Robot emergency response method, device and equipment in hydrogen detonation scene and medium

Abstract

The invention discloses a hydrogen deflagration scene oriented robot emergency response method, device and equipment and a medium, and the method comprises the steps that a robot obtains real-time hydrogen concentration data and carries out concentration analysis, and if the concentration analysis result is abnormal, a preset first algorithm is used for determining whether the hydrogen concentration data is abnormal or not according to the hydrogen concentration data and a preset real-time map; determining a hydrogen explosion venting area; the robot obtains real-time temperature data, fire intensity data and smoke concentration data, and obtains fire position coordinates through a preset second algorithm; the robot plans an emergency response path according to the hydrogen explosion venting area and the fire position coordinates; and the robot arrives at the fire scene according to the emergency response path and monitors field environment data in real time so as to execute a fire extinguishing action. According to the method, the robot is used for replacing rescue personnel to execute the whole-process action of emergency response in the hydrogen detonation scene, so that the casualty probability of the rescue personnel in the hydrogen detonation scene is reduced.

Description

Technical Field

[0001] The present invention relates to the field of emergency response, and particularly to a robot emergency response method, device, equipment and medium for hydrogen deflagration scenarios. Background Art

[0002] In modern society, disaster management and emergency response are of great importance, which require efficient monitoring, early warning and handling capabilities. The rapid development of robot technology provides new solutions for dealing with emergencies. By carrying various sensors and devices on an aerial platform, it provides real-time data and emergency support for the ground. At the same time, using robots for emergency response can replace rescue workers to detect and handle emergencies at the disaster site, thereby reducing casualties and lowering the cost of human and material resources.

[0003] Currently, traditional robot systems mainly focus on image acquisition and monitoring, and fail to fully utilize advanced perception and navigation technologies, as well as intelligent decision support systems. At the disaster site, such as fires and hydrogen leaks, more comprehensive emergency response capabilities are required, rather than just image data. In the prior art, although there are already some robots used for disaster monitoring and emergency response, they lack comprehensive integrated functions and fail to provide comprehensive solutions for complex scenarios such as fires and hydrogen leaks. Therefore, in many rescue links for complex scenarios such as fires and hydrogen leaks, especially the fire extinguishing link, manual intervention is still required. During this process, rescue workers may face the situation of secondary deflagration at any time, and their personal safety cannot be fully guaranteed. Summary of the Invention

[0004] The present invention provides a robot emergency response method, device, equipment and medium for hydrogen deflagration scenarios, which can reduce the casualty probability of rescue workers in hydrogen deflagration scenarios, thereby improving rescue safety.

[0005] In a first aspect, an embodiment of the present invention provides a robot emergency response method for hydrogen deflagration scenarios, including:

[0006] The robot obtains real-time hydrogen concentration data and conducts concentration analysis. If the result of the concentration analysis is abnormal, the hydrogen deflagration area is determined according to the hydrogen concentration data and a preset real-time map through a preset first algorithm;

[0007] The robot obtains real-time temperature data, fire intensity data and smoke concentration data, and obtains the fire location coordinates through a preset second algorithm;

[0008] The robot conducts emergency response path planning according to the hydrogen deflagration area and the fire location coordinates;

[0009] The robot reaches the fire scene according to the emergency response path and monitors the on-site environmental data in real time to perform fire extinguishing actions.

[0010] In the embodiment of the present invention, by determining the hydrogen explosion venting area and marking the fire location, data support is provided for subsequent planning of the emergency response route for the robot to reach the fire scene; by reasonably planning the route for the robot to reach the fire scene according to the hydrogen explosion venting area and the fire location coordinates, the efficiency of the robot to reach the fire scene is improved; by monitoring the on-site environmental data in real time to perform fire extinguishing actions, it can ensure that the fire extinguishing actions performed achieve the fire extinguishing effect. In this application, from obtaining on-site data to planning the emergency response path and then to performing fire extinguishing actions, the entire process is operated by the robot without manual intervention. Compared with the prior art, using the robot to replace rescue personnel can greatly reduce the probability of casualties of rescue personnel when taking emergency response measures in the hydrogen deflagration scenario.

[0011] Further, the robot obtains real-time hydrogen concentration data and performs concentration analysis. If the result of the concentration analysis is abnormal, the hydrogen explosion venting area is determined through a preset first algorithm according to the hydrogen concentration data and a preset real-time map. Specifically:

[0012] The robot obtains real-time hydrogen concentration data and analyzes whether the concentration data exceeds a preset first threshold;

[0013] If the concentration data exceeds the first threshold, the result of the concentration analysis is abnormal. The robot will analyze the hydrogen concentration distribution through a preset first algorithm according to the hydrogen concentration data and the preset real-time map, and superimpose a gas concentration distribution map on the real-time map according to the analysis result of the hydrogen concentration distribution; wherein, the first algorithm includes a preset spatial interpolation algorithm and a Gaussian mixture model;

[0014] According to the gas concentration distribution map and a preset safety regulation, the hydrogen explosion venting area is determined through a preset boundary detection algorithm.

[0015] In the embodiment of the present invention, the hydrogen explosion venting area is accurately determined through a preset algorithm and safety regulation, providing accurate data support for subsequent emergency response path planning of the robot.

[0016] Further, the robot obtains real-time temperature data, fire intensity data, and smoke concentration data, and obtains the fire location coordinates through a preset second algorithm. Specifically:

[0017] The robot obtains real-time temperature distribution data, flame intensity data, and smoke concentration data, and all perform normalization processing to obtain initial fire data;

[0018] Perform data fusion on the fire data through a preset second algorithm to obtain the fire location coordinates; wherein, the second algorithm includes a preset neural network algorithm and a Bayesian network model.

[0019] In the embodiment of the present invention, by using rich on-site real-time data and a preset algorithm, the fire location coordinates are accurately determined, providing accurate data support for the subsequent emergency response path planning of the robot.

[0020] Further, the robot performs emergency response path planning according to the hydrogen explosion relief area and the fire location coordinates, specifically:

[0021] Align and fuse the hydrogen explosion relief area and the fire location coordinates in space to obtain comprehensive on-site data;

[0022] The robot plans an emergency response path according to the comprehensive on-site data through a preset path planning algorithm, a preset genetic algorithm, and a preset reinforcement learning algorithm.

[0023] In the embodiment of the present invention, by aligning and fusing the hydrogen explosion relief area and the fire location coordinates in space, the main influencing factors of the emergency response are comprehensively considered, and the emergency response path is planned from different dimensions one by one through multiple algorithms to ensure the safety, effectiveness, and efficiency of the emergency response path.

