achieving fine segmentation of the tunnel risk area.

[0021] The cable operation state data of each monitoring section is collected in real time through the sensor network deployed in the tunnel and integrated with the power grid monitoring system (such as SCADA). The operation period reflects the peak or seasonal power supply strategy at the system level (such as summer peak and night light load), the operation mode distinguishes whether the cable is in full load, overload or light load typical working condition, and the operation intensity quantifies the actual working pressure through the current load rate of the cable in the section. After normalization, the three types of parameters are fused into an environmental load index using the square sum and square root algorithm, which comprehensively represents the comprehensive stress degree of the robot equipment in a specific section. The preset threshold value of the environmental load index is determined based on the failure statistics quantile related to environmental factors in the historical operation and maintenance data of the cable tunnel, and the section with a load index exceeding the limit is identified as a high-load monitoring section, laying the foundation for precise scheduling of robot resources.

[0022] Compared with the traditional static partitioning method, this step solves the defect of the prior art that is not sensitive to environmental changes through multi-source data fusion and dynamic load evaluation, providing data support for the subsequent adaptive inspection strategy.

[0023] S2: Obtain the battery endurance index, sensor performance index, and body health index of the robot, and obtain the comprehensive risk index of the robot according to the battery endurance index, sensor performance index, and body health index; Monitor the battery endurance index, sensor performance index, and body health index of the robot located in the high-load monitoring section; According to the battery endurance index, the current power of the robot is obtained in real time, and the battery power drop rate of the current power in the current monitoring section is calculated, and the battery power drop rate is taken as the battery risk coefficient, which is positively correlated with the battery attenuation risk; According to the sensor performance index, a minimum value in each sensor confidence is obtained, and a sensor risk coefficient is obtained according to the minimum value, ; wherein, is a sensor risk coefficient, is a confidence, and a lower confidence indicates a higher sensor risk; According to the body health index, a body risk coefficient is calculated, and the calculation formula is: body risk coefficient = (current health index - baseline health index) / (maximum allowed health index - baseline health index), wherein the body health index includes motor temperature and vibration amplitude, and the current health index, baseline health index and maximum allowed health index are measurement values of the same physical parameter; The maximum value among the battery risk coefficient, sensor risk coefficient and body risk coefficient is taken as the comprehensive risk index of the robot.

[0024] It should be noted that this step constructs a robot dynamic risk assessment system through multi-source sensor data fusion and bottleneck decision mechanism. Specifically: the battery endurance index collects real-time voltage and current data through the power management chip, calculates the battery power consumption rate based on the power consumption per unit time, and this rate directly reflects the energy consumption intensity of the robot in the current harsh environment. The sensor performance index relies on the self-diagnosis module of each perception unit, for example, the vision system calculates the confidence through the image signal-to-noise ratio, and the laser radar evaluates the reliability through the point cloud integrity. The body health index is monitored through embedded sensors (temperature and vibration sensors) to monitor the key physical parameters of the core components. The higher the body risk coefficient, the greater the risk of the body state; At the data processing level: a maximum value aggregation strategy is used to determine the comprehensive risk index, which is derived from the system risk theory, that is, the overall risk level of a system is determined by the riskiest link. For example, assume that after state evaluation of the robot R1, the risk coefficients of the three aspects are calculated as follows: battery risk coefficient: 0.15 (energy consumption rate is moderate), sensor risk coefficient: 0.10 (sensor working state is good), and body risk coefficient: 0.65 (the driving motor is close to the safety upper limit due to high temperature environment), then the comprehensive risk index of the robot R1 is: max(0.15, 0.10, 0.65) = 0.65. This maximum value aggregation strategy ensures that the system can focus on the most dangerous state factor.

[0025] Compared with the traditional single threshold early warning mechanism, the method reveals the real risk boundary of the robot in a high temperature and high humidity environment through multi-dimensional risk coupling evaluation. For example, when the temperature of the robot drive motor rises to the limit value of the highest allowable health indicator, its body risk coefficient will reach 1, thus becoming the dominant risk and triggering the system protection mechanism immediately. This fine risk assessment method based on physical constraints effectively solves the one-sided problem of existing technology in evaluating the state of the robot, and provides accurate and reliable risk quantification basis for the dynamic allocation of inspection resources.

