Intelligent fire response system

Abstract

The present invention comprises: a BIM database storing building information model data including three-dimensional shape information of a building, spatial attribute information, and possible evacuation paths; a plurality of IoT sensors installed inside the building to detect at least one environmental data among temperature, smoke, and harmful gases in real time; a data transmission and reception unit receiving the environmental data and user information from the plurality of IoT sensors and user terminals; and a central processing unit operating in conjunction with the BIM database, IoT sensors, and the data transmission and reception unit. The central processing unit, when a fire is detected by the IoT sensors, performs a fire spread simulation to predict the spread of fire and harmful substances using the building information model data and the environmental data, and an evacuation movement simulation to predict the degree of congestion according to the evacuation paths of a plurality of users, thereby generating a dynamic risk map that quantifies the risk of each unit space within the building, and calculates an optimal evacuation path individualized for each user based on the dynamic risk map and the current location of each user.

Description

Technology Field

[0001] The present invention relates to an intelligent fire response system, and more specifically, to an intelligent fire response system that predicts fire spread using real-time fire data and dynamically generates and guides an optimal evacuation route for occupants according to danger situations that change moment by moment. Background Technology

[0003] In modern society, buildings are becoming increasingly larger, taller, and more complex, leading to a rising risk of casualties in the event of a fire. In particular, for super high-rise buildings and large-scale complex facilities, rapid and safe evacuation is becoming more difficult due to their complex internal structures and large number of occupants.

[0004] Traditional fire safety systems operate through fixed equipment such as fire alarms, sprinklers, and emergency exit signs. These systems operate by guiding evacuation along predetermined routes in the event of a fire, always indicating the same path regardless of the actual location of the fire or the direction of smoke spread. However, fires unfold in highly diverse ways depending on the point of ignition, the type of combustible material, and the building's ventilation conditions, and fixed, predetermined routes are not always safe.

[0005] Recently, with the advancement of building digitalization, Building Information Modeling (BIM) technology is being widely utilized in the fields of construction and facility management. BIM includes not only three-dimensional shape information of a building but also attribute information such as spatial usage, materials, and facility locations, allowing it to be used for various purposes throughout the building's entire lifecycle. Furthermore, advancements in Internet of Things (IoT) technology have made it possible to install various sensors inside buildings to collect and monitor environmental data, such as temperature, humidity, air quality, and fire, in real time.

[0006] In the field of fire safety, research utilizing computational simulation technology to predict and analyze fire situations is actively underway. Computational Fluid Dynamics (CFD)-based fire simulation tools, such as the Fire Dynamics Simulator (FDS), can precisely calculate the movement of heat and smoke caused by a fire, while evacuation simulation tools like EVAC can simulate evacuation processes by reflecting the movement characteristics of occupants. However, these simulation tools are primarily used to pre-verify building safety during the architectural design phase or to analyze the causes of accidents after a fire occurs.

[0007] In actual fire situations, the progression of the fire is difficult to predict and changes moment by moment, making real-time response impossible based solely on pre-calculated simulation results. A system is required that continuously predicts the spread of fire by reflecting real-time environmental data collected from IoT sensors upon its occurrence, and dynamically provides the safest evacuation route based on this analysis.

[0008] Conventional evacuation guidance systems primarily guided occupants to the shortest exit by considering only the location of the fire and the occupants. However, in actual fire situations, various risk factors such as smoke diffusion, temperature rise, and reduced visibility interact in a complex manner, and a physically short path is not necessarily the safest. Therefore, it is crucial to provide an optimal route that comprehensively considers the level of risk associated with the fire, rather than simply distance.

[0009] Furthermore, in large buildings, as a large number of people evacuate simultaneously, congestion may occur at specific exits or corridors. Such bottlenecks not only delay evacuation times but can also lead to additional safety accidents due to congestion. Therefore, it is necessary to enhance overall evacuation safety by predicting the congestion levels of each evacuation route and distributing personnel accordingly.

[0010] A fire is not a static event but a dynamic situation that continuously changes over time. Routes that were initially safe may become dangerous due to smoke or flames as time passes, while conversely, routes that were initially blocked may become safe again due to firefighting efforts. Therefore, a system is required that collects sensor data in real time even while the fire is in progress, continuously monitors the spread of the fire and the movement of evacuees, and recalculates and updates evacuation routes as necessary.

[0011] As a related prior art, Korean Registered Patent No. 10-2124067 discloses a system that predicts smoke spread and guides evacuation routes in the event of a fire using Internet of Things (IoT) sensors. This prior art is characterized by collecting smoke and flame data detected by IoT sensors, predicting smoke and flame spread paths by comparing them with spread result data based on fire scenarios in conjunction with BIM information, and determining evacuation routes in real time by considering routes that may be closed over time. However, this prior art has limitations in that it determines evacuation routes primarily by predicting only fire spread, and thus fails to simultaneously consider the movement of evacuating personnel and the resulting changes in congestion levels along each route. In particular, it does not present specific methodologies for providing individualized evacuation routes to multiple users within the same space based on their respective locations and attributes, or for avoiding bottlenecks through the dispersed deployment of evacuating personnel.