[0024] Further, the robot arrives at the fire scene according to the emergency response path and real-time monitors the on-site environmental data to perform fire extinguishing actions, specifically:

[0025] The robot arrives at the fire scene according to the emergency response path and real-time obtains the on-site temperature data, on-site flame intensity data, and on-site gas concentration data;

[0026] Perform data fusion on the on-site temperature, on-site flame intensity data, and on-site gas concentration data to obtain on-site environmental data;

[0027] Determine the on-site fire situation according to the on-site environmental data through a preset spatio-temporal convolutional neural network model and a preset inference system;

[0028] Predict the fire development trend according to the on-site environmental data through an algorithm that mixes a preset physical diffusion model and a machine learning regression model;

[0029] Determine the fire extinguishing strategy according to the on-site fire situation and the fire development trend, and perform fire extinguishing actions according to the fire extinguishing strategy; wherein, the fire extinguishing strategy includes selecting the fire extinguishing position, determining the type of fire extinguishing agent, determining the spraying flow rate of the fire extinguishing agent, and selecting the spraying angle of the fire extinguishing agent.

[0030] In an embodiment of the present invention, by monitoring on-site environmental data in real time, determining the on-site fire situation based on the on-site environmental data and predicting the fire development trend, determining a suitable fire extinguishing strategy according to the on-site fire situation and the fire development trend, and having a robot execute a fire extinguishing action according to the fire extinguishing strategy, the execution of the fire extinguishing strategy in an intelligent manner can improve the efficiency of strategy formulation, and by using a robot to replace rescue personnel in the traditional method to execute the fire extinguishing strategy, the safety of rescue personnel during emergency response for a hydrogen deflagration scenario can be improved, and the risk of personnel injury can be reduced.

[0031] Furthermore, while monitoring on-site environmental data in real time to execute a fire extinguishing action, the robot will also conduct real-time fire extinguishing effect evaluation and autonomously change the fire extinguishing strategy according to the evaluation result. Among them, the conduct of real-time fire extinguishing effect evaluation and autonomously changing the fire extinguishing strategy according to the evaluation result is specifically as follows:

[0032] While the robot is extinguishing the fire, it collects on-site temperature data, flame intensity data, and thermal imaging data in real time, and compares each data with the corresponding predicted data calculated by a preset fire spread prediction model to obtain a fire extinguishing effect evaluation result;

[0033] If the error between each data and the corresponding predicted data is less than a preset percentage, the evaluation result is valid, and the original fire extinguishing strategy is maintained;

[0034] Otherwise, the evaluation result is invalid, and the fire extinguishing strategy is adjusted; the adjustment of the fire extinguishing strategy includes increasing the spraying flow rate of the fire extinguishing agent, changing the spraying angle of the fire extinguishing agent, and optimizing the fire extinguishing movement route.

[0035] In an embodiment of the present invention, by evaluating the on-site fire extinguishing effect while the robot is extinguishing the fire and adjusting the fire extinguishing strategy if the expected fire extinguishing effect is not achieved, the effectiveness and reliability of the robot's fire extinguishing are ensured.

[0036] Furthermore, the execution of the fire extinguishing action also includes the robot obtaining the liquid level data of the fire extinguishing agent in real time. When the liquid level data is lower than a preset second threshold, a fire extinguishing agent replenishment strategy is executed. Among them, the execution of the fire extinguishing agent replenishment strategy is specifically as follows:

[0037] The robot accesses a preset high-precision map containing the locations of fire extinguishing agent replenishment stations, and plans a safe path to the nearest replenishment station through a preset path planning algorithm;

[0038] According to the safe path, it reaches the nearest replenishment station, replenishes the fire extinguishing agent, and obtains the liquid level data of the fire extinguishing agent in real time. When the liquid level data exceeds a preset third threshold, the replenishment of the fire extinguishing agent is stopped;

[0039] Access the preset real-time fire map to obtain the first fire location data, and plan a return path through the preset path planning algorithm;

[0040] Arrive at the first fire location according to the return path and continue the fire extinguishing operation.

[0041] In the embodiment of the present invention, the robot autonomously executes the fire extinguishing agent replenishment strategy to ensure that the fire extinguishing process does not require the intervention of rescue personnel throughout the process, thereby further reducing personnel injuries and ensuring the safety of rescue personnel.

[0042] In a second aspect, an embodiment of the present invention provides a robot emergency response device for a hydrogen deflagration scenario, including an explosion venting area determination module, a fire location acquisition module, a response path planning module, and a fire extinguishing operation execution module, where,

[0043] The explosion venting area determination module is used for the robot to obtain real-time hydrogen concentration data and perform concentration analysis. If the result of the concentration analysis is abnormal, the hydrogen explosion venting area is determined according to the hydrogen concentration data and the preset real-time map through a preset first algorithm;

[0044] The fire location acquisition module is used for the robot to obtain real-time temperature data, fire intensity data, and smoke concentration data, and obtain the fire location coordinates through a preset second algorithm;

[0045] The response path planning module is used for the robot to plan an emergency response path according to the hydrogen explosion venting area and the fire location coordinates;

[0046] The fire extinguishing operation execution module is used for the robot to reach the fire scene according to the emergency response path and monitor the on-site environment data in real time to execute the fire extinguishing operation.

[0047] In the embodiment of the present invention, the explosion venting area determination module determines the hydrogen explosion venting area, providing accurate data support for subsequent planning of the emergency response route for the robot to go to the fire scene; the fire location acquisition module obtains the accurate fire location, providing accurate data support for subsequent planning of the emergency response route for the robot to go to the fire scene; the response path planning module accurately plans the emergency response path for the robot to carry out fire rescue according to the hydrogen explosion venting area and the fire location, ensuring the safety and efficiency on the robot's response path; the fire extinguishing operation execution module controls the robot to execute the fire extinguishing operation, thereby replacing the situation where rescue personnel arrive at the hydrogen deflagration scene for rescue in the traditional way, thereby reducing the probability of personnel injuries during the rescue process.

[0048] In a third aspect, an embodiment of the present invention provides a terminal device, including: a processor, a memory, a communication interface, and a communication bus. The processor, the memory, and the communication interface complete communication with each other through the communication bus;

[0049] The memory is used to store at least one executable instruction, and the executable instruction causes the processor to execute the operations of the robot emergency response method for hydrogen deflagration scenarios as described in any one of the above.

[0050] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, which includes a stored computer program. When the computer program runs, it controls the device / device where the computer-readable storage medium is located to execute the robot emergency response method for hydrogen deflagration scenarios as described in any one of the above.

[0051] The above description is only an overview of the technical solutions of the embodiments of the present invention. In order to be able to understand the technical means of the embodiments of the present invention more clearly, it can be implemented according to the content of the description. And in order to make the above and other purposes, features and advantages of the embodiments of the present invention more obvious and understandable, the specific embodiments of the present invention are given below. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a schematic diagram of a robot emergency response method for hydrogen deflagration scenarios provided by an embodiment of the present invention;

[0053] Figure 2 It is a structural diagram of a robot emergency response device for hydrogen deflagration scenarios provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0054] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0055] Embodiment 1:

[0056] As Figure 1 shown, a robot emergency response method for hydrogen deflagration scenarios provided by an embodiment of the present invention includes the following steps:

[0057] S11. The robot obtains real-time hydrogen concentration data and performs concentration analysis. If the result of the concentration analysis is abnormal, the hydrogen deflagration area is determined according to the hydrogen concentration data and the preset real-time map through a preset first algorithm;

[0058] S12. The robot obtains real-time temperature data, fire intensity data, and smoke concentration data, and obtains the fire location coordinates through a preset second algorithm;

[0059] S13. The robot plans an emergency response path according to the hydrogen explosion relief area and the fire location coordinates.