[0026] S3-1: Determine the robot rotation requirement according to the comprehensive risk index; In the current high-load monitoring area, the following judgments are performed: Priority rotation judgment: calculate the load index change rate of the environment load index; If the battery risk coefficient is greater than the preset descending rate threshold, and at the same time the load index change rate is greater than the preset load index change rate threshold, the priority rotation requirement is triggered immediately, and the high-frequency rotation scheme is started; Regular rotation judgment: if the above priority rotation judgment is not triggered, calculate the product of the comprehensive risk index and the environment load index as the comprehensive risk load value; when any robot enters the high-load monitoring area, and the comprehensive risk load value is higher than the preset safety threshold, the regular rotation requirement is triggered; According to the regular rotation requirement, determine the robot rotation strategy according to the comprehensive risk load value: if the comprehensive risk load value is higher than the first level risk threshold, start the high-frequency rotation scheme; if the comprehensive risk load value is higher than the second level risk threshold but lower than the first level risk threshold, start the medium-frequency rotation scheme; if the comprehensive risk load value is lower than the second level risk threshold, start the low-frequency rotation scheme; wherein the first level risk threshold is higher than the second level risk threshold.

[0027] It should be noted that in the cable tunnel inspection system, the robot inspection path is a fixed trajectory set in advance according to the actual physical structure of the tunnel (such as the trend, slope, and branch) and the accurate positions of key asset points (such as cable joints and insulation terminals) through path planning algorithms (such as the optimal coverage path generated after SLAM mapping). This path ensures that all key equipment can be effectively covered. Taking a basic path containing straight sections, curved sections, and high-voltage sections as an example, first, according to the real-time rotation requirements, mark the rotation points that need to be temporarily rested in the path to form a preliminary inspection rotation marking path. The preset descending rate threshold and the preset load index change rate threshold are obtained based on clustering analysis of the robot battery energy consumption mode and environmental mutation characteristics in high-load working conditions in historical operation and maintenance data. In the high-load section, if the comprehensive risk load value continues to be higher than the first-level risk threshold, a high-frequency rotation scheme is started, for example, the robot is handed over at the rotation point every 30 minutes to respond to the sharply rising operation risk; if the risk load value is moderate (higher than the second-level threshold but not exceeding the first-level), a medium-frequency rotation scheme is started, such as rotating every 2 hours to balance safety and efficiency; and in the conventional inspection section with low risk load value, a low-frequency rotation scheme is started, such as rotating every 4 hours to optimize resource utilization efficiency.

[0028] By initializing the marked path, the preset standard risk threshold is multiplied by the reference path load to obtain the single-trip risk capacity, which represents the upper limit of the total risk that the robot can safely withstand in a single task. Starting from the first rotation point, the product of the comprehensive risk load value and the path length of each monitoring section is calculated to obtain the section risk value reflecting the actual risk accumulation of the section. By accumulating these risk values, the path cumulative risk value is obtained, and when the cumulative amount reaches the single-trip risk capacity, the current rotation point is upgraded to a replacement point.

[0029] In the specific implementation process, the generation position of the replacement point is dynamically adjusted by real-time monitoring of the actual risk accumulation of each inspection section. This risk accumulation-based early warning mechanism ensures that the robot always works within the safety risk capacity, avoiding both resource waste caused by premature rotation and operation safety hazards caused by excessive risk accumulation. The replacement point refers to the handover position where the robot needs to be replaced to continue the subsequent task.

[0030] For example, assuming that a certain inspection path contains three high-load sections, the initial rotation marking path sets rotation points at the end of each section. When it is detected that the path cumulative risk value calculated at the end of the second section reaches the single-trip risk capacity, the rotation point is automatically upgraded to a replacement point, forming a final inspection rotation marking path containing both rotation points and replacement points. This dynamic marking mechanism based on risk accumulation effectively solves the problem of uneven task allocation in traditional fixed path planning, achieving precise scheduling of multi-robot collaborative operation.