[0012] As previously discussed, existing systems have limitations, such as separating fire spread prediction and evacuation route guidance, relying on predefined scenarios without reflecting real-time data, or failing to consider the characteristics and location of individual users. Prior art literature

[0013] Korean Registered Patent No. 10-2124067 The problem to be solved

[0014] The present invention aims to provide an intelligent fire response system that enables safe and efficient evacuation in the event of a fire by integrating fire spread simulation and evacuation movement simulation to generate a dynamic risk map in real time and providing an optimal evacuation route individualized according to each user's location and attributes. means of solving the problem

[0016] To achieve the above objective, the present invention comprises: a BIM database storing building information model data including three-dimensional shape information of a building, spatial attribute information, and possible evacuation paths; a plurality of IoT sensors installed inside the building to detect at least one environmental data among temperature, smoke, and harmful gases in real time; a data transmission and reception unit receiving the environmental data and user information from the plurality of IoT sensors and user terminals; and a central processing unit operating in conjunction with the BIM database, IoT sensors, and data transmission and reception unit. The central processing unit, when a fire is detected by the IoT sensors, performs a fire spread simulation to predict the spread of fire and harmful substances using the building information model data and the environmental data, and an evacuation movement simulation to predict the degree of congestion according to the evacuation paths of a plurality of users, thereby generating a dynamic risk map that quantifies the risk of each unit space within the building, and calculates an optimal evacuation path individualized for each user based on the dynamic risk map and the current location of each user.

[0017] In one embodiment, the central processing unit can update the fire spread simulation and the evacuation movement simulation by reflecting user location information continuously received from the data transmission and reception unit and environmental data continuously received from the IoT sensor during the fire, update the dynamic risk map by integrating the updated simulation results, and recalculate and update the optimal evacuation route based on the updated dynamic risk map.

[0018] In one embodiment, the central processing unit converts the possible evacuation path extracted from the building information model data into a graph structure composed of nodes and edges, dynamically updates the weight (cost) of each edge according to the dynamic risk map, and calculates the optimal evacuation path for each user using a shortest path search algorithm based on the updated weight.

[0019] In one embodiment, the central processing unit may set the weight (Cost) by reflecting the risk level of the corresponding path and calculate the path with the minimum sum of weights as the optimal evacuation path using the shortest path search algorithm.

[0020] In one embodiment, the central processing unit may use the Dijkstra algorithm or the A* (A-star) algorithm as the shortest path search algorithm.

[0021] In one embodiment, the central processing unit calculates the personnel density of each path by time period from the evacuation movement simulation results, identifies a section where the personnel density exceeds a preset threshold density as a bottleneck section, and can reassign some of the users heading toward the bottleneck section to another path.

[0022] In one embodiment, the central processing unit may receive user attribute information including at least one of the user's age, disability status, and movement speed from a user terminal, or determine user attribute information through at least one of image analysis using an IoT sensor or a camera, and calculate an optimal evacuation path by differentiating path generation criteria according to the user attribute information.

[0023] In one embodiment, the central processing unit classifies multiple users located within the same space into multiple groups based on their distance from the nearest evacuation route, and assigns different evacuation routes to each group to avoid bottlenecks.

[0024] In one embodiment, the central processing unit may classify a plurality of users located within the same space into a proximity group, an intermediate group, and a distance group based on their distance from a first exit, and assign a path to the first exit to the proximity group, a path to the second exit to the distance group, and a path selected according to the dynamic risk map to the intermediate group.

[0025] In addition, the present invention is further characterized by a method for providing a dynamic evacuation route using an intelligent fire response system, comprising: receiving in real time at least one environmental data among temperature, smoke, and hazardous gas from a plurality of IoT sensors installed inside a building; loading building information model data when a fire is detected from the IoT sensors; performing a fire spread simulation that predicts the three-dimensional spatial spread of fire and hazardous substances over time using the building information model data and the environmental data, and an evacuation movement simulation that predicts the congestion level of each route by time according to the evacuation routes of a plurality of users; integrating the results of the fire spread simulation and the results of the evacuation movement simulation to generate a dynamic risk map in real time that quantifies the risk level of each unit space within the building; calculating an optimal evacuation route individualized for each user based on the dynamic risk map and each user's current location; and transmitting the calculated optimal evacuation route to the corresponding user terminal. Effects of the invention

[0027] According to the present invention, by simultaneously predicting fire spread and the movement of evacuees and providing an individual optimal evacuation route with minimized risk for each user in real time, there is an advantage of avoiding bottlenecks and maximizing evacuation safety, thereby minimizing casualties in the event of a fire. Brief explanation of the drawing

