Low-power real-time fire emergency rescue method and system based on edge computing

Abstract

The present invention discloses a low-power real-time auxiliary first aid method and system for fire based on edge computing, which relates to the field of auxiliary first aid technology for fire, and is applicable to crowded areas of public buildings such as schools, hospitals, high-rise buildings, hotels, and shopping malls. The present invention uses multi-sensor data fusion to collect environmental information such as population density, temperature, and smoke concentration in real time, and uses edge computing units for fire detection and dynamic evacuation path planning. The intelligent terminal adjusts the brightness and color based on the dynamic path, and uses sound reminders to achieve clear audio-visual guidance. In the event of a power outage due to a fire, the lighting strategy of each area is dynamically adjusted through optimized lighting strategies to reduce energy consumption, effectively extend the use time of the backup power supply, and effectively ensure the safety and efficiency of personnel evacuation.

Description

Technical Field

[0001] The present application relates to the field of fire assisted emergency rescue technology, and in particular to a low-power real-time fire assisted emergency rescue method and system based on edge computing. Background Art

[0002] Traditional fire emergency lighting systems usually use fixed lighting modes or preset scene switching modes. These systems can only trigger simple lighting changes when a fire is detected, such as starting emergency lights or turning on preset evacuation indicators. However, when faced with complex environments in crowded areas (such as theaters and high-rise buildings), existing systems often find it difficult to respond quickly to changes in population density, and evacuation instructions are difficult to adjust dynamically. The lack of clear audio-visual guidance may lead to reduced evacuation efficiency and increased safety risks for personnel.

[0003] For example, the Chinese patent with the authorization announcement number CN117939761B discloses a method for emergency evacuation lighting through path-guiding light, and dynamic path planning through the fitting of abnormal planning network and emergency response information, but there are still certain limitations in practical applications. This method relies on preset regional features and historical data models for planning and adjustment. When faced with sudden fires, it may not be able to respond quickly to real-time changing environments, resulting in lags in evacuation guidance. Secondly, although the solution introduces a variety of sensors and image acquisition devices, its processing and comprehensive analysis capabilities for sensor data are limited, especially when processing complex, multi-dimensional real-time data, it is difficult to effectively deal with uncertainty and complex fire scenes. In addition, the lighting control strategy of the solution is relatively simple, mainly focusing on the activation and control of path-guiding light, lacking fine-tuning of multiple dimensions, and it is difficult to provide the best evacuation guidance according to changes in the actual fire environment.

[0004] The above methods have obvious deficiencies in terms of intelligence level, dynamic response capability and low-power operation, and cannot effectively improve evacuation efficiency and safety in places with high-density crowds. Summary of the invention

[0005] In response to the deficiencies in the prior art, the present application discloses a low-power, real-time fire auxiliary first aid method and system based on edge computing.

[0006] In the first aspect, the present application provides a low-power fire real-time auxiliary first aid method based on edge computing, including:

[0007] Receive and collect environmental data using a sensor interface, the environmental data including: temperature data, smoke concentration data, personnel data, and ambient light intensity data; wherein the personnel data includes: personnel density data and personnel movement rate;

[0008] Based on the environmental data, using a preset fire detection algorithm, determining in the edge computing unit whether a fire event has occurred; in response to detecting the occurrence of the fire event, generating a dynamic evacuation path in real time based on the location of the fire source and the layout of the building;

[0009] Based on the evacuation path, a control signal is sent to the control module using the input / output interface, and the control module controls the intelligent terminal to issue auxiliary first aid information based on the control signal; wherein the auxiliary first aid information is used to guide personnel to evacuate; the auxiliary first aid information includes: visual information and sound information.

[0010] As an optional implementation manner, the determining whether a fire event occurs includes:

[0011] Perform multi-sensor data fusion on the temperature data, smoke concentration data, personnel data, and ambient light intensity data to generate fused sensor data;

[0012] Dynamically adjust the fire judgment threshold based on real-time fusion sensor data and preset historical data model;

[0013] Processing the fused sensor data based on a fuzzy logic algorithm to generate a fire risk score;

[0014] In response to the fire risk score being greater than the fire judgment threshold, a fire event alarm is triggered and an evacuation route is dynamically generated.

[0015] As an optional implementation, the edge computing unit forms a distributed cluster system by multiple edge computing units;

[0016] The cluster system includes: each edge computing unit is responsible for processing the sensor data connected to it, and sharing the data with other edge computing units in real time through a communication network to perform distributed data fusion.

[0017] As an optional implementation, generating a fire risk score includes:

[0018] Defining fuzzy variables, including: temperature variable, smoke concentration variable, personnel variable and ambient light intensity variable, and setting a fuzzy membership function for each of the fuzzy variables;

[0019] Using fuzzy inference rules, mapping the fused sensor data to the fuzzy membership of each of the fuzzy variables;

[0020] Based on the fuzzy inference rules and in combination with the fuzzy membership of each of the fuzzy variables, a fuzzy fire risk score is calculated;

[0021] The fuzzy fire risk score is defuzzified to generate the fire risk score.

[0022] As an optional implementation, the triggering of a fire event alarm and the dynamic generation of an evacuation route include:

[0023] In response to the fire risk score exceeding the fire judgment threshold, generating a prediction model of the fire source location and the fire spread direction in real time;

[0024] Based on the fire source location and fire spread direction, generate and dynamically adjust the evacuation path;

[0025] Triggering multi-mode fire alarms, including: audible alarm, visual indication, and sending remote alarm signals;

[0026] The visual indication includes: light flashing and light color change; wherein the light flashing and the light color change are executed by a smart terminal.

[0027] As an optional implementation, the method further includes: the edge computing unit generates the control signal based on the evacuation path, and the control signal includes: a brightness adjustment instruction and a color change instruction;

[0028] The brightness adjustment instruction dynamically adjusts the brightness of the smart terminal based on the personnel density data and the personnel movement rate;

[0029] The color change instruction dynamically adjusts the color of the smart terminal based on the location of the fire source and the direction of fire spread.

[0030] As an optional implementation manner, the dynamically adjusting the brightness of the smart terminal further includes:

[0031] Dynamically adjust the brightness of the smart terminal based on the smoke concentration data and the ambient light intensity data;

[0032] The dynamically adjusting the brightness of the smart terminal further includes:

[0033] Predicting where people gather based on the people density data and the movement speed of the people;

[0034] In a first time period before the predicted gathering of people occurs at the gathering place of people, the brightness of the smart terminal is adjusted in advance.

[0035] As an optional implementation manner, the dynamically adjusting the color of the smart terminal includes:

[0036] Based on the relative position of the fire source location and the evacuation path, dynamically select different colors to indicate safe paths and dangerous areas;

[0037] In response to the fire spreading direction intersecting the evacuation route, adjusting the color of the smart terminal close to the fire source to a preset warning color;

[0038] Adjust the lighting color of the area far away from the fire source to the preset guide color;

[0039] Based on the fire spread speed and path length, the color change frequency and gradient effect of different areas are adjusted in real time.