[0031] S3-2: Obtain a robot inspection path, and obtain a final inspection rotation marking path according to the rotation requirement; Mark in the robot inspection path according to the rotation requirement, and obtain a preliminary inspection rotation marking path; Initialize the inspection rotation marking path to calculate a robot single-trip risk capacity, which is a product of a preset standard risk threshold and a reference path load of the robot in a standard environment, wherein the reference path load is a product of a path length covered by the robot in a standard endurance time and a standard environment load index; Multiply a comprehensive risk load value of each monitoring section by a corresponding inspection rotation marking path length from a first rotation point to a last rotation point to obtain a section risk value, and accumulate the section risk values to obtain a path cumulative risk value; If the path cumulative risk value reaches the robot single-trip risk capacity, mark a corresponding rotation point as a relief point, update the inspection rotation marking path according to the relief point, and obtain a final inspection rotation marking path.

[0032] It should be noted that in the cable tunnel inspection system, the robot inspection path is a fixed trajectory preset according to the actual physical structure (such as the trend, slope, and branch) of the tunnel and the accurate positions of the key asset points (such as cable joints and insulation terminals) through a path planning algorithm (such as an optimal coverage path generated after SLAM mapping). The path ensures that all key equipment can be effectively covered. The standard environment load index is a benchmark constant used for path planning calculation. The standard environment load index is not a real-time measurement value, but is calculated based on long-term historical operation data (such as light load operation period and well-ventilated environmental conditions) of the cable tunnel under typical and moderate working conditions. The calculation method is the same as that of the real-time environment load index, but the operation period, mode, intensity, and other parameters used are preset standard values representing ideal inspection conditions. The index is used to quantify the baseline path load of the robot under no additional environmental pressure, providing a unified baseline for calculating the “robot single-trip risk capacity”.

[0033] Taking a basic path containing straight segments, curved sections, and high-voltage areas as an example, first, mark rotation points in the path according to real-time triggered rotation requirements to form a preliminary inspection rotation marking path. A rotation point is defined as a planned position where the robot needs to hand over tasks due to risk reaching a threshold.

[0034] Initialize the marked path to calculate the robot single-trip risk capacity, which is a product of a preset standard risk threshold and a reference path load of the robot in a standard environment, representing the upper limit of the total risk that the robot can safely withstand in a single task.

[0035] From the first rotation point, multiply the comprehensive risk load value of each monitoring section by the corresponding inspection rotation marker path length to obtain a section risk value, and accumulate the section risk values to obtain a path cumulative risk value.

[0036] When the path cumulative risk value reaches the single-trip risk capacity of the robot, the corresponding rotation point is marked as a handover point. The single-trip risk capacity is set based on analysis of the performance degradation law of the robot in historical operation data. Specifically, by continuously monitoring and recording the evolution of the comprehensive risk index and the environmental load index of the robot during task execution, and using a trend analysis algorithm to identify the inflection point at which the risk begins to accelerate, a reliable risk boundary is determined at the data level by correlating the identified risk inflection points in a large number of tasks with cases that ultimately lead to task interruption or failure. For example, analysis shows that when the total amount of risk accumulated by the robot reaches a certain critical value, performance inflection points generally begin to appear, so this critical value is set as the single-trip risk capacity. When the path cumulative risk value of the robot in a new task reaches this capacity, a handover point is marked at the corresponding location.

[0037] This risk accumulation-based early warning mechanism ensures that the robot always works within the safety risk boundary, avoiding both the waste of resources caused by premature rotation and the operational safety hazards caused by excessive risk accumulation. In specific implementation, by monitoring the path cumulative risk value of each inspection section in real time, the generation position of the handover point is dynamically adjusted to achieve an optimal balance in task handover between multiple robots. The handover point refers to the handover location where the robot needs to be replaced to continue executing subsequent tasks.