[0029] Figure 1 shows a configuration diagram of an intelligent fire response system according to an embodiment of the present invention. Figure 2 shows a schematic diagram of a data transmission and reception unit according to an embodiment of the present invention. FIG. 3 is a conceptual diagram showing dynamic evacuation route guidance provided to a user reflecting real-time fire conditions, where FIG. 3a shows the evacuation route in the initial state of a fire and FIG. 3b shows the updated evacuation route after fire spread information has been reflected. FIG. 4 is a conceptual diagram of a scenario showing a bottleneck that may occur during the evacuation process of multiple users located in different spaces and a route reassignment process to resolve it. FIG. 4a shows a situation where a bottleneck is formed as user groups located in different places move to the same evacuation route, and FIG. 4b shows a state in which alternative evacuation routes are reassigned for each user group to alleviate the bottleneck. FIG. 5 is a conceptual diagram of a scenario showing a bottleneck occurring during the evacuation process of multiple users located in the same space and a path reassignment process to resolve it. FIG. 5a shows a situation where a bottleneck is formed near an evacuation route as a group of users densely concentrated in the same space is concentrated in one direction, and FIG. 5b shows a state where the evacuation route is readjusted by dispersing the group of users toward different evacuation routes to alleviate the bottleneck. Figure 6 shows the results of simulating the evacuation process of multiple users located in the same space. Figure 6a shows the evacuation result where a bottleneck occurs as users concentrate in the same evacuation route based on the shortest distance, and Figure 6b shows the evacuation result where evacuation stability is improved by dispersing user groups to different evacuation route directions to alleviate the bottleneck. FIG. 7 shows a flowchart of a method for providing a dynamic evacuation route according to an embodiment of the present invention. Specific details for implementing the invention

[0030] The present invention will be described in detail below with reference to the contents described in the attached drawings. However, the present invention is not limited or restricted by exemplary embodiments. Identical reference numerals in each drawing indicate components that perform substantially the same function.

[0031] The purpose and effects of the present invention may be naturally understood or become clearer through the following description, and the purpose and effects of the present invention are not limited solely to the description below. Furthermore, in describing the present invention, if it is determined that a detailed description of known technology related to the present invention may unnecessarily obscure the essence of the present invention, such detailed description will be omitted.

[0032] The terms used in this invention are used merely to describe specific embodiments and are not intended to limit the invention. The singular expression includes the plural expression unless the context clearly indicates otherwise. In this application, terms such as "comprising" or "having" are intended to specify the presence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the description of the invention, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0033] Terms such as "first," "second," etc., may be used to describe various components, but said components should not be limited by said terms. These terms are used solely for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be named the second component, and similarly, the second component may be named the first component.

[0034] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which this invention pertains. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this invention.

[0035] In interpreting the components, they are interpreted to include a margin of error even without a separate explicit indication. In the case of descriptions regarding temporal relationships, for example, where the temporal sequence is described using 'after,' 'following,' 'next,' 'before,' etc., cases that are not continuous are included unless 'immediately' or 'directly' is used.

[0036] Hereinafter, the technical configuration of the present invention will be described in detail with reference to the attached drawings.

[0037] FIG. 1 shows a configuration diagram of an intelligent fire response system (10) according to an embodiment of the present invention. Referring to FIG. 1, the intelligent fire response system (10) may include a BIM database (100), an IoT sensor (300), a data transmission and reception unit (500), and a central processing unit (700). The intelligent fire response system (10) is an intelligent system that integrates and analyzes real-time environmental data inside a building and user location information when a fire occurs, simultaneously performs fire spread prediction and evacuation personnel movement prediction, and dynamically provides an optimal evacuation route individualized to each user based on this. Each component of the intelligent fire response system (10) is organically linked to respond in real-time to a complex and ever-changing disaster situation such as a fire, and can minimize casualties.

[0038] A BIM database (100) can store building information model data including three-dimensional shape information, spatial attribute information, and possible evacuation paths of a building. Building Information Modeling (BIM) is a database that includes attribute information of all elements constituting a building, going beyond simple three-dimensional shape information. The three-dimensional shape information can include the three-dimensional coordinates, dimensions, and shapes of all structural and architectural elements, such as the building's exterior walls, interior walls, columns, beams, slabs, stairs, ramps, elevators, doors, and windows. This three-dimensional shape information enables accurate reproduction of the spatial structure inside the building and can be used as base data to precisely calculate the movement paths of smoke and heat when performing fire spread simulations.

[0039] Spatial attribute information may include the use, area, volume, occupancy, ceiling height, ventilation system, type and combustion characteristics of interior finishing materials, type and quantity of combustible materials, and information on fire doors and fire compartments for each room. For example, data may be stored by distinguishing between spaces with a high risk of fire and a large amount of combustible materials, such as kitchens, and spaces with relatively fewer combustible materials, such as conference rooms; information regarding whether the interior materials of each space are non-combustible or combustible may also be included. This spatial attribute information can be used as important input data to predict the fire growth rate, smoke generation volume, and hazardous gas concentration in the event of a fire.

[0040] Evacuation routes include all paths that allow movement from inside the building to a safe external space or evacuation space, and each route may consist of corridors, stairs, ramps, emergency exits, general entrances, etc. For each route, detailed information such as the width, length, slope, stair step height, door opening / closing direction, and the location of guide lights and emergency lighting may be stored. In particular, information on barrier-free routes accessible to the disabled or the elderly is also displayed separately and can be used for route guidance for wheelchair users or users with limited mobility. This comprehensive information stored in the BIM database (100) plays an essential role in increasing the accuracy of fire spread simulations and generating realistic evacuation routes that reflect the actual building structure.

