Multi-device collaborative inspection safety conflict detection method and system

Abstract

The invention discloses a multi-device cooperative inspection safety conflict detection method and system, and belongs to the technical field of intelligent inspection and robot control. According to the method, a UWB positioning base station and a wired interconnected WIFI networking router are deployed in a tunnel, high-precision three-dimensional positioning is carried out on a robot dog, an unmanned aerial vehicle and intelligent wearable equipment which carry UWB tags and WIFI modules, and position and state information is returned in real time; and the ground station terminal dynamically calculates a dynamic safety distance between any two pieces of equipment based on equipment motion parameters and personnel operation states, evaluates a conflict risk probability in combination with an actual distance, and triggers corresponding graded early warning and control measures according to risk grades. According to the invention, high-precision cooperation and intelligent safety conflict prevention and control of multiple types of inspection main bodies in a closed environment are realized, the safety and reliability of inspection operation are remarkably improved, and the method is suitable for complex scenes such as hydropower station tunnels, urban underground pipe galleries and mine roadways.

Description

Technical Field

[0001] This invention belongs to the field of power plant inspection technology, and specifically relates to a method and system for detecting safety conflicts during multi-equipment collaborative inspection. Background Technology

[0002] With the rapid development of infrastructure construction such as water conservancy and transportation, the length and number of tunnels are increasing year by year, making their daily inspection and maintenance crucial for ensuring project safety. Traditional manual inspection suffers from low efficiency and high risk; therefore, collaborative inspection using multiple devices such as robot dogs, drones, and smart wearable devices is gradually becoming an effective automated inspection solution. However, several core problems still exist in the multi-device collaborative inspection process, which are the technical problems that this invention aims to solve: Problem 1: Insufficient dynamic positioning accuracy of multiple devices. Complex signal blockage and transmission interference in tunnels can easily lead to the failure of airborne sensors. In addition, the errors of positioning technologies such as GPS denial, traditional Bluetooth, and RFID generally exceed 1 meter. However, the safe navigation of robot dogs, drones and inspection personnel in collaborative operations requires maintaining a close safety distance of 0.5 meters. The existing positioning accuracy cannot meet this requirement.

[0003] Question 2: The imbalance between real-time performance and security in data interaction restricts collaborative efficiency. Existing inspection systems mostly use technologies such as ZigBee to transmit data, which have significant latency and cannot support dynamic safe distance detection.

[0004] Problem 3: The lack of quantitative support for conflict detection logic leads to a high false alarm rate (a key problem addressed by this invention). Traditional solutions only determine conflicts based on static distance thresholds, without considering equipment motion parameters such as the robot dog's acceleration, the drone's turning angular velocity, and personnel status. This results in an excessively high false alarm rate, which not only interferes with normal inspection processes but may also cause operators to become desensitized to alarm signals and ignore real risks.

[0005] Question 4: Weak anti-interference capability of network configuration affects system stability. In scenarios with multiple devices operating concurrently, single Bluetooth networking is susceptible to electromagnetic interference and signal competition between devices, resulting in excessive packet loss. This leads to interruptions in the transmission of location and status data, affecting the continuity of collaborative operations. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a method and system for detecting safety conflicts in multi-device collaborative inspection. The present invention realizes high-precision collaboration and intelligent safety conflict prevention and control of multiple types of inspection subjects in a closed environment, which significantly improves the safety and reliability of inspection operations and is applicable to complex scenarios such as hydropower station tunnels, urban underground pipe corridors, and mine roadways.

[0007] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A method for detecting safety conflicts during multi-device collaborative inspections, comprising the following steps: Step 1: Deploy UWB positioning base stations and WIFI network routers at preset intervals in the tunnel, wherein the WIFI network routers are connected to each other via wires and communicate with the ground station terminal. Step 2: Configure UWB tags and WIFI communication modules on multiple mobile devices participating in the inspection. The mobile devices include robot dogs, drones, and smart wearable devices. Step 3: Using the UWB positioning base station, the three-dimensional position information of each mobile device is obtained based on the TOF algorithm, and the positioning correction is performed by fusing the onboard sensor data of each mobile device to obtain a real-time positioning result with an accuracy better than 0.3 meters; Alternatively, an adaptive weighted Kalman filter algorithm can be used to fuse UWB positioning data with airborne sensor data. By dynamically adjusting the weights in real time, the positioning accuracy can be consistently better than 0.3 meters. Step 4: The location information and status parameters of each mobile device are transmitted back to the ground station terminal in real time through the WIFI network routing. The WIFI network backhaul is achieved by employing a time-slot scheduling mechanism: priority time slots (30%) are allocated to drone location data, and secondary time slots (25%) are allocated to robot dog status parameters. Data from personnel-worn devices is transmitted using compression to ensure that the data backhaul latency is <100ms, meeting the concurrent transmission requirements of multiple devices within the tunnel. Step 5: The ground station terminal dynamically calculates the dynamic safe distance between any two mobile devices based on the transmitted status parameters. The dynamic safe distance is determined by the basic safe distance, the device motion parameter adjustment coefficient, and the personnel status adjustment coefficient. Step 6: Based on the comparison results between the dynamic safety distance and the actual distance, calculate the probability of conflict risk, and classify the risk level according to the preset threshold, and trigger different levels of early warning measures accordingly.