[0038] Example: Suppose a certain inspection path contains three high-load sections, and the initial rotation marker path sets rotation points at the end of each section. When it is detected that the path cumulative risk value calculated at the end of the second section reaches the single-trip risk capacity of the robot, the rotation point is automatically upgraded to a handover point, forming a final inspection rotation marker path that contains both rotation points and handover points. This risk-based dynamic accumulation-based marking mechanism effectively solves the problem of uneven task allocation in traditional fixed path planning, achieving precise scheduling of multi-robot collaborative work.

[0039] S4: determining an environmental mutation point according to the environmental load index, and obtaining an inspection handover marker path according to the environmental mutation point; Comparing the absolute value of the difference between the environmental load indices of adjacent monitoring sections, marking the section junction point whose absolute value of the difference is higher than a preset mutation threshold as an environmental mutation point; According to the environmental mutation point, marking in the robot inspection path to obtain an inspection handover marker path.

[0040] It should be noted that in the cable tunnel dynamic inspection system, the identification of environmental mutation points is of key significance to ensure the continuous operation of the robot. The environmental mutation point is defined as the boundary position where the environmental load index of adjacent monitoring sections has a significant transition. By calculating the absolute value of the difference between the environmental load indexes of adjacent sections, the dramatic change of the environmental conditions in the tunnel can be captured, such as the sudden temperature rise caused by the sudden entry from the conventional heat dissipation area to the cable dense area. The preset mutation threshold is set according to the historical average fluctuation amplitude of the environmental load index of adjacent monitoring sections.

[0041] For example, assuming that a tunnel contains three consecutive monitoring sections with environmental load indexes of 0.3, 0.8 and 0.5 respectively. By calculating the absolute value of the difference between adjacent sections, it is found that the difference between the second and first sections is 0.5, which exceeds the preset threshold, and this boundary point is marked as an environmental mutation point. The inspection handover marking path generated based on these mutation points is complementary to the rotation marking path based on the state of the robot: the former is aimed at the operation risk caused by environmental mutation, and the latter is aimed at the problem of performance degradation of the robot. This dual-path coordination mechanism effectively overcomes the defect of the traditional single planning that the response to environmental mutation is lagging, and by setting the handover position in advance at the environmental critical point, it ensures the continuity and safety of the inspection task in complex tunnel environment.

[0042] S5: determining a robot inspection coordination method according to the final inspection rotation marking path and the inspection handover marking path.

[0043] Fusing the final inspection rotation marking path and the inspection handover marking path generates an individual inspection path for each robot, which includes all rotation points and handover points; Based on the maximum moving speed of the robot and the expected response time, the maximum allowed distance of the maintenance complementary constraint is obtained; The maximum allowed distance calculation formula is: ; Wherein, is the maximum allowed distance, is the maximum moving speed, is the expected response time.

[0044] Based on the cable voltage level, the basic safety distance is determined according to the electrical safety standard, and multiplied by the safety factor determined based on the robot shell material to obtain the minimum safety distance of the blind coordination constraint; The minimum safety distance calculation formula is: ; Wherein, is the minimum safety distance, is the safety factor, is the basic safety distance.

[0045] Based on the individual inspection path, the Euclidean distance between any two robots at any time is calculated, and the following constraints are ensured by the path adjustment algorithm: For all robot pairs and all times, the Euclidean distance is less than or equal to the maximum allowed distance; For all regions in the high-load monitoring area and all times when the electromagnetic load level exceeds the safety threshold, the Euclidean distance between all robot pairs in this region is greater than or equal to the minimum safety distance; The robot path set, speed scheduling scheme and task timing arrangement that satisfy the above constraints are output, forming the final robot inspection coordination method.