[0040] As an optional implementation manner, adjusting the color of the smart terminal near the fire source to a preset warning color includes:

[0041] In response to detecting that the distance between the intersection of the fire spreading direction and the evacuation path is less than a preset safety distance, triggering the warning color adjustment;

[0042] The preset warning colors include: red, orange, and yellow; wherein red is used to indicate a high-risk area; the high-risk area includes: the intersection distance is less than or equal to the first distance threshold and the fire spread speed is greater than or equal to the first speed threshold;

[0043] Orange is used to indicate a moderately dangerous area; the moderately dangerous area includes: the intersection distance is greater than the first distance threshold and less than the second distance threshold, and the fire spread speed is less than the first speed threshold and greater than the second speed threshold;

[0044] Yellow is used to indicate low-risk areas; the low-risk areas include: the intersection distance is greater than or equal to the second distance threshold or the fire spread speed is less than or equal to the second speed threshold.

[0045] On the second aspect, the present application also provides a low-power fire real-time auxiliary emergency rescue system based on edge computing, including:

[0046] A collection unit, used to receive and collect environmental data using a sensor interface, wherein the environmental data includes: temperature data, smoke concentration data, personnel data, and ambient light intensity data; wherein the personnel data includes: personnel density data and personnel movement rate;

[0047] A judgment unit, configured to judge whether a fire event has occurred in the edge computing unit based on the environmental data and using a preset fire detection algorithm; in response to detecting the occurrence of the fire event, generate a dynamic evacuation path in real time based on the location of the fire source and the layout of the building;

[0048] The processing unit is used to send a control signal to the control module using the input / output interface based on the evacuation path, and the control module controls the intelligent terminal to issue auxiliary first aid information based on the control signal; wherein the auxiliary first aid information is used to guide personnel to evacuate; the auxiliary first aid information includes: visual information and sound information.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] This invention significantly improves the dynamic response capability of the fire emergency system in complex environments by introducing edge computing technology and multi-sensor data fusion. For crowded places (such as theaters, high-rise buildings, etc.), it can collect and analyze multi-dimensional data such as population density, temperature, and smoke concentration in real time, so as to quickly generate dynamic evacuation path guidance when a fire occurs. Through intelligent audio-visual guidance, including the adjustment of brightness, color, and sound information, the evacuation guidance is clearer and more intuitive, which improves the efficiency and safety of personnel evacuation.

[0051] In addition, the present invention effectively extends the use time of the backup power supply while ensuring the guidance effect through optimized lighting strategies. The system can dynamically adjust the lighting brightness of each area according to information such as smoke concentration and ambient light intensity to reduce energy consumption and avoid unnecessary light pollution. While ensuring safe evacuation, the present invention has a high energy utilization efficiency and can continue to stably guide personnel to evacuate safely even in the event of a power outage. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 A flow chart of a low-power fire real-time auxiliary first aid method based on edge computing provided in an embodiment of the present disclosure;

[0053] Figure 2 A schematic diagram of an edge computing unit cluster provided in an embodiment of the present disclosure;

[0054] Figure 3 A flow chart of a method for triggering a fire event alarm and dynamically generating an evacuation path provided in an embodiment of the present disclosure;

[0055] Figure 4 A schematic diagram of the structure of a low-power real-time fire emergency assistance system based on edge computing provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0056] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0057] See also Figure 1 , Figure 1 The flowchart of the low-power fire real-time auxiliary first aid method based on edge computing provided in the embodiment of the present disclosure includes steps S101 to S103:

[0058] S101: Receive and collect environmental data using a sensor interface, the environmental data including: temperature data, smoke concentration data, personnel data, and ambient light intensity data; wherein the personnel data includes: personnel density data and personnel movement rate;

[0059] S102: Based on the environmental data, using a preset fire detection algorithm, determining whether a fire event occurs in the edge computing unit; in response to detecting the occurrence of the fire event, generating a dynamic evacuation path in real time based on the fire source location and building layout;

[0060] S103: Based on the evacuation path, a control signal is sent to the control module using the input / output interface, and the control module controls the intelligent terminal to issue auxiliary first aid information based on the control signal; wherein the auxiliary first aid information is used to guide personnel to evacuate; the auxiliary first aid information includes: visual information and sound information.

[0061] In a specific implementation, the sensor interface is connected to multiple environmental sensors for real-time collection of environmental data. The environmental data includes temperature data, smoke concentration data, personnel density data, personnel movement rate, and ambient light intensity data. The edge computing unit receives these data from the sensor interface and analyzes them through the built-in fire detection algorithm to determine whether a fire event has occurred.

[0062] The fire detection algorithm can pre-set multiple judgment conditions, including: a sharp rise in temperature, a significant increase in smoke concentration, etc. When the edge computing unit detects environmental data that meets the fire conditions, it immediately determines that a fire event has occurred and begins to plan the evacuation route for personnel.

[0063] After detecting a fire event, the edge computing unit first dynamically generates an evacuation path based on the location of the fire source, building layout, and real-time updated personnel data. The evacuation path is adjusted in real time according to the location of the fire source and the direction of fire spread to ensure the safety and effectiveness of the evacuation path.

[0064] The edge computing unit uses the input / output interface to convert the generated evacuation path information into a control signal and transmits it to the control module. After receiving the control signal, the control module controls the intelligent terminal to issue auxiliary first aid information based on the control signal. The auxiliary first aid information is used to guide personnel to evacuate; the auxiliary first aid information includes: visual information and sound information.

[0065] For example, by increasing lighting brightness and changing color, the system guides people toward safe evacuation routes and away from danger zones.

[0066] For example, when the temperature detected by the temperature sensor exceeds the set threshold and the smoke concentration sensor detects that the smoke concentration continues to rise, the edge computing unit determines that a fire event has occurred. The edge computing unit immediately plans a safe evacuation path through a preset evacuation strategy and sends a control signal to the control module through the input / output interface.

[0067] After receiving the signal, the control module adjusts the lighting brightness on the evacuation path to the maximum and sets the lighting color to green to guide people to evacuate quickly along this path. At the same time, the lighting color of the area close to the fire source is set to red to prompt people to avoid these areas.

[0068] In this way, the present invention can detect the fire status in real time when a fire occurs, dynamically plan the optimal evacuation route, and effectively guide personnel to evacuate by adjusting the lighting brightness and color, thereby significantly improving the efficiency and safety of fire emergency evacuation.

[0069] As an optional implementation, the edge computing unit collects a variety of environmental data through the sensor interface, and these data can be processed through multi-sensor data fusion technology to generate more accurate fused sensor data.

[0070] As an optional implementation, multiple edge computing units can work together to form a cluster system to achieve more efficient and reliable fire detection and emergency response.

[0071] In specific implementation, multiple edge computing units can be organized into a cluster, and each edge computing unit in the cluster is responsible for processing sensor data within a certain range. The cluster system connects each edge computing unit through a communication network to achieve real-time sharing and collaborative processing of data. Each edge computing unit can undertake different data processing tasks according to the type and number of sensors it is connected to, thereby improving the overall processing efficiency.

[0072] In a cluster system, each edge computing unit not only processes its own sensor data independently, but also regularly exchanges data with other units to achieve distributed data fusion and analysis. For example, an edge computing unit can be responsible for fusing temperature and smoke concentration data and send the fusion results to the edge computing unit responsible for personnel density data. The edge computing unit responsible for personnel density data further fuses and analyzes these fusion results with the personnel density data it processes to generate a comprehensive description of the environmental status.