[0060] S14. The robot reaches the fire scene according to the emergency response path and monitors the on-site environmental data in real time to perform fire extinguishing operations.

[0061] It should be noted that in the specific implementation process, the embodiments of the present invention are independently completed by several (each with the same function and its own patrol area) all-round robots throughout the entire emergency response process. The robot integrates multiple functions such as environmental monitoring, fire recognition, hydrogen explosion relief perception, decision-making, fire extinguishing execution, and data feedback, and has a high degree of autonomous operation ability. Its body structure integrates multiple sensor modules, including temperature, smoke, infrared, gas concentration and other sensing devices, and can comprehensively grasp the on-site situation. At the same time, a high-performance computing and control unit is installed inside the robot. After obtaining environmental data, it can immediately complete data analysis and situation judgment, and thus autonomously generate emergency strategies.

[0062] In this embodiment, step S11 is specifically as follows: The robot obtains real-time hydrogen concentration data and analyzes whether the concentration data exceeds a preset first threshold; if the concentration data exceeds the first threshold, the result of the concentration analysis is abnormal, and the robot will analyze the hydrogen concentration distribution according to the hydrogen concentration data and the preset real-time map through a preset first algorithm, and superimpose a gas concentration distribution map on the real-time map according to the analysis result of the hydrogen concentration distribution; wherein, the first algorithm includes a preset spatial interpolation algorithm and a Gaussian mixture model; according to the gas concentration distribution map and the preset safety regulations, the hydrogen explosion relief area is determined through a preset boundary detection algorithm.

[0063] In the specific implementation, in order to accurately identify hydrogen leakage, the robot is equipped with gas sensors with high selectivity and high sensitivity to hydrogen. These sensors can work based on different principles, such as semiconductor type, electrochemistry type or thermal conductivity type. In the embodiments of the present invention, the hydrogen sensor includes a sensitive material layer (such as doped metal oxide). When exposed to an air environment in the detection area, hydrogen molecules react with the sensitive material to cause a change in its resistivity. The processing module measures the resistance change in real time and calculates the corresponding hydrogen concentration according to the calibration curve. The detection range covers 50 ppm to 10,000 ppm, outputs concentration data in ppm units, and has a response time of less than 30 seconds and a concentration measurement error of less than ±5%. The sensor has a fast response and recovery time and outputs an analog or digital signal of the hydrogen concentration level.

[0064] Further, the robot will filter the original gas sensor data to reduce noise and improve the signal-to-noise ratio. Calibrate the gas sensor to ensure the accuracy of the concentration readings.

[0065] Further, in the embodiment of the present invention, the first threshold can be set to 4% volume of hydrogen concentration, and after determining that the result of the concentration analysis is abnormal, mark the area exceeding the first prediction as a potential high deflagration risk area.

[0066] Further, in the embodiment of the present invention, the process of using the spatial interpolation algorithm for data processing is specifically as follows: Adopt the Kriging interpolation algorithm or the inverse distance weighted interpolation (IDW) to estimate the hydrogen concentration in the area not directly covered according to the known sensor data. During the interpolation calculation process, use the spatial distribution and distance weighting of the sensor data to ensure that the estimated value has a high accuracy (the error does not exceed ±0.5% volume) in the area without data.

[0067] Further, in the embodiment of the present invention, the process of analyzing the hydrogen concentration distribution through the Gaussian mixture model (GMM) is specifically as follows: Model the spatial distribution of the gas, and use the Gaussian mixture model based on the maximum likelihood estimation (MLE) to fit the distribution of the hydrogen concentration. The initial parameters of the GMM model are optimized through the EM algorithm (expectation maximization algorithm), and the model outputs the probability density function (PDF) to provide the distribution probability of the hydrogen concentration in the environment.

[0068] Further, overlaying the gas concentration distribution map on the real-time map is specifically as follows: Based on the interpolation result and the GMM model, generate a two-dimensional gas concentration heat map or a three-dimensional concentration distribution map on the environmental map, and each pixel point or voxel represents the hydrogen concentration value of the corresponding area. The resolution of the gas concentration map is not less than 10 cm / pixel (two-dimensional map) or 10 cm 3 / voxel (three-dimensional map) to ensure the spatial accuracy.

[0069] It should be noted that the real-time map is obtained through a lidar system, and the acquisition process is specifically as follows: The lidar system includes a laser emission module and a receiving module. The emission module scans the environment with a pulsed laser beam of 905 nm wavelength within the horizontal angle range of 0° to 360°. The receiving module measures the time difference between the laser pulse from emission to reflection back to the receiver through the time of flight, and calculates the coordinates of each reflection point in the three-dimensional space in combination with the emission angle to generate three-dimensional point cloud data with a density of not less than 100,000 points / second. The processing module constructs a real-time environmental map based on the point cloud data, uses feature extraction and matching algorithms for precise positioning and navigation, and issues an obstacle avoidance control signal when detecting an obstacle with a distance less than a preset threshold (such as 0.5 meters). The system is applicable to complex industrial environments with dust, low light, and reflective surfaces, and the spatial measurement error is less than ±2 cm.

[0070] Further, according to the gas concentration distribution map and a preset safety regulation, the hydrogen explosion venting area is determined through a preset boundary detection algorithm, specifically: by comparing with a preset safety standard (such as a hydrogen concentration of 4% by volume or higher), based on the gas concentration map and the safety regulation, a boundary detection algorithm (such as a threshold-based contour extraction method) is used to determine the spatial boundary of the hydrogen explosion venting area. The area within the boundary is regarded as a high-risk area that needs to be monitored or avoided with emphasis, and is dynamically adjusted according to factors such as wind direction and climate.

[0071] In this embodiment, step S12 is specifically as follows: The robot acquires real-time temperature distribution data, flame intensity data, and smoke concentration data, and all are normalized to obtain initial fire data; the fire data is subjected to data fusion through a preset second algorithm to obtain the fire location coordinates; wherein, the second algorithm includes a preset neural network algorithm and a Bayesian network model.

[0072] In specific implementation, the robot includes a data acquisition module that periodically receives data streams from a thermal imaging camera (outputting temperature distribution data), a flame detector (outputting a flame presence / intensity signal), and a smoke sensor (outputting smoke concentration data) at a sampling frequency not lower than 10 Hz. The data acquisition module attaches a timestamp to each group of sensor data based on a unified high-precision clock source (with a precision better than ±1 millisecond). All timestamps are recorded in Coordinated Universal Time (UTC) format to ensure that data from different sensors can achieve time alignment and synchronous analysis in the subsequent data fusion processing stage.