[0046] It should be noted that in the cable tunnel multi-robot inspection system, the path fusion process uses space-time coding technology to organically integrate the rotation path based on the performance state of the robot and the replacement path based on environmental risk. Specifically, by establishing a unified spatial coordinate system, the two types of path markers are rearranged according to the actual position and operation timing. For example, in a typical inspection path containing straight-line segments, curved areas and high-voltage areas, the performance rotation points and environmental replacement points are combined to generate an enhanced inspection path that contains both cooling maintenance points and task handover points. This individual inspection path not only specifies the basic travel route of the robot, but also marks all necessary work adjustment positions, providing accurate navigation for multi-robot cooperative work.

[0047] The calculation of the maximum allowed distance under the maintenance complementary constraint is based on the motion performance of the robot and the system response requirement. The maximum moving speed is the safety running limit of the robot drive system in the tunnel environment, and the expected response time includes the sum of signal transmission delay, decision processing time and mechanical action time. The setting of this distance parameter ensures that the robot group forms an effective mutual assistance network: when a robot fails, the nearby support robot can reach the scene within a specified time. The maximum allowed distance defines the maximum spatial interval between mutual assistance robot pairs to maintain effective rescue capability.

[0048] The minimum safety distance calculation of blind coordination constraint strictly follows the high-voltage electrical safety specification. Among them, the cable voltage grade refers to the rated working voltage of the power transmission line, the electrical safety standard refers to the relevant provisions of the International Electrotechnical Commission (IEC), and the basic safety distance directly corresponds to the minimum air insulation gap under a specific voltage grade. In order to cope with the uncertainty brought by the robot body, a safety factor k based on the shell material is introduced: for example, for a conductive metal shell, take k=1.5 to fully cope with the risk of electric field distortion and induced charge; for a non-metal shell with good insulation performance, take k=1.0, that is, directly use the basic safety distance. Blind coordination refers to the special case where the communication is interrupted due to strong electromagnetic interference, and the robot relies on pre-loaded instructions and local sensors to achieve limited collaborative work. The core function of the minimum safety distance is to prevent the risk of arc discharge between robots due to potential difference. The safety threshold of the electromagnetic load is set according to the electrical safety standard corresponding to the cable voltage grade.

[0049] The path adjustment algorithm is a closed-loop optimization process consisting of four core steps: conflict prediction, speed adjustment, path re-planning, and result verification. First, in the conflict prediction phase, a spatiotemporal motion trajectory model is established for each robot. Specifically, based on the pre-set path and the regular inspection speed, the expected travel route of each robot is depicted in a three-dimensional coordinate system composed of two-dimensional space and time. This continuous trajectory is discretized into multiple spatiotemporal units and marked with occupancy status, thereby constructing a complete spatiotemporal occupancy grid. By traversing the position relationships of all robots at various time points, two types of potential conflicts are identified: one is when the actual distance between any two robots exceeds the maximum mutual assistance distance, marked as "exceeding mutual assistance range"; the other is when the distance between robots in a high-load area is less than the minimum safety distance, marked as "insufficient safety distance." After entering the speed adjustment phase, robots are assigned processing priorities based on task importance. In descending order, the spatiotemporal trajectory of the robots is changed by inserting waiting time or adjusting the travel speed in certain path segments. This priority-based speed regulation can effectively resolve most spatiotemporal conflicts while maintaining the original route largely unchanged. When speed adjustment alone cannot resolve complex conflicts, path re-planning is initiated. For example, in a narrow tunnel, a local area is defined centered on the conflict point, and other affected robots are treated as dynamic obstacles. A real-time motion planning algorithm is used to generate a local detour path that meets the safety distance requirement. The new path must ensure no collision with the tunnel structure and other robots, while also complying with the established distance constraints. Finally, convergence verification and iterative optimization are performed. After each round of adjustment, conflict detection is performed again until one of the following conditions is met: one is that all conflicts have been eliminated, indicating successful planning; the other is that conflicts still exist after reaching the preset iteration limit, at which point the system will initiate an emergency mechanism. The successful verification scheme is finally output, including a detailed conflict-free path set, precise speed control instructions for each segment, and a complete task timing arrangement. The electromagnetic load safety threshold is determined based on the cable insulation level and partial discharge characteristics. When the environmental electromagnetic intensity exceeds this threshold, strict safety distance control is automatically enabled.