[0073] In the specific implementation, the CAN communication module of the edge computing unit first sets the communication rate of the CAN bus by configuring the baud rate register. For example, a baud rate of 500kbps is selected to meet the needs of real-time data transmission. Then, during the initialization process of each edge computing unit, the data frame format of the CAN bus is set to a standard frame, and a filter is used to receive only data frames related to the current task.

[0074] When sending data, the edge computing unit can use the internal hardware timer to generate a 32-bit timestamp and attach it to a specific field of the CAN data frame. After the receiving edge computing unit reads the timestamp, it determines whether the data is the latest by comparing it with the current system time. If the timestamp of the data exceeds the set time window (for example, 10 milliseconds), the data will be discarded and requested to be resent. To ensure the reliability of communication, the system automatically triggers the retransmission mechanism when data packet loss or CRC check error is detected. The edge computing unit will resend the data frame in a short time (for example, within 2 milliseconds) until the correct ACK signal is received.

[0075] In specific implementation, when the system starts, the edge computing unit can set different priorities for each task based on the sensor type and task complexity. For example, the processing task of temperature and smoke sensor data is set to the highest priority to ensure the real-time performance of fire detection. Priority can be achieved by setting the priority level of the task in the RTOS, and the task with a high priority will be executed first during task scheduling.

[0076] See also Figure 2 , Figure 2 Schematic diagram of the edge computing unit cluster provided for the embodiment of the present disclosure; in order to dynamically adjust the task allocation, the CPU utilization, task queue length, and task execution time of each edge computing unit can be monitored. When the load of a certain edge computing unit exceeds 80%, the task migration mechanism is automatically triggered. For example, when the CPU utilization of edge computing unit A is too high and the length of its task queue exceeds 10 tasks, the temperature data processing task can be selectively migrated to edge computing unit C. Task migration is implemented through the message queue in the RTOS. Edge computing unit A packages the context data of the current task (including input data, processing intermediate results, etc.) and sends it to the message queue of edge computing unit C. After receiving the task data, edge computing unit C restores the task context and continues to process unfinished tasks to ensure a smooth transition of task migration.

[0077] In the specific implementation, during the multi-sensor data fusion process, the data of each type of sensor can be normalized first to eliminate the dimensional differences between different data types. Then, the system assigns different weights to each type of sensor data, and the weight setting can be based on the accuracy and historical performance of the sensor. For example, the weight of the temperature sensor is set to 0.4, the weight of the smoke concentration sensor is 0.3, the weight of the occupancy density sensor is 0.2, and the weight of the ambient light intensity sensor is 0.1.

[0078] Exemplarily, a Kalman filter algorithm may be used to fuse these data.

[0079] In the specific implementation, the state estimate at the next moment is first predicted, and then the estimate is updated by receiving the actual sensor data. In the Kalman filter algorithm, the covariance matrix is ​​initialized based on the measurement noise characteristics of the sensor and updated according to the changes in the sensor data in each filtering iteration. The accuracy of the filtered data is evaluated by the mean square error (MSE). If the MSE is within a certain threshold range, the data fusion is considered successful.

[0080] For example, a large shopping mall has multiple sensors installed in it. These sensors are organized into a cluster system through multiple edge computing units. The system is responsible for monitoring fire conditions and providing evacuation guidance. Various types of sensors are installed in different areas of the shopping mall, especially those with high crowd density, such as food areas, clothing areas, underground parking lots, main passages, cinemas, etc., including temperature sensors, smoke concentration sensors, crowd density sensors, ambient light intensity sensors, etc. These sensors are connected to multiple edge computing units through a high-speed communication network.

[0081] For example, edge computing unit A is responsible for processing the temperature and smoke density sensor data in the food area. Edge computing unit B is responsible for processing the temperature and smoke density sensor data in the clothing area. Edge computing unit C is responsible for processing the population density data and ambient light intensity data in the underground parking lot. Edge computing unit D is responsible for processing the population density data and ambient light intensity data in the main channel. Each edge computing unit independently processes the data from the sensors it is connected to and regularly exchanges data with other edge computing units in the cluster to achieve distributed data fusion and analysis.

[0082] For example, when an abnormally high temperature and increased smoke concentration are detected in the food area, edge computing unit A will immediately process the data and determine whether there is a fire risk in the area. After processing, edge computing unit A sends the fused temperature and smoke data to edge computing unit C so that edge computing unit C can further fuse and analyze the data with the population density data and ambient light intensity data of the underground parking lot it is responsible for. Edge computing unit B simultaneously processes the temperature and smoke concentration data of the clothing area and sends the results to edge computing unit D so that edge computing unit D can fuse and analyze the data with the population density data and ambient light intensity data of the main channel.

[0083] After receiving the fused data from other edge computing units, each edge computing unit will further fuse and analyze it with the data it is responsible for to generate a more comprehensive description of the environmental status. For example, after edge computing unit C receives the fused data of the food area temperature and smoke concentration sent by edge computing unit A, it further fuses it with the population density and ambient light intensity data of the underground parking lot to generate a comprehensive description of the environmental status. If edge computing unit C detects that the population density in the underground parking lot is high and the temperature and smoke concentration are at dangerous levels, it will determine that the fire risk is extremely high and send a fire alarm signal to the entire cluster. After edge computing unit D receives the fused data of the clothing area temperature and smoke concentration sent by edge computing unit B, it fuses and analyzes it with the population density and ambient light intensity data of the main channel. If the population density in the main channel is high but the fire risk in the clothing area is low, the edge computing unit can decide not to evacuate the people in that area temporarily, but to concentrate on evacuating people in other more dangerous areas.

[0084] In this way, cluster technology has achieved efficient fusion and analysis of multi-sensor data, significantly improving the system's processing power and fire detection accuracy. The adaptive weight adjustment within the cluster further enhances the reliability and stability of the system, providing technical support for fire emergency lighting and evacuation guidance.

[0085] Based on the generated fused sensor data, the edge computing unit will compare it with the preset historical data model. The historical data model is constructed from past fire event data and contains the change patterns of various sensor data when fires occur under different circumstances. By comparing with the historical data model, the edge computing unit can dynamically adjust the threshold for fire judgment. For example, when certain indicators of the fused sensor data (such as temperature and smoke concentration) gradually approach the typical characteristics of historical fire events, the edge computing unit will lower the threshold for fire judgment to detect possible fire events earlier. This dynamic adjustment mechanism ensures that the system can adapt to different environments and fire types, and improves the sensitivity and accuracy of fire detection.

[0086] In the specific implementation, the edge computing unit uses fuzzy logic algorithm to process the fused sensor data. The fuzzy logic algorithm defines multiple fuzzy variables, such as temperature, smoke concentration, human density and ambient light intensity, and sets fuzzy membership functions for each fuzzy variable to reflect the different states of these data (such as "high temperature", "moderate smoke", "high density of people", etc.).

[0087] The process of generating an evacuation path combines the layout information of the building and the real-time updated population density data to ensure that people can evacuate through the safest path in the shortest time. The edge computing unit converts the generated evacuation path information into a control signal, transmits it to the control module through the input / output interface, controls the brightness and color of the smart terminal, and guides people to evacuate. In addition, sound information can also be used to strengthen reminders.