[0008] Preferably, in step 5, the formula for calculating the dynamic safety distance is: ; in: Dynamic safety distance, which is the threshold ultimately used for conflict determination; Basic safety distance: an initial safety threshold set based on device type and interaction scenario; Equivalent adjustment factor: A correction factor that integrates equipment movement and personnel status; its value range is [range missing]. This is to avoid excessively increasing the safety distance, which could reduce inspection efficiency.

[0009] Preferably, the equivalent adjustment factor Calculation method: Adjustment coefficient of equipment motion Personnel status adjustment coefficient The result is a combination of two factors, and the "larger factor principle" is adopted. That is, when both factors exist simultaneously, the coefficient with the greater influence is prioritized to avoid excessive safety distance caused by repeated superposition. ; ① Equipment motion adjustment coefficient : Based on calculations using the equipment's acceleration, turning angular velocity, and flight altitude, the derivation is performed for different scenarios: Scenario 1: Robot dog / drone accelerating in a straight line: ; Equipment acceleration: maximum acceleration of robot dog ≤ 1.5m / s², and of drone ≤ 1.0m / s²; Scenario 2: Turning motion of robot dog / drone: ; ω: Turning angular velocity, the maximum turning angular velocity of the robot dog is ≤90° / s, and that of the drone is ≤60° / s; Scenario 3: Drone flying at low altitude: ; h: Flight altitude of the drone, minimum flight altitude ≥ 1.5m; final The value can be: ; ②Personnel Status Adjustment Coefficient : Calculations are performed based on the personnel's work posture and movement speed, and are divided into different scenarios: Scenario 1: Personnel working in a bent-over / squatting position ; Scenario 2: Rapid movement of people ; Personnel movement speed: normal walking speed 0.8-1.2m / s, emergency evacuation speed ≤2.5m / s; final Values: .

[0010] Preferably, in step 5, the basic safety distance is set according to the interaction scenario, and the calculation formula is: The basic safety distance D = (equipment braking distance D1 + obstacle avoidance reaction distance D2 + environmental disturbance increment D3) × safety redundancy coefficient K. This includes: The robot dog is 1.0 meter away from the inspection personnel; The distance between the drone and the inspection personnel is 0.8 meters; The distance between the robot dog and the drone is 1.2 meters; The distance between multiple drones is 1.5 meters.

[0011] Preferably, in step 5, the calculation rules for the device motion parameter adjustment coefficient include: When the equipment accelerates in a straight line, the adjustment coefficient increases by 0.05 for every 0.2 m / s² increase in acceleration, with a maximum of 0.2. When the equipment turns, the adjustment coefficient increases by 0.03 for every 15° / s increase in steering angular velocity, with a maximum of 0.15. When the drone flies at an altitude of less than 3 meters, the adjustment factor increases by 0.08 for every 0.5 meters decrease in altitude.

[0012] Preferably, in step 5, the calculation rules for the personnel status adjustment coefficient include: When a smart wearable device detects that a person is working in a bent-over or squatting position and the body tilt angle is greater than 45°, the adjustment coefficient is 0.3. When the movement speed of personnel increases by 0.3 m / s, the adjustment coefficient increases by 0.06, with a maximum of 0.18.

[0013] Preferably, in step 6, the conflict risk probability assessment is divided into the following risk levels and corresponding early warnings are issued: Risk probability is 0: No risk, normal inspection. Risk probability between 0.01 and 0.99: low risk, triggering a low-frequency audible and visual alarm and controlling the relevant equipment to slow down; Risk probability equals 1: High risk, triggering high-frequency sound, light and vibration alarms, and instructing the equipment to stop or avoid obstacles in an emergency; Risk probability between 0.01 and 1 and expected conflict time ≤ 1 second: potential risk, provide 1 second advance warning and adjust equipment operating trajectory.

[0014] Preferably, risk probability The formula used to quantify conflict risk levels is as follows: ; The actual distance between the equipment and personnel is calculated from the corrected positioning data; The predicted time of intersection between the device and personnel trajectories is calculated by the edge gateway based on the speed and direction of movement. Current time; : Conflict risk probability, with a value range of 0-1, where 0 = no risk and 1 = extremely high risk.