[0073] Further, before performing multi-source data fusion, the robot standardizes the original sensor data through a preprocessing module. The preprocessing module first applies the Kalman filter algorithm to each sensor data to filter out random noise caused by environmental interference and measurement errors, and improve data accuracy. Subsequently, the temperature data (unit: °C), flame intensity data (normalized to 0 - 1), and smoke concentration data (unit: μg / m 3 ) output by different sensors are normalized so that each data item is within a unified dimension and standard range (such as linearly normalized to the [0, 1] interval), ensuring compatibility with the input format requirements of subsequent fusion algorithms based on weighted average and feature extraction.

[0074] Further, the fire data is fused through a preset second algorithm to obtain the fire location coordinates, specifically as follows: The fire recognition module in the robot uses a machine learning model based on a convolutional neural network (CNN) for thermal imaging image analysis. The CNN model consists of multiple convolutional layers, pooling layers, and fully connected layers, and the input is a two-dimensional temperature distribution map collected by a thermal imaging camera. The CNN extracts the regional features of areas with high temperatures (above 150 °C) and continuous flame contours in the image through supervised learning with an annotated data set containing fire and non-fire samples during the training phase. The recognition module determines the location and size (expressed in area m 2 2) and intensity (expressed by the highest temperature and the integrated temperature of the region) of the fire area based on the classification probability (probability of fire presence) output by the CNN and the region segmentation mask. In addition, the recognition result is fused with the detection results of the flame detector and the smoke sensor to further improve the accuracy and robustness of fire recognition. Then, the fire detection module in the robot system uses a Bayesian network model for inference. The Bayesian network includes multiple nodes, and each node corresponds to an input feature, including the abnormal temperature distribution collected by the thermal imaging camera, the flame intensity signal output by the flame detector, and the particle concentration value detected by the smoke sensor. The directed edges between the nodes represent the conditional dependence relationships between the features. The inference module calculates the posterior probability of the presence of fire based on the input features. When this probability is higher than a preset threshold (such as 90%), it is determined that a fire exists, and the spatial coordinate position (such as three-dimensional coordinates X, Y, Z) where the fire occurs is determined through maximum a posteriori estimation (MAP estimation) in combination with the spatial position information of each sensor.

[0075] In this embodiment, step S13 is specifically: spatially aligning and fusing the hydrogen explosion relief area and the fire location coordinates to obtain comprehensive on-site data; the robot plans an emergency response path according to the comprehensive on-site data through a preset path planning algorithm, a preset genetic algorithm, and a preset reinforcement learning algorithm.

[0076] Optionally, the path planning algorithm provided by the embodiments of the present invention is the A* algorithm. When used for path planning, by introducing a comprehensive cost function, it not only considers the traditional spatial distance but also combines multiple risk factors, including the fire hazard level and the risk of hydrogen leakage and explosion area. Specifically, the fire hazard level is quantified by the distance from the fire source and the temperature value. The closer to the fire source and the higher the temperature, the greater the corresponding cost value. For example, for every 1-meter increase in the distance from the fire source, the cost increases by 10 units; for every 1-degree Celsius increase in temperature, the cost increases by 5 units. The risk of the hydrogen leakage area is evaluated by the distance from the leakage source and the leakage concentration. For every 1-meter increase in distance and every 1-ppm increase in leakage concentration, the costs increase by 15 units and 3 units respectively. The A* algorithm dynamically adjusts the cost values of each cell, enabling the robot to preferentially select a safe path that avoids fire and hydrogen leakage areas during path planning, while still ensuring the shortest and most efficient path. This method ensures that in a complex environment, the robot can find the best path that is both short and safe while ensuring safety.

[0077] Optionally, the genetic algorithm (GA) is used to optimize the path based on multiple criteria (such as time, temperature exposure, and risk). GA can explore a wider range of potential paths and find solutions that balance multiple conflicting objectives, such as minimizing travel time and exposure to hazards. By simulating the process of natural selection, GA can iteratively improve a set of potential paths and finally converge to a near-optimal solution that takes into account all defined constraints.

[0078] Optionally, the reinforcement learning (RL) algorithm is used to learn the optimal path through trial and error in a dynamic environment. RL enables the robot to adjust its path planning strategy according to the real-time feedback of the environment, thus obtaining a more efficient and safe route in unpredictable situations. By receiving rewards or punishments for its actions, the RL agent can learn to associate certain environmental conditions with the best path planning decisions.

[0079] In this embodiment, step S14 is specifically as follows: The robot arrives at the fire scene according to the emergency response path and real-time obtains on-site temperature data, on-site flame intensity data, and on-site gas concentration data; performs data fusion on the on-site temperature, the on-site flame intensity data, and the on-site gas concentration data to obtain on-site environmental data; determines the on-site fire situation according to the on-site environmental data through a preset spatio-temporal convolutional neural network model and a preset inference system; predicts the fire development trend according to the on-site environmental data through an algorithm that mixes a preset physical diffusion model and a machine learning regression model; determines the fire extinguishing strategy according to the on-site fire situation and the fire development trend, and performs fire extinguishing actions according to the fire extinguishing strategy; wherein the fire extinguishing strategy includes selecting the fire extinguishing position, determining the type of fire extinguishing agent, determining the spraying flow rate of the fire extinguishing agent, and selecting the spraying angle of the fire extinguishing agent.

[0080] In a specific implementation, the system continuously monitors environmental parameters by deploying multiple sensors, including: a temperature sensor that collects environmental temperature data once per second, with a measurement range of -40°C to +150°C and an accuracy of ±0.5°C; a flame detector and a thermal imaging camera that detect flame intensity. The response time of the flame detector is less than 1 second, and it can detect open flames above 0.1m 2 The above open flames, the resolution of the thermal imaging camera is not less than 640×480 pixels, the temperature measurement range is -20°C to +500°C, and the temperature resolution is better than 0.1°C; a hydrogen sensor is used to monitor the hydrogen concentration, with a detection range of 0 - 10,000 ppm, a sensitivity error not exceeding ±2%, and a sampling frequency of 2 times per second; in addition, an oxygen sensor is configured to monitor the oxygen content, with a range of 0 - 25%VOL and an accuracy of ±0.1%VOL, and specific gas sensors such as carbon monoxide and methane are selected according to application requirements. The detection ranges and accuracies of the specific gas sensors are 0 - 1000 ppm and the error is less than ±5% respectively according to the gas types. The system updates the environmental status in real time through data collection and processing at least once per second, providing high-precision data support for path planning and risk assessment.