[0088] In the specific implementation, dynamic path planning can be achieved through reinforcement learning algorithm. Reinforcement learning is a machine learning method based on the interaction between agents and the environment. Agents continuously optimize their decision-making strategies through trial and feedback to achieve the goal of maximizing cumulative rewards. In fire evacuation path planning, each edge computing unit is modeled as an agent, and factors such as building layout, personnel distribution, fire source location and fire spread direction in the evacuation area constitute the environment in which the agent is located. The reinforcement learning model mainly includes state space, action space, reward function and strategy update mechanism. Among them, the state space consists of information describing the current environment, including the location of the fire source, the direction of fire spread, the density of people in each area, the location of the evacuation exit and the current fire risk score. The state can be represented as a multidimensional vector, and each dimension corresponds to a key parameter in the environment. The action space corresponds to the path planning decisions that the agent (edge ​​computing unit) can take. At each time step, the agent can choose an action, that is, specify an evacuation path for the personnel in the current state. The choice of action is based on the current state, and the agent can continuously optimize the action selection strategy through learning. The reward function defines the feedback obtained by the agent after taking an action at each time step. The reward design focuses on encouraging the agent to choose a path that can effectively reduce personnel risks and evacuation time.

[0089] For example, the reward function can be defined as follows: when people successfully evacuate through a safe path, a high reward is given. When the path chosen by the agent is close to the fire source or the direction of fire spread, a penalty is given. The reward is adjusted according to the length of the evacuation time, and the shorter the evacuation time, the higher the reward.

[0090] After training is completed, the trained reinforcement learning model is deployed to the edge computing unit cluster system. In an actual fire scenario, the system obtains the status information of the current environment in real time, inputs it into the trained model, and the model outputs the optimal evacuation path decision. Since the model has been trained extensively, it can quickly respond to dynamically changing environments and generate and adjust evacuation paths in real time.

[0091] In this way, the present invention can detect fire events more sensitively and accurately in complex and changing environments. The introduction of fuzzy logic algorithms enables the system to effectively process fuzzy and uncertain data, thereby generating accurate fire risk scores.

[0092] In the specific implementation, in order to generate the fire risk score, multiple fuzzy variables need to be introduced, including temperature variable, smoke concentration variable, personnel variable and ambient light intensity variable. Each fuzzy variable represents a different dimension of environmental data to better reflect different aspects of fire risk.

[0093] For example, the temperature variable can be divided into three fuzzy states: "low temperature", "medium temperature" and "high temperature", and each state has its corresponding fuzzy membership function. The membership function is usually represented by a triangle, trapezoid or other suitable distribution function. At a certain temperature, it may be partially in the "medium temperature" and "high temperature" states at the same time, and the specific membership degree is determined by the membership function. Similarly, the smoke concentration variable is also divided into three fuzzy states: "low concentration", "medium concentration" and "high concentration", and the corresponding membership function is set for each.

[0094] The fuzzy states and membership functions of the personnel variables and ambient light intensity variables are defined according to the specific application scenarios. For example, personnel variables may include "low density", "medium density" and "high density", while the ambient light intensity variable may be defined as "weak light", "medium light" and "strong light". Once the fuzzy variables and their membership functions are defined, the edge computing unit will process the fused sensor data through fuzzy inference rules.

[0095] Fuzzy inference rules are a set of "if-then" statements that combine the membership of fuzzy variables to generate a fire risk score. For example, a rule might be: If the temperature is "hot" and the smoke concentration is "high", then the fire risk score is "high". The system maps the current environmental data into these rules to calculate the fuzzy set output of the fire risk.

[0096] The edge computing unit will defuzzify the fire risk fuzzy set generated by the fuzzy inference rules to generate a specific fire risk score value. The defuzzification methods include the centroid method, the maximum membership method, etc. In specific implementation, the centroid method is usually used to calculate the centroid position of the fuzzy set membership function curve to obtain the final fire risk score. This score will be used for subsequent fire event judgment and response strategies, such as evacuation route planning.

[0097] For example, when a fire occurs in a building, the system will collect environmental data such as temperature and smoke concentration. These data will be mapped to their respective fuzzy variables, and their membership will be calculated according to the predefined membership function. For example, the current temperature may belong to the "high temperature" state, and the smoke concentration may partially belong to the "medium concentration" and "high concentration" states. Based on these memberships, the system evaluates the fire risk through fuzzy inference rules, and finally generates an accurate fire risk score through the defuzzification process.

[0098] The definition and division of each fuzzy variable are based on different environmental parameters and the actual needs of fire risk assessment. The fuzzy states of variables such as temperature, smoke concentration, occupant density and ambient light intensity are determined through a comprehensive analysis of historical data, actual fire cases and expert experience.

[0099] Temperature variables: usually divided based on the change in ambient temperature before and after a fire occurs. Low temperatures usually indicate a normal or non-fire state; medium temperatures indicate possible equipment heat dissipation or minor abnormalities; high temperatures indicate the presence of a fire or other dangerous heat source. These divisions can be adjusted based on the typical temperature range in a specific fire scenario.

[0100] Smoke concentration variable: The classification is based on the concentration level of smoke particles or harmful gases such as carbon monoxide in the air. Low concentrations are usually normal air quality, medium concentrations may indicate deteriorating air quality or initial smoke accumulation, and high concentrations significantly indicate increased concentrations of fire smoke or harmful gases. The classification criteria can be adjusted based on national or industry health and safety standards, as well as typical data in fire environments.

[0101] Density of people variable: The division is based on the density of people in the space. Low density usually corresponds to a sparse or unoccupied state, medium density is a normal activity state, and high density may occur in emergency evacuation or crowd gathering scenarios. This division can refer to the design capacity and actual usage of the building.

[0102] Ambient light intensity variables: The divisions are based on light intensity variations. Low light usually indicates a dim or no light environment, medium light is normal lighting conditions, and high light may be the result of light intensity variations caused by external light sources or fire. These divisions can be set based on the lighting requirements of different scenarios or the typical light variations during a fire.

[0103] Compared with the fire detection and evacuation guidance methods in the prior art that do not use fuzzy logic, the present invention significantly improves the system's adaptability and real-time performance by introducing fuzzy logic algorithms. The fuzzy logic system is highly interpretable, and the decision-making process is based on the "if-then" rule, which makes the system highly transparent and debuggable in practical applications. At the same time, since the fuzzy logic rule formulation is based on domain knowledge and engineering experience, it does not need to rely on a large amount of data for training, avoiding the common data dependency problems in machine learning and neural network methods, especially when sample data is scarce.

[0104] Compared with machine learning or neural network methods, the calculation process of fuzzy logic is relatively simple and suitable for scenarios with high real-time requirements, such as emergency evacuation route planning when a fire occurs. Although machine learning and neural networks perform well in complex pattern recognition, they usually require a large amount of data training and are easily limited by the quality and quantity of training data in actual deployment. Fuzzy logic, on the other hand, can maintain efficient and stable performance with limited data by defining well-defined membership functions and inference rules. In addition, the fuzzy logic system is extremely stable and can always output consistent results under the same input conditions, avoiding the randomness and uncertainty that may occur in machine learning models. This stability is particularly important for fire detection and emergency response systems that require high reliability.