[0015] A multi-device collaborative inspection safety conflict detection system, employing the aforementioned multi-device collaborative inspection safety conflict detection method, includes: Multiple UWB positioning base stations are deployed at intervals along the tunnel sidewall to perform the positioning source deployment in step 1; Multiple WIFI network routers are deployed at intervals along the tunnel, connected to each other via wires, and connected to the ground station terminal to perform the network construction in step 1; Multiple mobile terminal devices, including robot dogs equipped with UWB tags and WIFI modules, drones, and smart wearable devices, are used to perform the terminal configuration in step 2. The ground station terminal is equipped with a data receiving module, a positioning fusion module, a dynamic safety distance calculation module, a conflict risk assessment module, and an early warning execution module, and is configured to execute steps 3 to 6.

[0016] Preferably, the WIFI networking router supports dual-band communication and is interconnected through Cat6 Ethernet cables to form a wired-wireless hybrid transmission network, which is used to ensure low-latency and high-reliability backhaul of location information and status parameters in step 4.

[0017] Preferably, the smart wearable device integrates a posture sensor and an audio-visual vibration alarm, which is used to sense the status of personnel in real time and provide early warning information to the inspection personnel in steps 5 and 6.

[0018] Preferably, the system is applicable to collaborative inspection scenarios in hydropower station tunnels, urban underground utility tunnels, or mine underground roadways, and the basic safety distance parameters can be adjusted in step 5 according to the specific application scenario.

[0019] The present invention can achieve the following beneficial effects: 1. By integrating UWB positioning with the device's own sensor data, it effectively overcomes GPS rejection and signal interference problems, achieving high-precision real-time three-dimensional positioning and meeting the safety requirements of close-range collaborative operations.

[0020] 2. A wired WIFI networking architecture is adopted to build a low-latency, interference-resistant data transmission channel, ensuring real-time transmission of location and status information of multiple devices and supporting dynamic conflict detection.

[0021] 3. By combining the dynamic safety distance model of equipment movement status and personnel working posture, the accuracy of risk identification is significantly improved, and false alarms and missed alarms are reduced.

[0022] 4. Implement graded early warning and control strategies based on the conflict risk level. This can not only avoid high-risk collisions in a timely manner, but also intervene in potential risks in advance, taking into account both safety and inspection continuity.

[0023] 5. Supports collaborative operations between various terminals such as robot dogs, drones, and inspection personnel, and is suitable for various complex and enclosed scenarios such as hydropower stations, underground pipe corridors, and mines, with good versatility and scalability. Attached Figure Description

[0024] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 This is a system structure diagram of the present invention; Figure 2 This is a diagram of the WIFI deployment structure of the present invention; Figure 3 This is a diagram of the monitoring interface of the ground station terminal management software of the present invention. Detailed Implementation

[0025] Existing Technology: In hydropower station projects, tunnels are crucial channels for energy transmission and water flow, and their operation and maintenance safety directly determines the power generation efficiency and operational stability of the hydropower station. As of 2024, my country had built over 8,000 kilometers of hydropower station tunnels, with a significant proportion exceeding 1 kilometer in length. These tunnels not only face the structural complexity of conventional tunnels but also the harsh environment unique to hydropower station scenarios. The drawbacks of traditional manual inspection methods are even more pronounced in hydropower station tunnels: on the one hand, manual inspection is inefficient, requiring more than 2 hours for a single inspection of a 1-kilometer-long pressure tunnel; on the other hand, personnel must directly face multiple risks, such as slipping from leaking water, oxygen deficiency in confined spaces, etc., and manual observation is affected by light and humidity, as well as potential hazards such as valve wear and concrete cracks. To overcome this challenge, multi-device collaborative inspection models are being gradually adopted in hydropower station tunnels. However, the unique environment of hydropower station tunnels presents even more severe technical bottlenecks for multi-device collaboration: First, positioning becomes significantly more difficult. In addition to the common GPS failure issue, strong electromagnetic interference from high-voltage equipment amplifies the errors of existing Bluetooth RSSI and RFID positioning technologies. Furthermore, the dense array of valves and pipelines within hydropower station tunnels necessitates close-range collaboration between robotic sensors and drones, which current accuracy standards cannot fully meet. Second, data transmission stability is poor. The transmission rate of ZigBee technology is insufficient to support real-time data transmission from drones and robotic sensors in hydropower station tunnels. Third, collision detection adaptability is inadequate. Existing fixed distance thresholds do not consider the specific risks of hydropower station tunnels, resulting in excessively high false alarm rates and interference with normal inspection procedures. Therefore, it is urgent to construct a multi-device collaborative technology system with high-precision positioning, highly stable transmission, and intelligent collision detection specifically designed for the harsh environment of hydropower station tunnels to ensure the safe operation and maintenance of hydropower station tunnels.