[0081] Furthermore, the data fusion of the on-site temperature, the on-site flame intensity data, and the on-site gas concentration data to obtain on-site environmental data is specifically as follows: fuse real-time sensor data to comprehensively understand the current situation. Construct a dynamic map that includes the location and intensity of the fire, the hydrogen explosion venting area, the temperature gradient, and other relevant environmental information. The system creates an environmental map in real time by integrating multi-source sensor data, specifically including: constructing a two-dimensional or three-dimensional spatial structure based on a lidar (360° scan, ranging from 0.1 meter to 100 meters, accuracy ±2 centimeters, refresh rate 10Hz); calibrating the location and intensity of the fire heat source in combination with a thermal imaging camera (resolution 640×480 pixels, temperature measurement range -20°C to 500°C, sensitivity better than 0.1°C), with the fire source location positioning error less than ±0.5 meters and the intensity quantified in the range of 0 - 100%; a hydrogen sensor (detection range 0 - 10,000 ppm, error ±2%) identifies the location of the leakage source, and when the hydrogen concentration is detected to be higher than 200 ppm for 3 consecutive seconds, a high-risk area is automatically marked; the concentrations of gases such as oxygen and carbon monoxide in the environment are dynamically updated with air quality information through dedicated sensors (sampling frequency ≥1Hz, error less than ±5%). All data is updated at least once per second, and data fusion is performed through an Extended Kalman Filter (EKF) or a graph optimization-based method to generate a real-time environmental map with a resolution not less than 0.1 meter × 0.1 meter. Based on this real-time environmental map, the robot completes local environmental modeling and risk assessment within 500 milliseconds, providing complete, accurate, and real-time data support for fire extinguishing decisions (such as path planning, fire extinguishing agent release, and priority target selection).

[0082] Furthermore, through a preset spatio-temporal convolutional neural network model and a preset inference system, the on-site fire situation is determined based on the on-site environmental data, and the fire development trend is predicted based on the on-site environmental data through an algorithm that mixes a preset physical diffusion model and a machine learning regression model. Specifically: The system uses a machine learning model (such as the spatio-temporal convolutional neural network ST-CNN trained based on 50,000 sets of historical fire and gas leakage data, with a prediction accuracy of ≥92%) or a rule-based inference system to predict the spread and intensity change of the fire. The fire prediction model takes the current environmental temperature (-40°C to +500°C, updated per second), flame intensity (quantified from 0 to 100%), wind speed and direction (measured by an airflow sensor, wind speed range 0-20 m / s, accuracy ±0.5 m / s, wind direction error less than ±5°) as inputs, combines with the historical fire expansion pattern to infer the fire source expansion direction and speed within the next 5 minutes (prediction time step 30 seconds, spatial resolution 0.5 meters), and marks the high-risk areas. The hydrogen deflagration expansion prediction is based on the current hydrogen concentration (0-10,000 ppm, sampled per second), airflow pattern and environmental temperature, and applies a hybrid method based on a physical diffusion model and machine learning regression (such as an XGBoost regressor, mean squared error MSE less than 0.01) to predict the hydrogen concentration change area within the next 1 minute, mark the dangerous areas where the explosion limit (range of 4,000-75,000 ppm) may be reached, and control the error range within ±10%. The system updates the prediction results once per second, and superimposes the potential fire spread path and hydrogen expansion area on the real-time environmental map, enabling the robot to actively adjust the fire extinguishing strategy based on the risk evolution within the next 30 seconds to 5 minutes, including preferentially deploying fire extinguishing agent spraying to block the fire spread, adjusting the movement path to avoid the predicted high-risk areas, and optimizing the fire extinguishing resource allocation.

[0083] Furthermore, based on the evaluated environmental situation and the predicted development trend, the robot selects an appropriate fire extinguishing strategy. The factors affecting the decision-making include the type and intensity of the fire, the distance from the hydrogen deflagration area, the availability of different fire extinguishing agents (such as water, foam, carbon dioxide), environmental conditions (such as the presence of sensitive equipment), the size and accessibility of the fire. The decision-making process may involve decision trees, fuzzy logic systems or machine learning classifiers trained based on various fire scenarios and their optimal fire extinguishing strategies.

[0084] In this embodiment, while the robot is monitoring the on-site environmental data in real time to perform fire extinguishing operations, it will also conduct real-time fire extinguishing effect evaluation and autonomously change the fire extinguishing strategy according to the evaluation results. Among them, the specific process of conducting real-time fire extinguishing effect evaluation and autonomously changing the fire extinguishing strategy according to the evaluation results is as follows: While the robot is extinguishing the fire, it will collect on-site temperature data, flame intensity data, and thermal imaging data in real time, and compare each data with the corresponding predicted data calculated by the preset fire spread prediction model to obtain the fire extinguishing effect evaluation result. If the error between each data and the corresponding predicted data is less than the preset percentage, the evaluation result is valid, and the original fire extinguishing strategy is maintained; otherwise, the evaluation result is invalid, and the fire extinguishing strategy is adjusted. The adjustment of the fire extinguishing strategy includes increasing the fire extinguishing agent spraying flow rate, changing the spraying angle of the fire extinguishing agent, and optimizing the fire extinguishing movement route.

[0085] In a specific implementation, the robot receives real-time feedback from sensors monitoring the fire (e.g., temperature sensors near the fire source, thermal imaging images, flame intensity measurements). It monitors the spread of the fire or any changes in intensity. It may also be necessary to monitor the concentration of hydrogen in the vicinity to ensure that the fire extinguishing work does not increase the risk of deflagration.

[0086] Furthermore, the system evaluates the effectiveness of the current fire extinguishing strategy by collecting and analyzing sensor data in real time. Specifically, it includes: a temperature sensor (measurement range -40°C to +150°C, accuracy ±0.5°C, sampling frequency ≥1Hz) is used to monitor the temperature change in the fire extinguishing area. If the average temperature drop is ≥10°C within 30 consecutive seconds, it is determined that the temperature reduction is effective; a flame detector (response time <1 second, flame detection sensitivity covering the above fire source) is used to measure the change in flame intensity. If the flame signal intensity drops ≥20%, it is considered that the flame intensity has weakened; the fire area is calculated through a thermal imaging camera (resolution 640×480 pixels, temperature resolution better than 0.1°C) and an image processing algorithm (fire area segmentation accuracy rate ≥95%). If the fire area shrinks ≥15% within 1 minute after fire extinguishing, the area change meets the standard. The system synchronously predicts the fire extinguishing effect based on the fire spread prediction model established before fire extinguishing (such as based on the spatio-temporal convolutional network ST-CNN, prediction accuracy ≥92%) and compares the actual observed data with the predicted values. When the errors in the actual temperature distribution, fire source range, and prediction results are all less than ±10%, the strategy is considered effective; if the temperature drop <5°C, the flame intensity drop <10%, the fire area does not shrink significantly, and the deviation from the prediction result >15% within 30 consecutive seconds, the system automatically triggers the strategy adjustment mechanism, including increasing the fire extinguishing agent spraying flow rate (adjustment range +0.5L / min each time, range 0 - 10L / min), changing the spraying angle (adjustment of 5° each time), or optimizing the movement path to dynamically optimize the fire extinguishing effect. 2 As above, the system evaluates the effectiveness of the current fire extinguishing strategy by collecting and analyzing sensor data in real time. Specifically, it includes: a temperature sensor (measurement range -40°C to +150°C, accuracy ±0.5°C, sampling frequency ≥1Hz) is used to monitor the temperature change in the fire extinguishing area. If the average temperature drop is ≥10°C within 30 consecutive seconds, it is determined that the temperature reduction is effective; a flame detector (response time <1 second, flame detection sensitivity covering the above fire source) is used to measure the change in flame intensity. If the flame signal intensity drops ≥20%, it is considered that the flame intensity has weakened; the fire area is calculated through a thermal imaging camera (resolution 640×480 pixels, temperature resolution better than 0.1°C) and an image processing algorithm (fire area segmentation accuracy rate ≥95%). If the fire area shrinks ≥15% within 1 minute after fire extinguishing, the area change meets the standard. The system synchronously predicts the fire extinguishing effect based on the fire spread prediction model established before fire extinguishing (such as based on the spatio-temporal convolutional network ST-CNN, prediction accuracy ≥92%) and compares the actual observed data with the predicted values. When the errors in the actual temperature distribution, fire source range, and prediction results are all less than ±10%, the strategy is considered effective; if the temperature drop <5°C, the flame intensity drop <10%, the fire area does not shrink significantly, and the deviation from the prediction result >15% within 30 consecutive seconds, the system automatically triggers the strategy adjustment mechanism, including increasing the fire extinguishing agent spraying flow rate (adjustment range +0.5L / min each time, range 0 - 10L / min), changing the spraying angle (adjustment of 5° each time), or optimizing the movement path to dynamically optimize the fire extinguishing effect.