[0105] See also Figure 3 , Figure 3 A flow chart of a method for triggering a fire event alarm and dynamically generating an evacuation path provided in an embodiment of the present disclosure. As an optional implementation, the method for triggering a fire event alarm and dynamically generating an evacuation path includes steps S201 to S203:

[0106] S201: In response to the fire risk score exceeding the fire judgment threshold, generating a prediction model of the fire source location and the fire spread direction in real time;

[0107] S202: Based on the fire source location and the fire spread direction, generate and dynamically adjust the evacuation path;

[0108] S203: triggering a multi-mode fire alarm, including: sound alarm, visual indication, and sending a remote alarm signal; wherein the visual indication includes: light flashing and light color change.

[0109] In a specific implementation, when the edge computing unit receives the fused sensor data and generates a fire risk score based on a preset fuzzy logic algorithm, it can monitor in real time whether the score exceeds the preset fire judgment threshold. Once the fire risk score exceeds the threshold, the edge computing unit will immediately trigger a fire event alarm and enter emergency mode.

[0110] After the fire alarm is triggered, the edge computing unit first uses the existing sensor data to generate a prediction model for the fire source location and the direction of fire spread. The model is based on the current sensor data, including temperature, smoke concentration, personnel movement rate, and ambient light intensity, and dynamically predicts the changing trend of the fire source location and the direction of fire spread by comparing and analyzing with the historical data model.

[0111] For example, the edge computing unit can calculate the speed and direction of fire spread by comparing the temperature and smoke density changes at different times, and predict the possible spread path of the fire. This prediction model will be used to update the evacuation path in real time to ensure that people can avoid the area where the fire spreads and evacuate smoothly.

[0112] Based on the prediction model of the fire source location and the direction of fire spread, the edge computing unit will generate and dynamically adjust the evacuation path in real time. The path generation algorithm takes into account the location of the fire source, the speed of fire spread, the layout of the building, and the density data and movement rate of people in the current evacuation area.

[0113] For example, the edge computing unit will prioritize paths that are far away from the fire source and have a low density of people as the main evacuation route. If the direction of fire spread intersects with the existing evacuation path, the evacuation route will be automatically replanned to ensure that all personnel can avoid the dangerous area and evacuate quickly and safely.

[0114] In the specific implementation, when a fire event occurs, data from multiple sensors are collected in real time, including: temperature, smoke concentration, population density, ambient light intensity, etc. Data preprocessing is performed by each edge computing unit in the edge computing unit cluster system. The generation of the prediction model is based on multiple key input variables, which come from the data fusion results of each sensor. For example, temperature change rate: extract temperature data from temperature sensors at multiple locations, and calculate the temperature change rate at each location. Smoke diffusion speed: based on the smoke concentration sensor data, calculate the diffusion speed of smoke at different locations. Population density distribution: through the data fusion of the population density sensor, obtain the population density distribution in different areas. Ambient light intensity change: based on the data of the ambient light intensity sensor, evaluate the impact of fire on ambient lighting.

[0115] In the specific implementation, machine learning algorithms (such as decision trees, support vector machines, or neural networks) are used to generate prediction models for the location of fire sources and the direction of fire spread. First, the system uses historical fire data to train the model. During the training process, historical sensor data and known fire source locations and fire spread paths are input as label data, and the model learns the dynamic characteristics of fire by continuously adjusting its own parameters.

[0116] The training set can use known fire case data, input historical sensor data and its corresponding fire source location and fire spread direction to form a training set.

[0117] It should be noted that the fuzzy logic mentioned above is mainly used to process real-time fire risk assessment after multi-sensor data fusion. Fuzzy logic is particularly good at processing uncertainty and fuzzy data (such as fuzzy states of high, medium, and low temperature), and it has the advantages of simple calculation and strong interpretability in real-time systems that require fast response. Therefore, fuzzy logic is suitable for quickly generating fire risk scores so that judgments can be made quickly and emergency responses can be triggered in the early stages of a fire incident.

[0118] Machine learning is used here to predict the location of the fire source and the direction of fire spread. This part of the work involves learning and analyzing a large amount of historical data to generate complex prediction models. Machine learning methods, such as neural networks and support vector machines, can obtain accurate prediction models through large amounts of data training, which is very effective when it is necessary to predict the path of fire spread and adjust the evacuation path.

[0119] In the event of a fire, the system inputs real-time sensor data into the trained model, which predicts the location of the fire source and the direction of fire spread based on the input real-time sensor data. The model can generate the coordinates of the fire source location and the fire spread direction vector in real time based on key factors such as temperature change rate and smoke diffusion speed.

[0120] The model predicts the current location coordinates of the fire source by analyzing variables such as the temperature change rate and smoke diffusion speed. The model predicts the direction of fire spread based on the historical changes in the location of the fire source and the current fire spread speed.

[0121] The evacuation route is dynamically adjusted based on the predicted location of the fire source and the direction of fire spread. Specifically, the route that is far away from the fire source and has a slower fire spread is selected as the preferred evacuation route. When the model predicts that the fire may change direction, the evacuation route is automatically recalculated and personnel are guided to avoid dangerous areas through smart terminals.

[0122] When the fire event alarm is triggered, the system activates the multi-mode fire alarm device. Specific implementation methods include: Sound alarm: A high-decibel alarm sound is emitted through the alarm installed on site to remind all personnel in the building to pay attention to the fire and evacuate immediately. The sound alarm mode can include continuous alarm sound, intermittent alarm sound, etc. to enhance the alarm effect. Visual indication: The edge computing unit controls the smart terminals at the evacuation channel and exit to flash the lights, and changes the color of the lights according to the real-time adjustment of the fire spread direction and evacuation path. For example, the area close to the fire source can use red flashing lights to indicate danger; the evacuation channel uses green or white lights to guide the safe evacuation path. Remote alarm signal: The edge computing unit remotely sends the fire alarm information to the monitoring center or other relevant agencies through a wireless network or wired connection. The alarm information includes key information such as the current fire source location, the direction of fire spread, and the real-time updated evacuation path, so that the relevant departments can respond in time.

[0123] In this way, through the above-mentioned multi-level alarm and dynamic path planning, the present invention can respond quickly when a fire occurs, predict the location of the fire source and the direction of fire spread in real time, and promptly notify the personnel in the building and the external emergency response team through a multi-mode alarm mechanism. In particular, in terms of dynamically generating and adjusting evacuation paths, the present invention provides highly flexible and safe evacuation guidance, which maximizes the safety of personnel's lives.

[0124] As an optional implementation, the edge computing unit generates the control signal based on the evacuation path, and the control signal includes: a brightness adjustment instruction and a color change instruction; the brightness adjustment instruction dynamically adjusts the brightness of the smart terminal based on personnel density data and personnel movement rate; the color change instruction dynamically adjusts the color of the smart terminal based on the fire source location and the direction of fire spread.

[0125] In a specific implementation, the generation of brightness adjustment instructions is based on the real-time collected personnel density data and personnel movement rate. The edge computing unit receives personnel density data and personnel movement rate data from the sensor interface, which reflect the distribution of personnel in various areas of the building and the movement speed of personnel.

[0126] For example, when the density of people in an evacuation channel is high and the speed of movement is slow, the edge computing unit will generate an adjustment instruction to increase the brightness and pass it to the smart terminal in the channel to increase the lighting brightness, thereby ensuring that people can clearly see the evacuation route and avoid congestion and trampling. On the contrary, if the density of people is low and the movement speed is fast, the edge computing unit may generate an instruction to moderately reduce the brightness to save energy and reduce unnecessary light pollution.