[0050] Taking the inspection task in the extra-high voltage section as an example, by dynamically adjusting the relative positions and travel speeds of the two robots, the minimum safety distance to prevent arc discharge is maintained, and the maximum allowed distance for timely mutual assistance is ensured, achieving the best balance between safety protection and operational efficiency. This intelligent coordination mechanism effectively solves the technical problem of balancing safety requirements and operational efficiency in traditional inspection systems, significantly improving the system reliability and operational continuity in complex environments.

[0051] Example: During the cable tunnel inspection process, the environmental load index and the comprehensive risk index of each monitoring section are continuously monitored. For example, when robot R1 enters a high-load monitoring area, its comprehensive risk load value continuously increases due to the high-temperature environment. The value is calculated in real time, and when it is lower than the first-level rotation threshold, the robot rotation requirement is automatically triggered. The value is calculated in real time, and when it is higher than the first-level risk threshold, the robot rotation requirement is automatically triggered. According to this requirement, a pre-marked rotation point is dynamically updated as a replacement point on the preset robot inspection path, combined with the calculated path cumulative risk value, to generate a final inspection rotation marking path. At the same time, the standby robot R2 is dispatched to the vicinity of the environmental mutation point according to the inspection replacement marking path. During the task handover, through the path adjustment algorithm, the Euclidean distance between R1 and R2 is ensured to always meet the double constraints: not exceeding the maximum allowed distance calculated based on the maximum moving speed and expected response time to ensure mutual assistance; and not less than the minimum safety distance determined based on the cable voltage level and shell material to prevent electrical hazards. Finally, the speed scheduling scheme and task timing arrangement of this rotation are output, realizing the dynamic, safe and efficient coordination of the physical resources and inspection workflow of the robot.

[0052] Example 2: Based on example 1, a robot inspection coordination system for a cable tunnel, as shown in Figure 2 , comprises: An environmental situation awareness module for obtaining cable tunnel engineering features and historical failure data, identifying key asset points and dividing monitoring sections, integrating operating parameter calculation environmental load index and its change rate, and identifying high-load monitoring areas; A robot state monitoring module for real-time monitoring of battery endurance indicators, sensor performance data and ontology health indicators of robots in high-load areas, determining a comprehensive risk index through risk coefficient calculation and maximum value screening; A collaborative path planning module for prioritizing rotation based on environmental load index change rate and battery risk coefficient, and for regular rotation based on comprehensive risk index and environmental load index to calculate comprehensive risk load value; according to the rotation requirement, combined with the calculation of robot one-way risk capacity and path cumulative risk value, a final inspection rotation marking path is generated; A safety coordination optimization module for calculating the maintenance complementary distance and electrical safety distance between robots, ensuring that the double constraints are met through a path adjustment algorithm, and outputting the collaborative inspection path, speed scheduling scheme and task timing arrangement.

[0053] It should be understood that the above description is for exemplary purposes only and is not intended to limit the present application. Those skilled in the art will understand that variations of the present application will be included within the scope of the claims herein.

Claims

1. A collaborative robot inspection method for cable tunnels, characterized in that, Includes the following steps: S1: Obtain the key asset points of the cable tunnel and determine the environmental load index based on the key asset points; S2: Obtain the robot's battery life index, sensor performance index, and robot health index, and obtain the robot's comprehensive risk index based on the battery life index, sensor performance index, and robot health index. S3: Determine the robot rotation requirements based on the comprehensive risk index; obtain the robot inspection path; and obtain the final inspection rotation marking path based on the rotation requirements. S4: Determine the environmental mutation point based on the environmental load index, and obtain the inspection replacement marking path based on the environmental mutation point; S5: Determine the robot inspection collaboration method based on the final inspection rotation marking path and the inspection succession marking path.