[0087] Furthermore, if the fire extinguishing strategy fails to achieve the expected effect, the robot will autonomously adjust its method. Possible adjustments include changing the type or quantity of the fire extinguishing agent used, adjusting the spraying direction or pattern of the fire extinguishing agent (e.g., aiming the water flow at the base of the flame, using a wider spray pattern), and moving to a different advantageous position to better approach the fire source. The decision to adjust the strategy may be based on predefined rules, fuzzy logic, or a machine learning model trained to optimize real-time fire extinguishing. There are three ways to determine the basis for adjusting the strategy: First, based on predefined rules, when the temperature sensor detects that the local temperature exceeds 80°C, the hydrogen concentration is higher than 500 ppm, or the flame detector detects a flame signal for more than 3 seconds continuously, immediately switch to the high-priority fire extinguishing mode; Second, based on a fuzzy logic inference system, input variables such as temperature (-40°C to +150°C), flame intensity (quantified from 0 - 100%), and hydrogen concentration (0 - 10,000 ppm), the fuzzy membership function is divided into three levels: low, medium, and high. The inference system outputs the fire extinguishing agent release rate (unit: L / min, range 0 - 10 L / min) and the adjustment ratio of the moving speed (0 - 100%); Third, based on a trained machine learning model (such as an XGBoost classifier trained with 10,000 sets of fire scenario data, with an accuracy greater than 95%), according to the real-time sensor input (sampling frequency ≥ 1 Hz per second), predict the best fire extinguishing strategy, including fire extinguishing agent selection, spraying angle (0° - 90°, error not exceeding ±2°), adjustment of the moving trajectory, and the priority processing area. The system takes no more than 100 milliseconds for each prediction decision, ensuring real-time response and decision accuracy during the fire extinguishing process.

[0088] In this embodiment, the execution of the fire extinguishing action further includes the robot's real-time acquisition of the fire extinguishing agent liquid level data. When the liquid level data is lower than a preset second threshold, the fire extinguishing agent replenishment strategy is executed. Specifically, when executing the fire extinguishing agent replenishment strategy, the robot accesses a preset high-precision map containing the positions of the fire extinguishing agent replenishment stations, and through a preset path planning algorithm, plans a safe path to the nearest replenishment station; according to the safe path, reaches the nearest replenishment station, replenishes the fire extinguishing agent and real-time acquires the fire extinguishing agent liquid level data. When the liquid level data exceeds a preset third threshold, stops the fire extinguishing agent replenishment; accesses the preset real-time fire situation map, acquires the first fire location data, and plans a return path through a preset path planning algorithm; according to the return path, reaches the first fire location and continues the fire extinguishing action.

[0089] In a specific implementation, the system continuously monitors the extinguishing agent liquid level at a frequency of 1 Hz through a liquid level sensor (measurement range 0 - 100%, measurement accuracy ±1%) or a pressure sensor (measurement range 0 - 10 bar, accuracy ±0.1 bar) installed in the extinguishing agent storage tank. When it detects that the extinguishing agent liquid level is lower than the set minimum threshold (usually set to 20% of the total capacity, for example, the liquid level in a 10 - liter storage tank is lower than 2 liters), the system determines within 1 second that a replenishment operation needs to be performed.

[0090] Furthermore, when the robot decides to replenish, it accesses a high - precision map (map resolution ≤ 0.1 m) pre - stored with the locations of all replenishment stations and uses the weighted A* path planning algorithm (setting high weights considering obstacles, fire source high - temperature areas > 80°C, and high - risk hydrogen leakage areas > 1000 ppm) to calculate the optimal safe route to the nearest effective replenishment station within 500 milliseconds, with a path update frequency ≥ 1 Hz. After arriving at the replenishment station, it completes the replenishment of the extinguishing agent through an automatic docking system at a flow rate ≥ 5 L / min, with the replenishment accuracy controlled within ±0.1 L, and the replenishment operation (including interface calibration and liquid injection) completed within no more than 60 seconds.

[0091] Furthermore, after the replenishment is completed, the system recalculates and plans the optimal path to return to the fire - fighting operation area based on the real - time fire situation map (combining thermal imaging and gas detection data, update frequency ≥ 1 Hz), giving priority to going to the area with the most concentrated fire or the highest risk of spread. The overall replenishment and return process ensures that the total downtime of the robot is controlled within 2 minutes, maximizing the continuity of the fire - fighting task and the operation efficiency.

[0092] It should be noted that the system is equipped with an artificial takeover mechanism. When the robot detects abnormal situations that it cannot handle (such as the environmental temperature exceeding 120°C, hydrogen concentration higher than 5000 ppm, or the path being completely blocked for more than 30 seconds), it automatically sends an abnormal alarm to the operator through an encrypted communication channel (AES - 256 - bit encryption) and waits for an artificial instruction to take over within 5 seconds. In the manual control mode, the operator can view the thermal imaging screen (resolution 640×480, update frequency ≥ 10 fps), the environmental map (update frequency ≥ 1 Hz) in real - time through a remote terminal, and remotely issue motion control instructions (delay ≤ 100 milliseconds) and extinguishing agent release control instructions (control accuracy 0.1 L / min). The communication link has dual redundancy (5G + Wi - Fi 6) to ensure that when a single link fails, the standby link automatically switches within 2 seconds, guaranteeing uninterrupted communication and improving the human - machine collaboration ability and operation safety of the robot in complex emergency environments.