[0127] In addition to brightness adjustment, the generation of color change instructions is equally important. The generation of color change instructions is based on data on the location of the fire source and the direction of fire spread. These data are analyzed and updated in real time by the edge computing unit to ensure that the evacuation path is always away from the danger zone.

[0128] For example, when a fire source is detected close to an evacuation path, the edge computing unit will generate a color change instruction to adjust the color of the smart terminal in the area to a warning color (such as red or orange) to remind people to stay away from the area. For safe evacuation channels away from the fire source, the edge computing unit will generate an instruction to adjust the lighting color to a guiding color (such as green or blue) to indicate a safe evacuation path.

[0129] In a specific implementation, the edge computing unit receives data from a human density sensor, a human movement rate sensor, a temperature sensor, and a smoke concentration sensor through a sensor interface. Based on the preprocessed human density data and human movement rate data, the edge computing unit generates a brightness adjustment instruction using a preset logic rule or algorithm (such as a threshold judgment or a ratio adjustment algorithm). For example, if the human density of a certain channel exceeds a preset high-density threshold and the human movement rate is lower than a preset movement rate threshold, the edge computing unit will generate a brightness increase instruction. Otherwise, a brightness reduction instruction is generated.

[0130] The edge computing unit uses similar logic rules or algorithms to generate color change instructions based on the data of the fire source location and the direction of fire spread. For example, when the direction of fire spread intersects with an evacuation path, the edge computing unit will generate an instruction to adjust the lighting color of the path to a warning color and notify people in the relevant area to speed up evacuation.

[0131] The generated brightness adjustment instructions and color change instructions are transmitted to the control module through the input / output interface. The control module adjusts the brightness and color of each smart terminal according to the received instructions to ensure that the evacuation path is always clearly visible and changes dynamically according to the fire situation.

[0132] In this way, during a fire emergency, the present invention can dynamically adjust the brightness and color of the lighting system according to real-time updated data such as population density, movement speed, fire source location, and fire spread direction, thereby effectively guiding personnel to evacuate safely and orderly and avoiding congestion and accidents.

[0133] As an optional implementation manner, the dynamically adjusting the brightness of the smart terminal further includes: dynamically adjusting the brightness of the smart terminal based on the smoke concentration data and the ambient light intensity data.

[0134] In a specific implementation, the edge computing unit receives and processes data from the smoke concentration sensor and the ambient light intensity sensor in real time through the sensor interface.

[0135] It is understandable that when a fire occurs, smoke usually spreads quickly and may reduce visibility in certain areas. To ensure that personnel can still clearly see the evacuation route in areas with high smoke concentration, the edge computing unit adjusts the brightness of the smart terminal based on the smoke concentration data.

[0136] For example, when the smoke concentration exceeds a preset safety threshold (e.g. 200 ppm), the edge computing unit will generate an instruction to increase the brightness to enhance the lighting effect and ensure that the path is clearly visible. If the concentration continues to rise, the edge computing unit can further increase the brightness or even set it to the maximum value to cope with severe visibility reduction.

[0137] The intensity of ambient light directly affects the overall brightness of the evacuation passage. The edge computing unit dynamically adjusts the brightness of the smart terminal based on the real-time ambient light intensity data to ensure that the evacuation passage can maintain adequate lighting under different lighting conditions.

[0138] For example, during the day or when there is sufficient light, if the ambient light intensity sensor detects a high light intensity (such as more than 1000 lux), the edge computing unit may reduce the brightness of the smart terminal to avoid energy waste while maintaining good visibility. At night or when the ambient light intensity is low (such as less than 100 lux), the edge computing unit will appropriately increase the lighting brightness to ensure that the evacuation path is still clearly visible in low-light environments.

[0139] Exemplarily, the edge computing unit receives smoke concentration data and ambient light intensity data through a sensor interface. After preprocessing, the data enters a subsequent brightness adjustment decision module.

[0140] The edge computing unit uses preset logic rules or algorithms to analyze the received smoke concentration data. When the smoke concentration is high, the system will increase the lighting brightness to enhance the visibility of the evacuation path.

[0141] At the same time, the edge computing unit dynamically adjusts the brightness based on the ambient light intensity data. If the ambient light intensity is low, the system will increase the lighting brightness; if the ambient light intensity is high, the system will appropriately reduce the brightness to avoid light pollution and unnecessary energy consumption.

[0142] The edge computing unit transmits the generated brightness adjustment command to the control module through the input / output interface. After receiving the command, the control module adjusts the brightness of the smart terminal in real time to ensure that the evacuation path maintains optimal visibility under various environmental conditions.

[0143] In this way, the present invention can effectively utilize smoke concentration data and ambient light intensity data, dynamically adjust the brightness of the smart terminal, adapt to different fire and environmental conditions, improve the visibility of the evacuation passage, and thus ensure the safe evacuation of personnel in various situations.

[0144] It should be noted that the above description of low-power operation related to energy saving and avoiding energy waste also has corresponding beneficial effects. Because when a disaster occurs, the power supply of the entire scene may be damaged. At this time, the entire lighting system needs to operate with backup power. Therefore, the balance of energy is crucial to whether the evacuation system can operate normally for a long time to guide the evacuation of people.

[0145] As an optional implementation, the dynamic adjustment of the brightness of the smart terminal also includes: predicting a gathering place of people based on crowd density data and the movement rate of the people; and adjusting the brightness of the smart terminal in advance within a first time period before people gather at the predicted gathering place of people.

[0146] In the specific implementation, the edge computing unit receives and processes data from the crowd density sensor and the crowd movement rate sensor in real time through the sensor interface. These data are used to predict where people gather and adjust the lighting brightness of the relevant areas in advance.

[0147] The edge computing unit receives real-time data from the occupant density sensor and analyzes the current evacuation route or the distribution of occupants in each area of ​​the building. When the occupant density in a certain area gradually increases and may exceed the ideal threshold for safe evacuation, the system predicts that a gathering of occupants may occur in that area.

[0148] For example, if the density of people near an evacuation exit increases rapidly in a short period of time, the edge computing unit can determine that the exit may become a place where people gather, and the lighting brightness of the area needs to be adjusted in advance to ensure the safe evacuation of people.

[0149] In addition to the density of people, the edge computing unit also monitors the movement rate of people. By analyzing the changes in the movement rate of people, the system can predict the possible flow direction of people and future gathering places.

[0150] For example, when multiple people move toward an evacuation exit at a faster rate, the edge computing unit can determine that people may gather in the area, and therefore it is necessary to adjust the lighting brightness of the area within the first time period before the people arrive.

[0151] When the edge computing unit predicts that a certain area may be crowded with people, the system will adjust the brightness of the smart terminal in advance within the first period of time when the people arrive in the area. The purpose of this is to ensure that the area has enough light before the people arrive, to avoid confusion or accidents caused by insufficient light.

[0152] In the specific implementation, the edge computing unit receives real-time data on population density and population movement rate through the sensor interface. Based on this data, the edge computing unit analyzes the flow of people in each area and determines whether there is a possible risk of gathering of people.

[0153] Based on current and historical data, the edge computing unit uses algorithms to predict areas where people may gather. For example, when the density of people gradually increases and the movement rate slows down, the edge computing unit may predict that the area will become a gathering point.