2. The robotic collaborative inspection method for cable tunnels according to claim 1, characterized in that, The acquisition of key asset points for cable tunnels, and the determination of environmental load indices based on these key asset points, include: Identify and label key asset points within cable tunnels based on their engineering characteristics and historical fault data. The cable tunnel is divided into N monitoring sections based on the adjacent critical asset points; Collect data on the operating time, operating mode, and operating intensity of cables within each monitoring section; The data on the operating period, operating mode, and operating intensity are normalized to obtain normalization results. Based on these results, the environmental load index for each monitoring segment is calculated. The formula for calculating the environmental load index is: ,in, This is the environmental load index. For normalized runtime, For normalized operation mode, To normalize the operational intensity value, , and Using the weighting coefficient, monitoring sections with environmental load indices exceeding a preset threshold are identified as high-load monitoring areas.

3. The robotic collaborative inspection method for cable tunnels according to claim 2, characterized in that, The process of acquiring the robot's battery life indicators, sensor performance indicators, and robot health indicators, and obtaining the robot's comprehensive risk index based on these indicators, includes: Monitor the robot's battery life, sensor performance, and overall health indicators within the high-load monitoring area; Based on the battery endurance index, the robot's current battery level is obtained in real time, and the rate of battery power decline within the current monitoring segment is calculated. The rate of battery power decline is used as the battery risk coefficient, which is positively correlated with the risk of battery degradation. The minimum confidence level of each sensor is obtained based on the sensor performance indicators, and the sensor risk coefficient is obtained based on the minimum value. ;in, For sensor risk coefficient, The confidence level is the highest level of confidence, with lower confidence levels indicating higher sensor risk. Based on the aforementioned physical health indicators, the physical risk coefficient is calculated using the following formula: Physical risk coefficient = (Current health indicator - Baseline health indicator) / (Maximum permissible health indicator - Baseline health indicator). The physical health indicators include motor temperature and vibration amplitude, and the current health indicator, the baseline health indicator, and the maximum permissible health indicator are measurements of the same physical parameter. The maximum value among the battery risk coefficient, sensor risk coefficient, and body risk coefficient is taken as the robot's comprehensive risk index.

4. A collaborative robot inspection method for cable tunnels according to claim 2, characterized in that, The process of determining robot rotation requirements based on the comprehensive risk index includes: Within the current high-load monitoring area, the following judgments are made: Priority rotation determination: Calculate the load index change rate of the environmental load index; If the battery risk coefficient is greater than the preset rate of decline threshold, and at the same time the load index change rate is greater than the preset load index change rate threshold, then the priority rotation requirement is immediately triggered and the high-frequency rotation scheme is initiated. Regular rotation judgment: If the above priority rotation judgment is not triggered, the product of the comprehensive risk index and the environmental load index is calculated as the comprehensive risk load value; when any robot enters the high load monitoring area and the comprehensive risk load value is higher than the preset safety threshold, the regular rotation requirement is triggered. Based on the aforementioned regular rotation requirements, a robot rotation strategy is determined according to the comprehensive risk load value: if the comprehensive risk load value is higher than the first-level risk threshold, a high-frequency rotation scheme is initiated; if the comprehensive risk load value is higher than the second-level risk threshold but lower than the first-level risk threshold, a medium-frequency rotation scheme is initiated; if the comprehensive risk load value is lower than the second-level risk threshold, a low-frequency rotation scheme is initiated; wherein, the first-level risk threshold is higher than the second-level risk threshold.

5. A collaborative robot inspection method for cable tunnels according to claim 2, characterized in that, The process of obtaining the robot inspection path and obtaining the final inspection rotation marking path based on the rotation requirements includes: According to the rotation requirements, the robot inspection path is marked to obtain a preliminary inspection rotation marking path; By initializing the inspection rotation marking path, the robot's single-trip risk capacity is calculated. The robot's single-trip risk capacity is the product of a preset standard risk threshold and the robot's reference path load in a standard environment. The reference path load is the product of the path length that the robot can cover within the standard endurance time and the standard environmental load index. From the first rotation point to the last rotation point, the comprehensive risk load value of each monitoring segment is multiplied by the corresponding inspection rotation mark path length to obtain the segment risk value. The segment risk values ​​are then summed to obtain the path cumulative risk value. If the cumulative risk value of the path reaches the single-trip risk capacity of the robot, the corresponding rotation point is marked as a replacement point, and the inspection rotation marking path is updated according to the replacement point to obtain the final inspection rotation marking path.