[0041] The IoT sensor (300) is installed inside the building and can detect at least one of the following environmental data in real time: temperature, smoke, and harmful gases. The IoT (Internet of Things) sensor (300) is strategically distributed across each floor, each zone, and each space of the building and can perform the role of detecting the occurrence of a fire early and continuously monitoring the spread of the fire. The temperature sensor detects a rise in temperature in a specific space, the smoke sensor measures smoke concentration, and the harmful gas sensor measures the concentration of harmful gases such as carbon monoxide and carbon dioxide.

[0042] The IoT sensor (300) can measure various environmental data in real time and transmit it to the data transmission / reception unit (500) via wireless communication (e.g., Zigbee, LoRa, WiFi, Bluetooth, etc.). Each sensor has a unique identifier (ID) and location information, allowing for accurate identification of which environmental changes have occurred at which location. The placement density of the sensors is adjusted according to the building's purpose and risk level; sensors are placed at a higher density in areas with a high risk of fire, such as kitchens, boiler rooms, and electrical rooms, while they can be placed at appropriate intervals in general office spaces or corridors. The sensor network has a self-diagnostic function, so that if a sensor malfunction or communication interruption occurs, it immediately sends a notification to the central processing unit (700) to maintain the reliability of the system.

[0043] The data transmission and reception unit (500) can receive environment data and user information from a plurality of IoT sensors (300) and user terminals (1).

[0044] FIG. 2 shows a schematic diagram of a data transmission and reception unit (500) according to an embodiment of the present invention. Referring to FIG. 2, the data transmission and reception unit (500) can function as a communication hub that performs a relay role between an IoT sensor (300) and a user terminal (1). The data transmission and reception unit (500) can collect and preprocess received environmental data and transmit it to a central processing unit (700). The preprocessing process may include noise removal of sensor data, detection and correction of outliers, and standardization of data formats. For example, if a temporarily abnormal value is measured at a specific sensor, the value may be compared with that of an adjacent sensor to determine whether it is an outlier, and if necessary, the data may be corrected or excluded.

[0045] The data transmission and reception unit (500) can receive user information from the user terminal (1). The user information may include location information and user attribute information. The location information is obtained using technologies such as GPS, WiFi-based indoor positioning (WiFi Positioning System), BLE (Bluetooth Low Energy) beacons, and UWB (Ultra-Wideband) of the user terminal (1), and the user's current location can be expressed as three-dimensional coordinates (x, y, z) or a specific area identifier of a building. The user attribute information includes the user's age, gender, disability status, usual movement speed, health status, etc., and may be entered in advance by the user through a mobile application or automatically collected from a wearable device.

[0046] The data transmission and reception unit (500) performs bidirectional communication functions to receive the optimal evacuation route calculated by the central processing unit (700) and transmit it to the corresponding user terminal (1). Route information can be provided to the user in various forms, such as text, voice, graphics (maps), and augmented reality (AR), and a suitable form can be selected depending on the user's preference or situation. For example, voice guidance may be provided to visually impaired users, and a map displaying the route on a mobile screen may be provided to general users. In addition, since communication traffic may surge in a fire situation, a priority-based data transmission method is adopted so that emergency information (e.g., route closure notifications, emergency evacuation instructions) is transmitted preferentially.

[0047] The central processing unit (700) can operate in conjunction with the BIM database (100), IoT sensor (300), and data transmission / reception unit (500). The central processing unit (700) is a computing device corresponding to the brain of the intelligent fire response system (10) and can perform all high-level processing processes in real time, such as fire detection, fire spread simulation, evacuation movement simulation, dynamic risk map generation, optimal evacuation path calculation, and path updating. The central processing unit (700) can be physically implemented as a server or cloud computing platform and can be equipped with high-performance hardware such as a multi-core processor, a GPU (Graphics Processing Unit), large-capacity memory, and a high-speed network interface.

[0048] Fire spread simulation and evacuation movement simulation are compute-intensive tasks that require processing hundreds of thousands to millions of computation cells depending on the size and complexity of the building, making parallel computing capabilities essential. GPUs are particularly effective for such parallel computing, and parallel computing platforms such as CUDA or OpenCL can significantly improve simulation speed. Additionally, the central processing unit (700) is equipped with a stream processing engine for real-time data processing, enabling it to immediately analyze and respond to a large amount of sensor data continuously flowing in from IoT sensors (300). The central processing unit (700) can load building information from the BIM database (100), receive real-time environmental data from IoT sensors (300), and receive location and attribute information from user terminals (1), thereby enabling integrated analysis of all this data.