[0026] The existing publicly available patent CN202510152880.2, "An Intelligent Control System for a Land-Air Dual-Use Tunnel Inspection Robot" (Publication No. CN119620667A), is the closest technical solution in the current field of multi-equipment collaborative tunnel inspection. It was published by Beijing University of Aeronautics and Astronautics in March 2025, and its core design is as follows: 1. Hardware configuration: The system adopts a dual-terminal collaborative architecture of "drone + unmanned vehicle". Each terminal is equipped with a camera, lidar, inertial measurement sensor and GPS positioning sensor. A signal repeater is deployed every 80 meters along the side wall inside the tunnel. Data is collected to the back-end management platform through a network communication module. At the same time, environmental sensors such as smoke and toxic gas are also provided. 2. Positioning and Collaborative Mode: Positioning is based on GPS and supplemented by inertial navigation. It supports two collaborative modes: integrated (UAV mounted on unmanned vehicle) and split (UAV high-altitude scanning + unmanned vehicle ground detection). It can achieve coverage of multiple areas such as the top and side walls of the tunnel, with a positioning accuracy of 1.0-1.8 meters and a data update frequency of 3 times / second. 3. Safety control mechanism: The obstacle observation module detects the distance between devices in real time and sets a fixed safety threshold (UAV-Vehicle ≥ 3 meters, equipment-tunnel wall ≥ 1.5 meters). When the lidar detects that the distance exceeds the limit, the intelligent control module sends an emergency hovering / braking command. There is no dynamic threshold adjustment logic. 4. Data transmission scheme: A hybrid transmission of 4G network and industrial Ethernet is adopted. Positioning and environmental data are transmitted back in real time via 4G, while high-definition images of drones are uploaded in batches via Ethernet. The measured data latency is 350-520ms. 4G signal interruption is prone to occur when multiple devices are connected concurrently.

[0027] To address the shortcomings of existing technologies, this invention provides a method and system for detecting security conflicts during multi-device collaborative inspections, with the main improvements being: 1. High-precision 3D positioning: By combining UWB with TOF positioning algorithm, airborne sensor fusion algorithm and error correction algorithm, it achieves 3D positioning accuracy within 0.3 meters, adapts to complex tunnel environments, and solves the problems of GPS failure and inertial navigation drift. 2. High real-time data transmission: Based on WIFI networking technology, it achieves high transmission rate and high data update frequency, supporting real-time backhaul and dynamic safe distance calculation of drones, robot dogs and smart wearable devices, and matching the response requirements of intelligent management and control systems; 3. Quantitative Conflict Detection: Through formulas, the risk level (low / high / potential) can be quantitatively judged, reducing the probability of false judgment and breaking through the limitations of existing static thresholds; 4. Stable and efficient networking: The dual-mode networking of Ethernet wired and WIFI wireless extends the communication distance to 100 meters, increases the communication deployment distance, reduces costs, and improves the stability of coverage in long tunnels.

[0028] Example 1: Specifically, such as Figure 1 , 2 As shown in Figure 3, this invention uses external UWB as the foundation to construct a collaborative system covering the entire process of "positioning-data transmission-detection-early warning," with conflict detection as the core control link. The design of each module revolves around improving the accuracy and real-time performance of conflict detection. The specific scheme is as follows: A method for detecting safety conflicts during multi-device collaborative inspections, comprising the following steps: Step 1: Deploy UWB positioning base stations and WIFI network routers at preset intervals in the tunnel, wherein the WIFI network routers are connected to each other via wires and communicate with the ground station terminal. In terms of hardware system, one set of WIFI network routers is deployed at regular intervals (e.g., 100m, which can be adjusted according to the site environment) inside the tunnel, and UWB positioning base stations are deployed at regular intervals (e.g., 30m, which can be adjusted according to the site environment). Step 2: Configure UWB tags and WIFI communication modules on multiple mobile devices participating in the inspection. The mobile devices include robot dogs, drones, and smart wearable devices. Step 3: Using the UWB positioning base station, the three-dimensional position information of each mobile device is obtained based on the TOF algorithm, and the positioning correction is performed by fusing the onboard sensor data of each mobile device to obtain a real-time positioning result with an accuracy better than 0.3 meters; Step 4: The location information and status parameters of each mobile device are transmitted back to the ground station terminal in real time through the WIFI network routing. Step 5: The ground station terminal dynamically calculates the dynamic safe distance between any two mobile devices based on the transmitted status parameters. The dynamic safe distance is determined by the basic safe distance, the device motion parameter adjustment coefficient, and the personnel status adjustment coefficient. A closed-loop management system is formed through "dynamic threshold calculation - risk quantification assessment - tiered early warning execution": First, the edge computing gateway sets a basic safety distance based on device type and scenario (e.g., 1.0 meter between a robot dog and a person, 0.8 meters between a drone and a person, which can be adjusted as a calibration parameter). Then, a dynamic safety distance formula is used... By integrating equipment motion parameters with personnel status adjustment coefficients, real-time safety thresholds are calculated. (1) Calculation of dynamic safety distance: The dynamic safety distance is the core basis for judging whether there is a conflict between equipment and personnel. It needs to integrate three major elements: basic safety distance, equipment motion parameter adjustment coefficient, and personnel status adjustment coefficient. The formula is as follows:

[0029] Where: the symbols are defined as follows Dynamic safety distance (unit: meters, m), which is the final threshold used for conflict determination; Basic safety distance (unit: m), the initial safety threshold set according to device type and interaction scenario; Equivalent adjustment factor (dimensionless), a correction factor that integrates equipment motion and personnel status, with a value range of [value missing]. (Avoid excessively increasing the safety distance, which could reduce inspection efficiency).