[0154] Before predicting a gathering of people, the edge computing unit generates a brightness adjustment instruction according to a preset first time period (for example, 10 seconds before the arrival of people). The edge computing unit sends the instruction to the control module through the input / output interface, requiring the system to increase the lighting brightness of the predicted gathering location in advance.

[0155] When people actually arrive at the area, the lighting brightness has been adjusted to an appropriate level to ensure visibility and safety of the evacuation path. If people gather for a long time, the edge computing unit can further adjust the brightness according to the on-site situation to ensure continuous lighting effects.

[0156] In this way, the present invention can dynamically predict where people gather and adjust the brightness of the smart terminal in advance to ensure that the lighting conditions in each area during the evacuation process meet actual needs, thereby effectively improving evacuation efficiency and safety.

[0157] As an optional implementation, the edge computing unit dynamically adjusts the color of the smart terminal to indicate safe paths and dangerous areas. This function is implemented based on the relative position of the fire source, the direction of fire spread, and the evacuation path.

[0158] The system first determines the current location of the fire source and the direction of fire spread through sensor data and a prediction model of the fire source location. The edge computing unit compares this information with the pre-generated evacuation path to determine the relative relationship between the fire source location and the evacuation path. For example, the building's floor plan and real-time sensor data can be used to calculate the positional relationship of the fire source location relative to each node of the evacuation path through a geometric algorithm (such as the shortest path algorithm or ray projection method). If the fire source location is close to an evacuation path, the system will mark the path as a high-risk area.

[0159] According to the relative position of the fire source and the evacuation path, different colors are dynamically selected to indicate safe paths and dangerous areas. Specifically: when the evacuation path is far away from the fire source and there is no risk of fire spreading, the system will adjust the color of the smart terminal to a preset safety color (such as green) to indicate that the path is a safe passage. When the evacuation path is close to the fire source or the direction of fire spread, the system will adjust the color of the smart terminal to a preset warning color (such as red or orange) to prompt people to avoid the path.

[0160] The system dynamically adjusts the lighting colors of different areas by analyzing the relative position changes between the fire source and the path in real time to ensure the visual guidance effect of the evacuation path. When it is detected that the direction of fire spread intersects the evacuation path, the edge computing unit adjusts the color change frequency and gradient effect of the smart terminal in real time according to the speed of fire spread and the length of the path. For example: if the fire spreads quickly and the path length is short, the system will increase the flashing frequency of the warning color (such as red) to attract the attention of personnel. On the contrary, if the fire spreads slowly and the path length is long, the system can reduce the frequency of color change to make the warning effect more stable. The application of gradient effect is used to remind personnel of the extent of fire spread. For example, the system can remind personnel that the fire source is approaching a specific area by gradually deepening the color. The trigger of color change is determined by the speed of fire spread and the length of the path. The system monitors the speed of fire spread in real time. When it exceeds a certain set threshold (for example, the speed of fire spread exceeds 2 meters / second), the system will immediately trigger the color change and adjust the color of nearby smart terminals.

[0161] In addition, different color change thresholds can be preset according to the length of the evacuation path in different areas. For example, in a short path area, the color change can be triggered when the fire spreads slowly, while in a long path area, the color change is triggered only when the fire spreads quickly.

[0162] In this way, the present invention can dynamically adjust the lighting color of the evacuation path based on the location of the fire source and the direction of fire spread, prompting personnel to avoid dangerous areas in real time, and ensure that personnel can obtain clear and intuitive evacuation guidance during the evacuation process through reasonable color change frequency and gradient effect. This solution not only improves the intelligence level of the fire emergency lighting system, but also significantly improves the safety of personnel evacuation.

[0163] As an optional implementation, a preset safety distance D can be set. When the distance between the intersection and the fire source is less than or equal to D, the system will determine that the area is in a dangerous state and is ready to trigger the warning color adjustment. Red is used to indicate a high-risk area. When the intersection distance is less than or equal to the first distance threshold and the fire spread speed is greater than or equal to the first speed threshold, the system adjusts the color of the smart terminal to red. This indicates that the fire risk in the area is extremely high and people should stay away as soon as possible. Orange is used to indicate a moderately dangerous area. When the intersection distance is greater than the first distance threshold and less than the second distance threshold, and the fire spread speed is less than the first speed threshold but greater than the second speed threshold, the system adjusts the color of the smart terminal to orange, indicating that there is a medium fire risk in the area. Yellow is used to indicate a low-risk area. When the intersection distance is greater than or equal to the second distance threshold or the fire spread speed is less than or equal to the second speed threshold, the system adjusts the color of the smart terminal to yellow, indicating that the fire risk in the area is low.

[0164] By setting different distance and speed thresholds, the system can provide graded warnings for fire risks. The division of high-risk areas, medium-risk areas and low-risk areas enables the system to provide different levels of warnings according to the development of the fire. This graded warning not only helps to optimize evacuation strategies, but also ensures appropriate responses in different fire scenarios.

[0165] Without a reasonable threshold setting, the system may frequently trigger alarms due to the slight proximity of the fire source to the evacuation path or the slow spread of the fire, causing unnecessary panic and false alarms. Setting a threshold ensures that the warning color adjustment is triggered only in real danger, thereby improving the reliability of the system.

[0166] The preset distance and speed thresholds can be adjusted according to factors such as the layout, materials, and density of people in different buildings to meet the fire emergency needs in different scenarios. For example, in places with dense crowds, the thresholds may need to be set more sensitive to ensure early warning.

[0167] The system can convey the severity of fire risk to personnel through color classification and frequency of change. The three warning colors of red, orange and yellow represent different risk levels. This intuitive visual guidance helps personnel quickly understand the current dangerous state and make the right response, reducing hesitation and delays during evacuation.

[0168] Through these technical means, the present invention can dynamically adjust the lighting colors of different areas in the fire emergency system based on the location of the fire source, the speed of fire spread and its relative position to the evacuation path. This multi-level color warning can not only significantly improve the intelligence level of the system, but also combine with sound information to provide evacuees with clearer and more intuitive escape path guidance, thereby effectively improving the safety of fire emergency evacuation.

[0169] Based on the same inventive concept, the embodiment of the present application also provides a low-power real-time assisted fire first aid system based on edge computing corresponding to the low-power real-time assisted fire first aid method based on edge computing. Since the principle of solving the problem by the system in the embodiment of the present application is similar to the above-mentioned low-power real-time assisted fire first aid method based on edge computing, the implementation of the system can refer to the implementation of the method, and the repeated parts will not be repeated.

[0170] Reference Figure 4 As shown, it is a structural diagram of a low-power real-time fire emergency assistance system based on edge computing provided by an embodiment of the present disclosure, and the system includes:

[0171] The collecting unit 10 is used to receive and collect environmental data using a sensor interface, wherein the environmental data includes temperature data, smoke concentration data, personnel data, and ambient light intensity data; wherein the personnel data includes personnel density data and personnel movement rate;

[0172] A judgment unit 20 is used to judge whether a fire event occurs in the edge computing unit based on the environmental data and using a preset fire detection algorithm; in response to detecting the occurrence of the fire event, a dynamic evacuation path is generated in real time based on the location of the fire source and the layout of the building;

[0173] The processing unit 30 is used to send a control signal to the control module using the input / output interface based on the evacuation path, and the control module controls the intelligent terminal to issue auxiliary first aid information based on the control signal; wherein the auxiliary first aid information is used to guide personnel to evacuate; the auxiliary first aid information includes: visual information and sound information.