[0049] When a fire is detected by an IoT sensor (300), the central processing unit (700) can generate a dynamic risk map that quantifies the risk of each unit space within the building by performing a fire spread simulation that predicts the spread of fire and hazardous substances using building information model data and environmental data, and an evacuation movement simulation that predicts the degree of congestion along the evacuation routes of multiple users. Fire detection can be performed by monitoring the measurements of the IoT sensor (300) in real time. The central processing unit (700) sets thresholds for each sensor type, and while it may determine a fire based solely on the threshold exceeded by a single sensor, it may also determine a fire if multiple adjacent sensors exceed the threshold simultaneously or continuously within a short period of time to prevent false alarms. Additionally, the central processing unit (700) can determine the growth rate of the fire by analyzing the rate of change of sensor values, and, for example, if the temperature rises at a rate of 10°C or more per minute, it may determine it as rapid fire growth and take more urgent action.

[0050] Fire spread simulations can be performed using physical models based on Computational Fluid Dynamics (CFD). A representative fire simulation tool is the Fire Dynamics Simulator (FDS), and the present invention can utilize CFD-based models such as FDS or Zone models. FDS calculates fluid flow, heat transfer, and combustion reactions based on the Navier-Stokes equations, and divides the internal space of a building into a three-dimensional grid to calculate the temperature, pressure, flow velocity, smoke concentration, hazardous gas concentration, oxygen concentration, etc., in each grid cell at each time step.

[0051] The initial conditions for the simulation may include the fire ignition point, ignition time, and heat release rate (HRR). The ignition point is determined based on the location of the sensor where the fire was first detected, and the heat release rate can be estimated based on the type and amount of combustible material in the space. The results of the fire spread simulation yield temperature distribution, smoke concentration distribution, carbon monoxide concentration distribution, visibility distribution, and oxygen concentration distribution in each space over time, which can be visualized as a 3D or 2D (floor-by-floor) distribution map.

[0052] Evacuation movement simulations can simulate the process of users within a building moving along evacuation routes following a fire. Evacuation simulation tools may include EVAC modules integrated with FDS, or specialized software such as Pathfinder, EXODUS, and Simulex. Evacuation movement simulations utilize an agent-based model, modeling each user as an independent agent to simulate individual behaviors. Each agent possesses attributes such as location, movement speed, body size, field of view, and reaction time, and can perceive the surrounding environment (smoke, temperature, the location of other users, etc.) to determine actions accordingly.

[0053] Movement speed is set differently depending on the user's age, gender, and physical condition; generally, adult males have a free movement speed of approximately 1.35 m / s, adult females 1.15 m / s, the elderly 0.8 m / s, and children 0.9 m / s. However, since movement speed may decrease depending on congestion in actual evacuation situations, the EVAC module can calculate the speed reduction based on congestion using a velocity-density relationship. Response time refers to the time from when a user recognizes a fire alarm until they actually begin evacuation, and it can generally be modeled as a random variable following a normal distribution.

[0054] In the evacuation movement simulation, each user moves along the guided path, and the time taken to pass through each point (corridor, stairs, exit, etc.) on the path and the density of people at that point can be calculated. As a result of the simulation, the density of people by time of each path, the evacuation completion time for each user, the total evacuation completion time (RSET, Required Safe Egress Time), and the location and duration of bottleneck sections can be calculated.

[0055] The central processing unit (700) can generate a dynamic risk map by integrating the fire spread simulation results and the evacuation movement simulation results. The dynamic risk map is dynamic data that is updated in real time over time, in which a comprehensive risk index is assigned to each unit space and each path within the building. The risk can be quantified by comprehensively considering factors such as temperature, smoke concentration, hazardous gas concentration, and personnel density. Each of these factors can be normalized to a value between 0 and 1, and the comprehensive risk index can be calculated by assigning weights and summing them. The weights can be adjusted according to the fire stage or building characteristics; for example, the weights for temperature and smoke can be set high during the initial stage of the fire, and the weights for hazardous gases can be set high during the fire growth stage.

[0056] The dynamic risk map can be visualized in the form of a heatmap that divides each floor of a building into a two-dimensional grid and displays a risk index for each grid cell in color, for example, green can indicate safety, yellow caution, orange warning, and red danger. The dynamic risk map is continuously updated as fire spread simulations and evacuation movement simulations are repeatedly performed, thereby allowing the ever-changing fire situation to be reflected in real time.

[0057] The central processing unit (700) can calculate an optimal evacuation route individualized for each user based on a dynamic risk map and each user's current location. The central processing unit (700) can convert the received user location information into the coordinate system of the BIM data and determine which floor and which space the user is currently located in. For each user, the central processing unit (700) can calculate an optimal route considering the risk level reflected in the dynamic risk map, with the current location as the starting point and a safe zone outside the building as the target point. At this time, even users within the same building may be provided with different optimal evacuation routes because their respective locations differ. For example, a route moving away from the fire may be provided preferentially to a user close to the point of fire, and a route to the nearest exit may be provided to a user far from the fire. In addition, user attributes (age, disability status, etc.) may also be considered, for example, a route that does not include stairs may be provided to a wheelchair user, and a route with a short travel distance may be provided preferentially to the elderly.

[0058] The central processing unit (700) can update the fire spread simulation and evacuation movement simulation by reflecting user location information continuously received from the data transmission and reception unit (500) and environmental data continuously received from the IoT sensor (300) during the fire, update the dynamic risk map by integrating the updated simulation results, and recalculate and update the optimal evacuation route based on the updated dynamic risk map.