[0030] (2) Basic safety distance Typical parameter values ​​are shown in Table 1. These can be used as parameters to be calibrated and adjusted during field deployment.

[0031] Table 1 Basic Safety Distance Value table Typical value parameter table

[0032] (3) Equivalent adjustment coefficient Calculation method: Adjustment coefficient of equipment motion Personnel status adjustment coefficient It is formed by superposition, and the "larger factor principle" is adopted (that is, when two factors exist at the same time, the coefficient with the greater influence is taken as the main factor to avoid excessive safety distance due to repeated superposition): ; ① Equipment motion adjustment coefficient

[0033] Calculations based on the device's acceleration, turning angular velocity, and flight altitude (for drones only), derived for different scenarios: Scenario 1: Linear acceleration of robot dog / drone ; Equipment acceleration (unit: m / s²), maximum acceleration of robot dog ≤ 1.5 m / s², drone ≤ 1.0 m / s²; Note: For every 0.2 m / s² increase in acceleration, the adjustment coefficient increases by 0.05, up to a maximum of 0.2 (e.g., when the robot dog accelerates suddenly, the safe distance increases by 20%).

[0034] Scenario 2: Turning motion of robot dog / drone: ; ω: Turning angular velocity (unit: ° / s), the maximum turning angular velocity of the robot dog is ≤90° / s, and that of the drone is ≤60° / s; Note: For every 15° / s increase in turning angular velocity, the adjustment coefficient increases by 0.03, up to a maximum of 0.15 (e.g., when the drone turns rapidly, the safe distance increases by 15%).

[0035] Scenario 3: Drone flying at low altitude (height < 3 meters)

[0036] h: Flight altitude of the UAV (unit: m), minimum flight altitude ≥ 1.5m; Note: The adjustment factor increases by 0.08 for every 0.5m decrease in height. final The value can be:

[0037] ②Personnel Status Adjustment Coefficient

[0038] Calculations are performed based on the personnel's work posture and movement speed, and are divided into different scenarios: Scenario 1: Personnel working in a bent-over / squatting position

[0039] Note: When bending over, the person's reaction speed is reduced by 50%, and the safe distance needs to be increased by an additional 30% (determined by the posture sensor of the wearable device, triggered when the tilt angle is >45°).

[0040] Scenario 2: Rapid movement of people

[0041] Personnel movement speed (unit: m / s): Normal walking speed 0.8-1.2 m / s, emergency evacuation speed ≤2.5 m / s; Note: For every 0.3 m / s increase in speed, the adjustment coefficient increases by 0.06, up to a maximum of 0.18 (for example, when personnel are evacuating quickly, the safe distance increases by 18%).

[0042] final Values:

[0043] Step 6: Based on the comparison results between the dynamic safety distance and the actual distance, calculate the probability of conflict risk, and classify the risk level according to the preset threshold, and trigger different levels of early warning measures accordingly.

[0044] according to The system implements tiered early warning: In low-risk situations, the smart wearable device emits a low-frequency audible and visual alarm, and the robot dog / drone automatically reduces its speed by 50% to increase the safe distance; in high-risk situations, the smart wearable device triggers a high-frequency audible, visual, and vibration alarm, and the device immediately performs emergency actions; in the event of potential risks, a conflict direction prompt is pushed to the terminal 1 second in advance, and the device's movement trajectory is automatically adjusted, upgrading from "passive alarm" to "active obstacle avoidance," maximizing the reduction of the risk of equipment collisions and personnel injuries in hydropower station tunnels.

[0045] To quantify the conflict risk level (low / high / potential), a conflict risk probability is introduced. The formula is as follows:

[0046] The actual distance between the equipment and personnel (unit: m) is calculated from the corrected positioning data; The predicted time of intersection between the device and personnel trajectories (in seconds) is calculated by the edge gateway based on the speed and direction of movement. Current time (in seconds); Conflict risk probability (dimensionless), with a value range of 0-1 (0 = no risk, 1 = extremely high risk).

[0047] The risk level correspondence is shown in Table 2: Table 2 Risk Level Correspondence Table

[0048] The advantages of this invention are: (1) Hardware networking architecture: WIFI networking routers are deployed every 100 meters in the tunnel, and UWB positioning base stations are deployed every 30 meters. The terminal equipment is equipped with UWB tags and WIFI modules. The ground station connects the external signal source and the terminal to realize the UWB collaboration with the airborne sensor to provide accurate positioning information. The WIFI networking router and network cable form a data transmission network to ensure the stability and real-time performance of the system status feedback. (2) Dynamic safety distance calculation method: integrate equipment motion parameters (machine dog acceleration a, equipment turning angular velocity ω, UAV flight altitude h) with personnel status (personnel posture, personnel movement speed) By using a dynamic safety distance calculation method, the safety threshold can be dynamically adjusted, breaking through the limitations of the traditional static threshold. (3) Conflict detection and early warning system: The ground station quantifies the risk based on the returned status and triggers three levels of early warning: low / high / potential, and alerts the monitoring personnel through sound and light. In terms of positioning accuracy, the existing CN202510152880.2 scheme is mainly based on two-dimensional positioning, with an accuracy of only 1.0-1.8 meters. However, this invention achieves three-dimensional positioning (including Z-axis height) within 0.3 meters by using UWB's TOF positioning algorithm and airborne sensor fusion algorithm.