6. A collaborative robot inspection method for cable tunnels according to claim 2, characterized in that, The step of determining environmental abrupt change points based on the environmental load index and obtaining inspection succession marking paths based on the environmental abrupt change points includes: Compare the absolute values ​​of the differences in the environmental load indices of adjacent monitoring sections, and mark the boundary points of sections where the absolute values ​​of the differences are higher than a preset abrupt change threshold as environmental abrupt change points; Based on the environmental mutation points, a replacement marking path is obtained by marking the robot inspection path.

7. A collaborative robot inspection method for cable tunnels according to claim 1, characterized in that, The method for determining robot inspection collaboration based on the final inspection rotation marker path and the inspection succession marker path includes: The final inspection rotation marker path and the inspection replacement marker path are merged to generate an individual inspection path for each robot, which includes all rotation points and replacement points. The maximum allowable distance for maintaining complementary constraints is determined based on the robot's maximum moving speed and expected response time. Based on the cable voltage level, the basic safety distance is determined according to electrical safety standards, and then multiplied by a safety factor determined based on the robot shell material to obtain the minimum safety distance for blind cooperative constraints; Based on the individual inspection path, calculate the Euclidean distance between any two robots at any time, and ensure that the following constraints are met through a path adjustment algorithm: For all robot pairs and at all times, the Euclidean distance is less than or equal to the maximum allowable distance; For all areas in high-load monitoring zones where the electromagnetic load level exceeds the safety threshold, and for all robots operating within these zones at all times, the Euclidean distance must be greater than or equal to the minimum safe distance. The system outputs a set of robot paths, a speed scheduling scheme, and a task timing arrangement that satisfy the above constraints, forming the final robot inspection collaboration method.

8. A collaborative robot inspection method for cable tunnels according to claim 7, characterized in that, The method for obtaining the maximum permissible distance for maintaining complementary constraints based on the robot's maximum moving speed and expected response time includes: the formula for calculating the maximum permissible distance is as follows: ； in, For the maximum allowable distance, Maximum movement speed, This represents the expected response time.

9. A collaborative robot inspection method for cable tunnels according to claim 7, characterized in that, The process of determining the basic safety distance based on cable voltage level and electrical safety standards, and then multiplying it by a safety factor determined based on the robot's shell material to obtain the minimum safety distance for blind cooperative constraints, includes: the minimum safety distance calculation formula is as follows: ； in, For minimum safe distance, For safety reasons, Based on the basic safe distance.

10. A robotic inspection collaborative system for cable tunnels, used to implement the robotic inspection collaborative method for cable tunnels as described in any one of claims 1-9, characterized in that, include: The environmental situation awareness module is used to acquire cable tunnel engineering characteristics and historical fault data, identify key asset points and divide monitoring sections, integrate operating parameters to calculate the environmental load index and its rate of change, and identify high load monitoring areas. The robot status monitoring module monitors the robot's battery life, sensor performance data, and body health indicators in high-load areas in real time, and determines the comprehensive risk index through risk coefficient calculation and maximum value screening. The collaborative path planning module makes priority rotation judgments based on the environmental load index change rate and battery risk coefficient, and makes regular rotation judgments based on the comprehensive risk index and environmental load index to calculate the comprehensive risk load value; according to the rotation requirements, combined with the calculation of the robot's single-trip risk capacity and the cumulative risk value of the path, the final inspection rotation marking path is generated. The safety collaboration optimization module calculates the maintenance complementarity distance and electrical safety distance between robots, and ensures that the dual constraints are met through path adjustment algorithms, outputting collaborative inspection paths, speed scheduling schemes and task timing arrangements.

Images