[0059] Since fire is a dynamic phenomenon that continuously spreads and changes over time, the initially guided route may no longer be safe as time passes. For example, a hallway that was initially safe may become filled with smoke, or the temperature of a specific staircase may rise to a dangerous level. Additionally, as users move along the guided route, the congestion level of each route may change, and unexpected bottlenecks may occur. Therefore, the central processing unit (700) can receive environmental data from the IoT sensor (300) in real time even while the fire is in progress and continuously update the current fire situation.

[0060] The central processing unit (700) can update the fire spread simulation by reflecting new sensor data as boundary conditions of the simulation. At the same time, it can determine the actual movement paths and speeds of users by continuously receiving location information from the user terminal (1). If a user moves to a different path instead of following the guided path, the central processing unit (700) can detect this and recalculate and provide a new path for the user based on their current location. The evacuation movement simulation is also re-executed by reflecting the updated sensor data and user location information, and the current and future congestion levels of each path can be recalculated.

[0061] The updated results of the fire spread simulation and the evacuation movement simulation are integrated to update the dynamic risk map, and the optimal evacuation route for each user can be recalculated based on the updated dynamic risk map. The route recalculation cycle can be adjusted according to the fire growth rate and the system's computational capacity. If the recalculated route differs from the existing route, i.e., if a route change is required, the central processing unit (700) can immediately transmit the updated route information to the corresponding user terminal (1) and provide a notification to the user informing them of the route change. Through this real-time update process, users can always be guided to the latest safe route according to changes in the fire situation, which can significantly improve evacuation safety.

[0062] FIG. 3 is a conceptual diagram showing dynamic evacuation route guidance provided to a user reflecting real-time fire conditions, where FIG. 3a shows the evacuation route in the initial state of a fire and FIG. 3b shows the updated evacuation route after fire spread information has been reflected.

[0063] Referring to FIG. 3a, a floor plan of the building is shown, and the point marked with a red circle in the upper left is the point of fire. The building consists of several rooms and corridors, and three exits are arranged: the first escape route, the second escape route, and the third escape route. Users U1, U2, U3, and U4 are each located in different positions. User U1 is located in the center near the point of fire, User U2 is located in the upper center, User U3 is on the right, and User U4 is located in the lower right. At the initial stage of the fire, the central processing unit (700) provides each user with an optimal evacuation route reflecting the initial fire situation. Since User U1 is close to the point of fire, a route (indicated by a dotted line) leading to the first escape route, which is the direction away from the fire, is provided. Since User U2 is located at the top, a route leading to the nearest first escape route is provided. Since Users U3 and U4 are located on the right, a route leading to the second escape route is provided. At this point, since the fire is confined to the point of origin and smoke spread is limited, each user can reach the nearest exit via a relatively simple path.

[0064] Referring to FIG. 3b, the state of the fire spreading over time is shown. A wide gray area is displayed centered on the point of origin of the fire, indicating the area where the smoke has spread. The smoke has spread along the left corridor from the fire point and is affecting the upper central section. In this case, the paths of users U1 and U2, who were initially guided to the first escape route, have become dangerous due to the smoke. The central processing unit (700) updates the dynamic risk map by reflecting the updated fire spread simulation results and provides a detour route to users U1 and U2. User U1 abandons the original path toward the first escape route and is guided to the second escape route or an alternative path at the bottom by detouring to the right. User U2 is guided to the path (dotted arrow) toward the second escape route after detouring to the right, avoiding the upper central corridor where the smoke has spread. Users U3 and U4 maintain their original path to the second escape route because they are located on the right side, which is less affected by the spread of fire and smoke.

[0065] In this way, the evacuation route is updated in real time according to the fire spread situation, allowing users to avoid dangerous zones and evacuate safely. The example in Fig. 3 illustrates the concept of providing a "dynamic" evacuation route, which is a feature of the present invention, and unlike fixed evacuation signs or guide lights, it visually shows the route changing in response to the real-time fire situation.

[0066] The central processing unit (700) can convert evacuation paths extracted from building information model data into a graph structure composed of nodes and edges, dynamically update the weight (Cost) of each edge according to a dynamic risk map, and calculate the optimal evacuation path for each user using a shortest path search algorithm based on the updated weights. The central processing unit (700) can set the weight (Cost) by reflecting the risk level of the path and calculate the path where the sum of the weights is minimized as the optimal evacuation path using a shortest path search algorithm. The conversion into a graph structure is a process of mathematically modeling evacuation paths within a building, and means converting a complex building structure into a form that a computer can process efficiently.

[0067] Nodes represent key points along a path, and corridor intersections, stair entrances and exits, door locations, elevator halls, entrances to each room, and final exits (evacuations) can be represented as nodes. Each node has a unique identifier (ID) and 3D coordinates (x, y, z), which can be automatically extracted from BIM data. Edges represent paths connecting two nodes, and corridor sections, stair sections, and connections passing through doors can be represented as edges. Each edge can have attributes such as a start node, an end node, physical length, width, slope, and path type (corridor / stairs / ramp, etc.).