[0049] In terms of data transmission performance, existing solutions use a hybrid transmission of 4G and industrial Ethernet, which is unsuitable for use in tunnel environments. The output transmission setup method of this invention can synchronize equipment motion parameters and personnel status data in real time, providing high-frequency data support for dynamic safe distance calculation.

[0050] The optimization of the conflict detection logic is particularly significant. Existing solutions rely on a static distance threshold of 1.5-3 meters, without considering equipment movement and environmental interference, resulting in a high false alarm rate. This invention constructs a quantitative detection system through a dynamic safety distance model and a conflict risk probability model, reducing the false alarm rate and upgrading from "passive threshold judgment" to "active risk quantification," which greatly improves the accuracy of early warning.

[0051] The scope of collaborative scenarios has also been significantly expanded. Existing solutions only support collaboration between two types of equipment: "drones + unmanned vehicles," which cannot adapt to personnel inspections and access to small devices. This invention achieves cross-dimensional collaboration between "ground robot dogs + high-altitude drones + personnel wearable devices," covering all types of inspection subjects and meeting the full-scenario operation and maintenance needs of complex tunnels.

[0052] Example 2: A multi-device collaborative inspection safety conflict detection system, the system comprising: Multiple UWB positioning base stations are deployed in groups every 30 meters along the tunnel sidewall to perform the positioning source deployment in step 1; Multiple WIFI network routers are deployed in groups every 100 meters along the tunnel, connected to each other via wires, and connected to the ground station terminal to perform the network construction in step 1; Multiple mobile terminal devices, including robot dogs equipped with UWB tags and WIFI modules, drones, and smart wearable devices, are used to perform the terminal configuration in step 2. The ground station terminal is equipped with a data receiving module, a positioning fusion module, a dynamic safety distance calculation module, a conflict risk assessment module, and an early warning execution module, and is configured to execute steps 3 to 6 as described in claim 1.

[0053] The WIFI networking router supports dual-band (2.4 GHz / 5 GHz) communication and is interconnected through Cat6 Ethernet cables to form a wired-wireless hybrid transmission network, which is used to ensure low-latency and high-reliability backhaul of location information and status parameters in step 4.

[0054] The smart wearable device integrates a posture sensor and an audio-visual vibration alarm, which is used to sense the status of personnel in real time and provide early warning information to the inspection personnel in steps 5 and 6.

[0055] The system is applicable to collaborative inspection scenarios in hydropower station tunnels, urban underground utility tunnels, or mine underground roadways, and the basic safety distance parameters can be adjusted in step 5 according to the specific application scenario.

[0056] The hardware selection in this embodiment is as follows: UWB positioning base station: It adopts DW1000 chip, transmits power of 12dBm, and deploys 1 group every 30 meters (7 groups in total). It is installed on the side wall of the tunnel 2 meters above the ground and is powered by an external power source. WIFI networking router: adopts dual-band (2.4GHz / 5GHz) enterprise-grade router, with 1 group deployed every 100 meters (3 groups in total). The routers are connected to each other via Cat6 Ethernet cable and connected to the ground station. Terminal equipment: Robot dog: Unitree Go2, equipped with UWB tag, IMU, and LiDAR; Drone: Halo P6000, equipped with UWB tag, LiDAR, and three-axis stabilized camera; Smart wearable devices: industrial control computer, integrating UWB tags, posture sensors, and sound, light and vibration alarms.

[0057] The implementation method is as follows: 1: After the UWB base station is powered on, it automatically completes time synchronization, calibrates the base station position through the calibration point at the tunnel entrance, and establishes a local coordinate system in the tunnel; 2: After the terminal device is powered on, the UWB tag sends a positioning signal to the base station. The base station calculates the distance between the tag and the base station using the TOF algorithm (Time of Flight algorithm) and performs positioning fusion with the airborne sensor information. 3: The ground station receives location and status information of mobile devices (robot dogs, drones, wearable devices).

[0058] ② Data transmission implementation (steps 4-6) 4: Status data such as the robot dog's walking speed, the drone's flight altitude, and the personnel's movement speed are transmitted to the router via the corresponding WIFI frequency band; 5: After the WIFI router aggregates the data, it transmits it to the ground station via network cable. The ground station preprocesses the data (filtering outliers, such as invalid data when the drone's altitude suddenly jumps to 5 meters). 6: The preprocessed data is divided into two categories: location data used in real time for collision detection, and high-definition images.