[0093] An emergency response method for a robot facing a hydrogen deflagration scenario provided by an embodiment of the present invention provides data support for subsequent planning of an emergency response route for the robot to reach the fire scene by determining the hydrogen explosion relief area and marking the fire location; by reasonably planning the route for the robot to reach the fire scene according to the hydrogen explosion relief area and the fire location coordinates, the efficiency of the robot to reach the fire scene is improved; by real-time monitoring of on-site environmental data to perform fire extinguishing actions, it can ensure that the fire extinguishing actions performed achieve the fire extinguishing effect. In this application, from obtaining on-site data to planning the emergency response path and then to performing fire extinguishing actions, all are full-process operations carried out by the robot without manual intervention. Compared with the prior art, using the robot to replace rescue personnel can greatly reduce the probability of personnel injury when rescue personnel take emergency response measures in a hydrogen deflagration scenario.

[0094] Embodiment 2:

[0095] As Figure 2 shown, this embodiment provides a robot emergency response device for a hydrogen deflagration scenario, including an explosion relief area determination module 001, a fire location acquisition module 002, a response path planning module 003, and a fire extinguishing action execution module 004. Among them,

[0096] The explosion relief area determination module 001 is used for the robot to obtain real-time hydrogen concentration data and perform concentration analysis. If the result of the concentration analysis is abnormal, the hydrogen explosion relief area is determined according to the hydrogen concentration data and a preset real-time map through a preset first algorithm;

[0097] The fire location acquisition module 002 is used for the robot to obtain real-time temperature data, fire intensity data, and smoke concentration data, and obtain the fire location coordinates through a preset second algorithm;

[0098] The response path planning module 003 is used for the robot to plan an emergency response path according to the hydrogen explosion relief area and the fire location coordinates;

[0099] The fire extinguishing action execution module 004 is used for the robot to reach the fire scene according to the emergency response path and real-time monitor the on-site environmental data to perform fire extinguishing actions.

[0100] In this embodiment, the process by which the explosion venting area determination module 001 determines the hydrogen explosion venting area is as follows: The explosion venting area determination module 001 obtains real-time hydrogen concentration data and analyzes whether the concentration data exceeds a preset first threshold; if the concentration data exceeds the first threshold, the result of the concentration analysis is abnormal, and the robot will analyze the hydrogen concentration distribution based on the hydrogen concentration data and a preset real-time map through a preset first algorithm, and superimpose a gas concentration distribution map on the real-time map according to the analysis result of the hydrogen concentration distribution; wherein, the first algorithm includes a preset spatial interpolation algorithm and a Gaussian mixture model; according to the gas concentration distribution map and a preset safety regulation, the hydrogen explosion venting area is determined through a preset boundary detection algorithm.

[0101] In this embodiment, the process by which the fire location acquisition module 002 obtains the fire location coordinates is as follows: The fire location acquisition module 002 obtains real-time temperature distribution data, flame intensity data, and smoke concentration data, and normalizes all of them to obtain initial fire data; the fire location coordinates are obtained through data fusion of the fire data by a preset second algorithm; wherein, the second algorithm includes a preset neural network algorithm and a Bayesian network model.

[0102] In this embodiment, the response path planning module 003 performs emergency response path planning based on the hydrogen explosion venting area and the fire location coordinates, specifically as follows: The response path planning module 003 spatially aligns and fuses the hydrogen explosion venting area and the fire location coordinates to obtain comprehensive on-site data; the robot plans an emergency response path based on the comprehensive on-site data through a preset path planning algorithm, a preset genetic algorithm, and a preset reinforcement learning algorithm.

[0103] In this embodiment, the fire extinguishing action execution module 004 arrives at the fire scene according to the emergency response path and monitors the on-site environmental data in real time to control the robot to execute fire extinguishing actions, specifically as follows: The fire extinguishing action execution module 004 arrives at the fire scene according to the emergency response path and obtains on-site temperature data, on-site flame intensity data, and on-site gas concentration data in real time; the on-site environmental data is obtained through data fusion of the on-site temperature, the on-site flame intensity data, and the on-site gas concentration data; the on-site fire situation is determined based on the on-site environmental data through a preset spatio-temporal convolutional neural network model and a preset inference system; the fire development trend is predicted based on the on-site environmental data through an algorithm that mixes a preset physical diffusion model and a machine learning regression model; the fire extinguishing strategy is determined based on the on-site fire situation and the fire development trend, and the robot is controlled to execute fire extinguishing actions according to the fire extinguishing strategy; wherein, the fire extinguishing strategy includes selecting a fire extinguishing position, determining the type of fire extinguishing agent, determining the spraying flow rate of the fire extinguishing agent, and selecting the spraying angle of the fire extinguishing agent.

[0104] For a more detailed working principle and step - by - step process of this embodiment, reference can be made, but not limited to, the relevant descriptions in Embodiment 1.

[0105] An embodiment of the present invention provides a robot emergency response device for a hydrogen deflagration scenario. The hydrogen explosion relief area determination module 001 determines the hydrogen explosion relief area, providing accurate data support for subsequent planning of the emergency response route for the robot to reach the fire site; the fire location acquisition module 002 acquires the accurate fire location, providing accurate data support for subsequent planning of the emergency response route for the robot to reach the fire site; the response path planning module 003 accurately plans the emergency response path for the robot to carry out fire rescue according to the hydrogen explosion relief area and the fire location, ensuring the safety and efficiency on the robot's response path; the fire - extinguishing action execution module 004 controls the robot to execute the fire - extinguishing action, thus replacing the situation where rescue personnel reach the hydrogen deflagration site for rescue in the traditional method, thereby reducing the probability of personnel injury during the rescue process.

[0106] Embodiment Three:

[0107] This embodiment provides a terminal device, including: a processor, a memory, a communication interface, and a communication bus. The processor, the memory, and the communication interface complete mutual communication through the communication bus;

[0108] The memory is used to store at least one executable instruction, and the executable instruction causes the processor to execute the operations of the robot emergency response method for a hydrogen deflagration scenario as described in any one of the above.

[0109] Embodiment Four:

[0110] An embodiment of the present invention provides a computer - readable storage medium. The computer - readable storage medium includes a stored computer program, wherein when the computer program runs, it controls the device / device where the computer - readable storage medium is located to execute the robot emergency response method for a hydrogen deflagration scenario as described in any one of the above.

[0111] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above - mentioned embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer - readable storage medium. When the program is executed, it can include the processes of the embodiments of the above - mentioned methods. Among them, the storage medium can be a magnetic disk, an optical disk, a Read - Only Memory (ROM), or a Random Access Memory (RAM), etc.