[0174] The preferred embodiments of the present invention disclosed above are only used to help explain the present application. The preferred embodiments do not describe all the details in detail, nor do they limit the present application to specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present application, so that those skilled in the art can understand and use the present application well. The present application is limited only by the claims and their full scope and equivalents.

Claims

1. A low-power fire real-time auxiliary first aid method based on edge computing, characterized in that: include: Receive and collect environmental data using a sensor interface, the environmental data including: temperature data, smoke concentration data, personnel data, and ambient light intensity data; wherein the personnel data includes: personnel density data and personnel movement rate; Based on the environmental data, using a preset fire detection algorithm, determining in the edge computing unit whether a fire event has occurred; in response to detecting the occurrence of the fire event, generating a dynamic evacuation path in real time based on the location of the fire source and the layout of the building; Based on the evacuation path, a control signal is sent to the control module using the input / output interface, and the control module controls the intelligent terminal to send auxiliary first aid information based on the control signal; wherein the auxiliary first aid information is used to guide personnel to evacuate; the auxiliary first aid information includes: visual information and sound information; The edge computing unit forms a distributed cluster system by multiple edge computing units; the determining whether a fire event occurs includes: Normalizing the temperature data, smoke concentration data, personnel data, and ambient light intensity data to eliminate dimensional differences between sensor data of different types; Set weights for different sensor data when performing multi-sensor data fusion based on sensor accuracy, historical performance, and measurement noise characteristics; The Kalman filter algorithm is used to perform multi-sensor data fusion on the weighted sensor data to generate fused sensor data; The generating of fused sensor data further includes: each edge computing unit is responsible for processing the sensor data connected to it, and sharing the data with other edge computing units in real time through a communication network to perform distributed data fusion; Dynamically adjust the fire judgment threshold based on real-time fusion sensor data and preset historical data model; Processing the fused sensor data based on a fuzzy logic algorithm to generate a fire risk score; In response to the fire risk score being greater than the fire judgment threshold, a fire event alarm is triggered and an evacuation route is dynamically generated.

2. The method according to claim 1, characterized in that: Generating a fire risk score includes: Defining fuzzy variables, including: temperature variable, smoke concentration variable, personnel variable and ambient light intensity variable, and setting a fuzzy membership function for each of the fuzzy variables; Using fuzzy inference rules, mapping the fused sensor data to the fuzzy membership of each of the fuzzy variables; Based on the fuzzy inference rules and in combination with the fuzzy membership of each of the fuzzy variables, a fuzzy fire risk score is calculated; The fuzzy fire risk score is defuzzified to generate the fire risk score.

3. The method according to claim 2, characterized in that The triggering of a fire event alarm and the dynamic generation of an evacuation route include: In response to the fire risk score exceeding the fire judgment threshold, generating a prediction model of the fire source location and the fire spread direction in real time; Based on the fire source location and fire spread direction, generate and dynamically adjust the evacuation path; Triggering multi-mode fire alarms, including: audible alarm, visual indication, and sending remote alarm signals; The visual indication includes: light flashing and light color change; wherein the light flashing and the light color change are executed by a smart terminal.

4. The method according to claim 3, characterized in that Also includes: The edge computing unit generates the control signal based on the evacuation path, and the control signal includes: a brightness adjustment instruction and a color change instruction; The brightness adjustment instruction dynamically adjusts the brightness of the smart terminal based on the personnel density data and the personnel movement rate; The color change instruction dynamically adjusts the color of the smart terminal based on the location of the fire source and the direction of fire spread.

5. The method according to claim 4, characterized in that The dynamically adjusting the brightness of the smart terminal further includes: Dynamically adjust the brightness of the smart terminal based on the smoke concentration data and the ambient light intensity data; The dynamically adjusting the brightness of the smart terminal further includes: Predicting where people gather based on the people density data and the movement speed of the people; In a first time period before the predicted gathering of people occurs at the gathering place of people, the brightness of the smart terminal is adjusted in advance.

6. The method according to claim 5, characterized in that The dynamically adjusting the color of the smart terminal includes: Based on the relative position of the fire source location and the evacuation path, dynamically select different colors to indicate safe paths and dangerous areas; In response to the fire spreading direction intersecting the evacuation route, adjusting the color of the smart terminal close to the fire source to a preset warning color; Adjust the lighting color of the area far away from the fire source to the preset guide color; Based on the fire spread speed and path length, the color change frequency and gradient effect of different areas are adjusted in real time.

7. The method according to claim 6, characterized in that The step of adjusting the color of the smart terminal near the fire source to a preset warning color includes: In response to detecting that the distance between the intersection of the fire spreading direction and the evacuation path is less than a preset safety distance, triggering the warning color adjustment; The preset warning colors include: red, orange, and yellow; wherein red is used to indicate a high-risk area; the high-risk area includes: the intersection distance is less than or equal to the first distance threshold and the fire spread speed is greater than or equal to the first speed threshold; Orange is used to indicate a moderately dangerous area; the moderately dangerous area includes: the intersection distance is greater than the first distance threshold and less than the second distance threshold, and the fire spread speed is less than the first speed threshold and greater than the second speed threshold; Yellow is used to indicate low-risk areas; the low-risk areas include: the intersection distance is greater than or equal to the second distance threshold or the fire spread speed is less than or equal to the second speed threshold.

8. A low-power fire real-time auxiliary emergency system based on edge computing, characterized in that: include: A collection unit, used to receive and collect environmental data using a sensor interface, wherein the environmental data includes: temperature data, smoke concentration data, personnel data, and ambient light intensity data; wherein the personnel data includes: personnel density data and personnel movement rate; A judgment unit, configured to judge whether a fire event has occurred in the edge computing unit based on the environmental data and using a preset fire detection algorithm; in response to detecting the occurrence of the fire event, generating a dynamic evacuation path in real time based on the location of the fire source and the layout of the building; A processing unit, configured to send a control signal to a control module using an input / output interface based on the evacuation path, and the control module controls the intelligent terminal to send auxiliary first aid information based on the control signal; wherein the auxiliary first aid information is used to guide personnel to evacuate; the auxiliary first aid information includes: visual information and sound information; The edge computing unit forms a distributed cluster system by multiple edge computing units; the determining whether a fire event occurs includes: Normalizing the temperature data, smoke concentration data, personnel data, and ambient light intensity data to eliminate dimensional differences between sensor data of different types; Set weights for different sensor data when performing multi-sensor data fusion based on sensor accuracy, historical performance, and measurement noise characteristics; The Kalman filter algorithm is used to perform multi-sensor data fusion on the weighted sensor data to generate fused sensor data; The generating of fused sensor data further includes: each edge computing unit is responsible for processing the sensor data connected to it, and sharing the data with other edge computing units in real time through a communication network to perform distributed data fusion; Dynamically adjust the fire judgment threshold based on real-time fusion sensor data and preset historical data model; Processing the fused sensor data based on a fuzzy logic algorithm to generate a fire risk score; In response to the fire risk score being greater than the fire judgment threshold, a fire event alarm is triggered and an evacuation route is dynamically generated.

Images