[0068] Each edge may be assigned a weight (Cost) representing the cost required to pass through the corresponding path. The weight does not merely represent physical distance, but comprehensively reflects the risk level of the path. When the risk index of each path segment is calculated from the dynamic risk map, it can be converted into the edge weight. Since the risk level of each path changes as the fire progresses, the central processing unit (700) can dynamically update the weight of each edge according to the update of the dynamic risk map. For example, initially, the risk of a specific corridor is low and the weight is 10, but if the risk level increases as smoke spreads, the weight can be increased to 50. Conversely, if the risk level decreases due to firefighting activities or the operation of the ventilation system, the weight can be lowered again. If a specific path is completely closed and passage is impossible, the weight of the corresponding edge is set to infinity (∞) or removed from the graph so that the path is not selected during path search.

[0069] The central processing unit (700) can apply a shortest path search algorithm to the graph structure configured in this way to search for a path from each user's current location (start node) to a safe zone outside the building (target node) that minimizes the sum of the weights. Here, the term "shortest" does not mean the path with the shortest physical distance, but rather the path with the minimum sum of the weights, that is, the optimal path when comprehensively considering risk and travel time.

[0070] The central processing unit (700) may use the Dijkstra algorithm or the A* (A-star) algorithm as the shortest path search algorithm. The Dijkstra algorithm calculates the shortest path from a starting node to all other nodes and works effectively in graphs where the weights are not negative. The algorithm starts from the starting node and visits adjacent nodes, updating the shortest distance to each node. It increases efficiency by using a priority queue to visit the node with the shortest shortest distance found so far first.

[0071] The A* algorithm is an improved version of Dijkstra's algorithm that enhances search efficiency by utilizing heuristic functions. Since the A* algorithm prioritizes searching in the direction of the target node using heuristics, it searches fewer nodes than Dijkstra's algorithm, allowing it to find the shortest path faster. Particularly in fire situations where real-time response is critical, the computationally fast A* algorithm can be effectively utilized.

[0072] The central processing unit (700) calculates the personnel density of each path by time period from the evacuation movement simulation results, identifies sections where the personnel density exceeds a preset threshold density as bottleneck sections, and can reassign some of the users heading toward the bottleneck sections to other paths.

[0073] As previously explained, evacuation movement simulations are performed using EVAC modules or agent-based models, and the number of people and the time taken to pass through each path segment are calculated as each user moves along the guided route. Personnel density is calculated by dividing the number of users located in a specific path segment (e.g., a certain section of a corridor or staircase) at a specific time by the area of ​​that segment, with the unit being persons / m². For example, if there are 40 users in a corridor segment (area 20m²) with a width of 2m and a length of 10m at a specific time, the personnel density is 40 persons / 20m² = 2 persons / m². As personnel density increases, movement speed decreases, and the risk of falls and crushing increases.

[0074] According to prior research, if the person density exceeds 2 people / m², movement begins to become difficult, if it exceeds 3 people / m², movement speed decreases significantly, and at 5 people / m² or more, there may be a risk of being crushed. Therefore, the central processing unit (700) sets a critical density, and generally, 3 people / m² can be used as the critical value. During the process of performing an evacuation movement simulation, the person density of all path sections is calculated at each time step (e.g., at 1-second intervals), and if a section exceeding the critical density is identified, it can be recorded as a bottleneck section. Since bottleneck sections are a major cause of reduced evacuation efficiency, resolving them is key to improving overall evacuation safety.

[0075] When a bottleneck is identified, the central processing unit (700) may reassign some of the users heading toward that section to another path. The criteria for selecting users to be reassigned may vary; for example, users who are far from the bottleneck may be reassigned first. This is because changing the path of users who have not yet reached the bottleneck is more efficient than changing the path of users who have already entered the bottleneck. Another criterion may be to prioritize reassigning users who have a lower cost of switching to an alternative path (additional travel distance or time). For example, users whose distance from their current location to an alternative exit does not increase significantly may be selected first. The proportion of users being reassigned may be adjusted according to the severity of the bottleneck.

[0076] The central processing unit (700) continuously performs this reassignment process while repeatedly performing the evacuation movement simulation, and can check whether the bottleneck has been resolved by running the simulation again after reassignment.

[0077] FIG. 4 is a conceptual diagram of a scenario showing a bottleneck that may occur during the evacuation process of multiple users located in different spaces and a route reassignment process to resolve it. FIG. 4a shows a situation where a bottleneck is formed as user groups located in different places move to the same evacuation route, and FIG. 4b shows a state in which alternative evacuation routes are reassigned for each user group to alleviate the bottleneck.

[0078] Referring to Fig. 4a, a floor plan of the building is shown, with a fire origin point marked by a red circle in the upper left corner. Three user groups are distributed across different locations within the building. Group G1 is located in the space near the fire in the upper left corner, Group G2 is located in the upper center space, and Group G3 is located in the large central space. The building has three exits: Exit 1, Exit 2, and Exit 3; Exit 1 is located in the upper center, Exit 2 is in the lower right corner, and Exit 3 is in the lower left corner. In the situation depicted in Fig. 4a, all three groups were guided toward the nearest Exit 1. In this case, the paths of the three groups converge in the corridor leading to Exit 1, forming a bottleneck section marked by hatching at this point. At the bottleneck section, the density of people increases rapidly, slowing down movement speed, and some users become completely stuck, delaying evacuation. This is a major problem that increases the overall evacuation time and reduces safety.