[0112] The specific embodiments described above further elaborate on the object, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only for the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. It is particularly pointed out that for those skilled in the art, any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A robot emergency response method for hydrogen deflagration scenarios, characterized in that, Applicable to a robot, the robot emergency response method includes: The robot acquires real-time hydrogen concentration data and conducts concentration analysis. If the result of the concentration analysis is abnormal, the hydrogen explosion venting area is determined through a preset first algorithm based on the hydrogen concentration data and a preset real-time map; The robot acquires real-time temperature distribution data, fire intensity data, and smoke concentration data, and obtains the fire location coordinates through a preset second algorithm; The robot conducts emergency response path planning based on the hydrogen explosion venting area and the fire location coordinates; The robot reaches the fire scene according to the emergency response path and monitors the on-site environmental data in real time to perform fire extinguishing actions.

2. The robot emergency response method for a hydrogen deflagration scenario according to claim 1, wherein, The robot acquires real-time hydrogen concentration data and conducts concentration analysis. If the result of the concentration analysis is abnormal, the hydrogen explosion venting area is determined through a preset first algorithm based on the hydrogen concentration data and a preset real-time map, specifically: The robot acquires real-time hydrogen concentration data and analyzes whether the concentration data exceeds a preset first threshold; If the concentration data exceeds the first threshold, the result of the concentration analysis is abnormal. The robot will analyze the hydrogen concentration distribution through a preset first algorithm based on the hydrogen concentration data and a preset real-time map, and superimpose a gas concentration distribution map on the real-time map according to the hydrogen concentration distribution analysis result; wherein, the first algorithm includes a preset spatial interpolation algorithm and a Gaussian mixture model; The hydrogen explosion venting area is determined through a preset boundary detection algorithm according to the gas concentration distribution map and a preset safety regulation.

3. The robot emergency response method for a hydrogen deflagration scenario according to claim 1, characterized in that, The robot acquires real-time temperature distribution data, fire intensity data, and smoke concentration data, and obtains the fire location coordinates through a preset second algorithm, specifically: The robot acquires real-time temperature distribution data, flame intensity data, and smoke concentration data, and all are normalized to obtain initial fire data; The fire data is data-fused through a preset second algorithm to obtain the fire location coordinates; wherein, the second algorithm includes a preset neural network algorithm and a Bayesian network model.

4. The robot emergency response method for a hydrogen deflagration scenario according to claim 1, wherein, The robot conducts emergency response path planning based on the hydrogen explosion venting area and the fire location coordinates, specifically: The hydrogen explosion venting area and the fire location coordinates are spatially aligned and fused to obtain comprehensive on-site data; The robot plans an emergency response path based on the comprehensive on-site data through a preset path planning algorithm, a preset genetic algorithm, and a preset reinforcement learning algorithm.

5. The robot emergency response method for a hydrogen deflagration scenario according to claim 1, characterized in that The robot reaches the fire scene according to the emergency response path and monitors the on-site environmental data in real time to perform fire extinguishing actions, specifically: The robot reaches the fire scene according to the emergency response path and acquires on-site temperature data, on-site flame intensity data, and on-site gas concentration data in real time; The on-site temperature, on-site flame intensity data, and on-site gas concentration data are data-fused to obtain on-site environmental data; The on-site fire situation is determined according to the on-site environmental data through a preset spatio-temporal convolutional neural network model and a preset inference system; Predict the fire development trend based on the on-site environmental data through an algorithm that mixes a preset physical diffusion model and a machine learning regression model; Determine the fire extinguishing strategy based on the on-site fire situation and the fire development trend, and execute the fire extinguishing action according to the fire extinguishing strategy; wherein, the fire extinguishing strategy includes selecting the fire extinguishing position, determining the type of fire extinguishing agent, determining the spraying flow rate of the fire extinguishing agent, and selecting the spraying angle of the fire extinguishing agent.

6. The robot emergency response method for a hydrogen deflagration scenario according to claim 1, wherein While the robot is monitoring the on-site environmental data in real time to execute the fire extinguishing action, it will also conduct real-time fire extinguishing effect evaluation and autonomously change the fire extinguishing strategy according to the evaluation result. Among them, the specific implementation of conducting real-time fire extinguishing effect evaluation and autonomously changing the fire extinguishing strategy is as follows: While the robot is extinguishing the fire, it collects the on-site temperature data, flame intensity data, and thermal imaging data in real time, and compares each data with the corresponding predicted data calculated by a preset fire spread prediction model to obtain the fire extinguishing effect evaluation result; If the error between each data and the corresponding predicted data is less than the preset percentage, the evaluation result is valid, and the original fire extinguishing strategy is maintained; Otherwise, the evaluation result is invalid, and the fire extinguishing strategy is adjusted; the adjustment of the fire extinguishing strategy includes increasing the spraying flow rate of the fire extinguishing agent, changing the spraying angle of the fire extinguishing agent, and optimizing the fire extinguishing movement route.

7. The robot emergency response method for a hydrogen deflagration scenario according to claim 1, wherein The execution of the fire extinguishing action also includes the robot obtaining the liquid level data of the fire extinguishing agent in real time. When the liquid level data is lower than the preset second threshold, the fire extinguishing agent replenishment strategy is executed. Among them, the specific implementation of the execution of the fire extinguishing agent replenishment strategy is as follows: The robot accesses a preset high-precision map containing the positions of fire extinguishing agent replenishment stations, and plans a safe path to the nearest replenishment station through a preset path planning algorithm; According to the safe path, reach the nearest replenishment station, replenish the fire extinguishing agent and obtain the liquid level data of the fire extinguishing agent in real time. When the liquid level data exceeds the preset third threshold, stop replenishing the fire extinguishing agent; Access the preset real-time fire situation map to obtain the first fire location data, and plan the return path through a preset path planning algorithm; According to the return path, reach the first fire location and continue the fire extinguishing action.

8. A robot emergency response device for hydrogen deflagration scenarios, characterized in that, It includes an explosion venting area determination module, a fire location acquisition module, a response path planning module, and a fire extinguishing action execution module. Among them, The explosion venting area determination module is used for the robot to obtain the real-time hydrogen concentration data and conduct concentration analysis. If the result of the concentration analysis is abnormal, the hydrogen explosion venting area is determined through a preset first algorithm based on the hydrogen concentration data and a preset real-time map; The fire location acquisition module is used for the robot to obtain the real-time temperature data, fire intensity data, and smoke concentration data, and obtain the fire location coordinates through a preset second algorithm; The response path planning module is used for the robot to conduct emergency response path planning according to the hydrogen explosion venting area and the fire location coordinates; The fire extinguishing action execution module is used for the robot to reach the fire scene according to the emergency response path and monitor the on-site environmental data in real time to execute the fire extinguishing action.

9. A terminal device, characterized in that, It includes: A processor, a memory, a communication interface, and a communication bus. The processor, the memory, and the communication interface complete communication with each other through the communication bus; The memory is used to store at least one executable instruction, and the executable instruction causes the processor to execute the operations of the robot emergency response method for a hydrogen deflagration scenario according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein when the computer program runs, it controls the device / equipment where the computer-readable storage medium is located to execute the robot emergency response method for a hydrogen deflagration scenario according to any one of claims 1 to 7.

Images