[0059] ③ Calculation of dynamic safety distance (steps 7-9) 7: Set a basic safety distance Drones - Personnel =0.8 meters, robot dog - personnel, =1.0 rice; 8: Calculate the equipment motion adjustment coefficient The drone flies at an altitude of 2.2m (<3m). =0.08×⌊(3−2.2) / 0.5⌋=0.08×1=0.08, no acceleration / steering, therefore =0.08; 9: Calculate the personnel status adjustment coefficient For personnel walking upright (angle of inclination 15° < 45°), the moving speed is 0.9 m / s ≤ 1.2 m / s, therefore... =0; calculate =0.8×(1+max(0.08,0))=0.864 meters.

[0060] ④ Risk quantification and early warning implementation (steps 10-12) 10: Ground station calculates the actual distance between the drone and personnel: =0.7 meters (<0.864 meters), substitute into Formula =(0.864−0.7) / (0.864×0.2)=0.164 / 0.1728≈0.95 (low-risk upper limit); 11: The risk level is determined to be low, triggering an early warning: the smart wearable device emits a low-frequency audible and visual alarm, and the drone automatically slows down; 12: If It was further reduced to 0.69 meters (≤0.8×0.864=0.691 meters). =1 (High Risk): The drone immediately hovers, the personnel's wristband triggers high-frequency vibration, and the robot dog stops moving and retreats 0.5 meters.

[0061] This embodiment, based on a 200-meter-long tunnel corridor (3m×3m cross-section, concrete lining) of a hydropower station, conducts a collaborative inspection experiment using a "robot dog + drone + inspection personnel" to verify the feasibility of the technical solution of this invention. Through experiments, a simulation experiment was conducted in a 200-meter corridor environment. By fusing UWB data with airborne sensor data, the location of mobile devices can be located in real time with an error of less than 0.3 meters. Simultaneously, data is transmitted back to the ground station for safe distance calculation and collision detection, with an error rate as low as 5%.

[0062] The above embodiments are merely preferred technical solutions of the present invention and should not be considered as limitations on the present invention. The scope of protection of the present invention should be limited to the technical solutions described in the claims, including equivalent substitutions of the technical features described in the claims. That is, equivalent substitutions and improvements within this scope are also within the scope of protection of the present invention.

Claims

1. A method for detecting safety conflicts during multi-device collaborative inspection, characterized in that, Includes the following steps: Step 1: Deploy UWB positioning base stations and WIFI network routers at preset intervals in the tunnel, wherein the WIFI network routers are connected to each other via wires and communicate with the ground station terminal. Step 2: Configure UWB tags and WIFI communication modules on multiple mobile devices participating in the inspection. The mobile devices include robot dogs, drones, and smart wearable devices. Step 3: Using the UWB positioning base station, the three-dimensional position information of each mobile device is obtained based on the TOF algorithm, and the positioning correction is performed by fusing the onboard sensor data of each mobile device to obtain a real-time positioning result with an accuracy better than 0.3 meters; Step 4: The location information and status parameters of each mobile device are transmitted back to the ground station terminal in real time through the WIFI network routing. Step 5: The ground station terminal dynamically calculates the dynamic safe distance between any two mobile devices based on the transmitted status parameters. The dynamic safe distance is determined by the basic safe distance, the device motion parameter adjustment coefficient, and the personnel status adjustment coefficient. Step 6: Based on the comparison results between the dynamic safety distance and the actual distance, calculate the probability of conflict risk, and classify the risk level according to the preset threshold, and trigger different levels of early warning measures accordingly.

2. The method for detecting safety conflicts during multi-device collaborative inspection according to claim 1, characterized in that, In step 5, the formula for calculating the dynamic safety distance is: ； in: Dynamic safety distance, which is the threshold ultimately used for conflict determination; Basic safety distance, an initial safety threshold set based on device type and interaction scenario; Equivalent adjustment factor: A correction factor that integrates equipment movement and personnel status; its value range is [range missing]. This is to avoid excessively increasing the safety distance, which could reduce inspection efficiency.

3. The method for detecting safety conflicts during multi-device collaborative inspection according to claim 2, characterized in that, Equivalent adjustment factor Calculation method: Adjustment coefficient of equipment motion Personnel status adjustment coefficient The result is a combination of two factors, and the "larger factor principle" is adopted. That is, when both factors exist simultaneously, the coefficient with the greater influence is prioritized to avoid excessive safety distance caused by repeated superposition. ； ① Equipment motion adjustment coefficient : Based on calculations using the equipment's acceleration, turning angular velocity, and flight altitude, the derivation is performed for different scenarios: Scenario 1: Robot dog / drone accelerating in a straight line: ； Equipment acceleration: maximum acceleration of robot dog ≤ 1.5m / s², and of drone ≤ 1.0m / s²; Scenario 2: Turning motion of robot dog / drone: ； ω: Turning angular velocity; the maximum turning angular velocity of the robot dog is ≤90° / s, and that of the drone is ≤60° / s. Scenario 3: Drone flying at low altitude: ； h: Flight altitude of the drone, minimum flight altitude ≥ 1.5m; final The value can be: ； ②Personnel Status Adjustment Coefficient : Calculations are performed based on the personnel's work posture and movement speed, and are divided into different scenarios: Scenario 1: Personnel working in a bent-over / squatting position ； Scenario 2: Rapid movement of people ； Personnel movement speed: normal walking speed 0.8-1.2m / s, emergency evacuation speed ≤2.5m / s; final Values: 。 4. The method for detecting safety conflicts during multi-device collaborative inspection according to claim 2, characterized in that, In step 5, the basic safety distance is set according to the interaction scenario, and the calculation formula is: Basic safety distance D = (equipment braking distance D1 + obstacle avoidance reaction distance D2 + environmental disturbance increment D3) × safety redundancy coefficient K.

5. The method for detecting safety conflicts during multi-device collaborative inspection according to claim 2, characterized in that, In step 5, the calculation rules for the device motion parameter adjustment coefficient include: When the equipment accelerates in a straight line, the adjustment coefficient increases by 0.05 for every 0.2 m / s² increase in acceleration, with a maximum of 0.

2. When the equipment turns, the adjustment coefficient increases by 0.03 for every 15° / s increase in steering angular velocity, with a maximum of 0.

15. When the drone flies at an altitude of less than 3 meters, the adjustment factor increases by 0.08 for every 0.5 meters decrease in altitude.

6. The method for detecting safety conflicts during multi-device collaborative inspection according to claim 2, characterized in that, In step 5, the calculation rules for the personnel status adjustment coefficient include: When a smart wearable device detects that a person is working in a bent-over or squatting position and the body tilt angle is greater than 45°, the adjustment coefficient is 0.

3. When the movement speed of personnel increases by 0.3 m / s, the adjustment coefficient increases by 0.06, with a maximum of 0.

18.

7. The method for detecting safety conflicts during multi-device collaborative inspection according to claim 1, characterized in that, In step 6, the conflict risk probability assessment is divided into the following risk levels and corresponding early warnings are issued: Risk probability is 0: No risk, normal inspection. Risk probability between 0.01 and 0.99: low risk, triggering a low-frequency audible and visual alarm and controlling the relevant equipment to slow down; Risk probability equals 1: High risk, triggering high-frequency sound, light and vibration alarms, and instructing the equipment to stop or avoid obstacles in an emergency; Risk probability between 0.01 and 1 and expected conflict time ≤ 1 second: potential risk, provide 1 second advance warning and adjust equipment operating trajectory.

8. The method for detecting safety conflicts during multi-device collaborative inspection according to claim 7, characterized in that, Risk probability The formula used to quantify conflict risk levels is as follows: ； The actual distance between the equipment and personnel is calculated from the corrected positioning data; The predicted time of intersection between the device and personnel trajectories is calculated by the edge gateway based on the speed and direction of movement. Current time; : Conflict risk probability, with a value range of 0-1, where 0 = no risk and 1 = extremely high risk.

9. A multi-device collaborative inspection safety conflict detection system, characterized in that, The system employs a multi-device collaborative inspection safety conflict detection method according to any one of claims 1-8, the system comprising: Multiple UWB positioning base stations are deployed at intervals along the tunnel sidewall to perform the positioning source deployment in step 1; Multiple WIFI network routers are deployed at intervals along the tunnel, connected to each other via wires, and connected to the ground station terminal to perform the network construction in step 1; Multiple mobile terminal devices, including robot dogs equipped with UWB tags and WIFI modules, drones, and smart wearable devices, are used to perform the terminal configuration in step 2. The ground station terminal is equipped with a data receiving module, a positioning fusion module, a dynamic safety distance calculation module, a conflict risk assessment module, and an early warning execution module, and is configured to execute steps 3 to 6.

10. The multi-device collaborative inspection safety conflict detection system according to claim 9, characterized in that, The WIFI networking router supports dual-band communication and is interconnected through Cat6 Ethernet cables to form a wired-wireless hybrid transmission network, which is used to ensure low-latency and high-reliability backhaul of location information and status parameters in step 4.

11. The multi-device collaborative inspection safety conflict detection system according to claim 9, characterized in that, The smart wearable device integrates a posture sensor and an audio-visual vibration alarm, which is used to sense the status of personnel in real time and provide early warning information to the inspection personnel in steps 5 and 6.

12. The multi-device collaborative inspection safety conflict detection system according to claim 9, characterized in that, The system is applicable to collaborative inspection scenarios in hydropower station tunnels, urban underground utility tunnels, or mine underground roadways, and the basic safety distance parameters can be adjusted in step 5 according to the specific application scenario.

Images