[0079] Referring to FIG. 4b, the central processing unit (700) identifies a bottleneck and reassigns the route to each user group. Group G1 is located closest to the fire and requires rapid evacuation, so it maintains the route to the first evacuation exit (dotted arrow). Group G2 changes its original route toward the first evacuation exit, detours to the right, moves downward, and is guided to the second evacuation exit or another alternative route (dotted arrow). Group G3 moves downward and is guided to the second evacuation exit or the third evacuation exit (dotted arrow). Through this reassignment, the number of people heading toward the first evacuation exit is reduced to only Group G1, and Groups G2 and G3 are dispersed to other exits, thereby resolving the bottleneck. Although the travel distance for Groups G2 and G3 may increase slightly, the actual evacuation time may actually be shortened as the congestion time caused by the bottleneck is eliminated. In addition, the dispersion of personnel reduces the risk of safety accidents caused by congestion and improves overall evacuation safety. The example in Fig. 4 illustrates a bottleneck that occurs when the paths of user groups in different spaces merge at a specific point and a solution to resolve it, and clarifies that the present invention does not merely consider the shortest path of individual users, but optimizes the evacuation flow of all users.

[0080] The central processing unit (700) classifies multiple users located within the same space into multiple groups based on their distance from the nearest evacuation route, and assigns different evacuation routes to each group to avoid bottlenecks. The same space refers to a single partitioned indoor space, such as the same office, the same conference room, or the same auditorium. When multiple users are present in such a space, if all users move simultaneously to the nearest exit, a bottleneck may occur near that exit.

[0081] In one embodiment, the central processing unit (700) classifies a plurality of users located within the same space into a proximity group, an intermediate group, and a distance group based on their distance from a first exit, and assigns a path to the first exit to the proximity group, a path to the second exit to the distance group, and a path selected according to a dynamic risk map to the intermediate group.

[0082] The proximity group is assigned a route to Exit 1, allowing for rapid evacuation as they are closest to the exit. The distance group is assigned a route to Exit 2 or Exit 3, which are further away from Exit 1; although the physical distance may be slightly greater, this avoids bottlenecks, enabling a safer and faster overall evacuation. The intermediate group is assigned to either Exit 1 or Exit 2 depending on the situation, which can be determined by comparing the expected congestion and risk levels of each route using a dynamic risk map. This group-based route assignment prevents the concentration of people at specific exits and effectively avoids bottlenecks by distributing all users across multiple exits.

[0083] The selection of a path for the intermediate group can be achieved through the following process. First, an evacuation movement simulation is performed assuming that the proximity group and the distance group move to the first and second exits, respectively. Through this, the expected personnel density and travel time for the path to the first exit and the path to the second exit are calculated. At the same time, the risk level (temperature, smoke concentration, hazardous gas concentration, etc.) of each path is evaluated from the results of the fire spread simulation. By synthesizing this information, the weight of each path is calculated. For example, if the path to the first exit is short in distance but has high congestion due to the movement of the proximity group, and the path to the second exit is somewhat longer but has low congestion and risk, the path to the second exit will have a lower weight, and the path to the second exit will be assigned to the intermediate group. Conversely, if the path to the first exit has acceptable levels of both congestion and risk, the path to the first exit may also be assigned to the intermediate group.

[0084] The dynamic selection process for intermediate groups is repeated in real-time during the fire, and the intermediate groups' paths may change depending on changes in the situation. For example, the intermediate group may initially be assigned to Exit 1, but if the fire spreads and the path to Exit 1 becomes dangerous, the intermediate group's path may be changed to Exit 2. Through this sequential assignment and dynamic reflection process, optimized paths are provided for each user even within the same space, thereby maximizing overall evacuation safety and efficiency.

[0085] FIG. 5 is a conceptual diagram of a scenario showing a bottleneck occurring during the evacuation process of multiple users located in the same space and a path reassignment process to resolve it. FIG. 5a shows a situation where a bottleneck is formed near an evacuation route as a group of users densely concentrated in the same space is concentrated in one direction, and FIG. 5b shows a state where the evacuation route is readjusted by dispersing the group of users toward different evacuation routes to alleviate the bottleneck.

[0086] Referring to Fig. 5a, there is a large space in the upper left corner of the building, where approximately 20 users are densely packed. A fire has broken out on the left side of the same space (red circle), and the users attempt to move to the right to escape the fire. The building has a first escape route (top center) and a second escape route (bottom right). In the situation depicted in Fig. 5a, all users were instructed to move to the nearest first escape route. When all 20 users crowd into the corridor leading to the first escape route, a bottleneck forms in the corridor near the first escape route. Since the width of the corridor is limited, attempting to pass through simultaneously causes the density of people to increase rapidly and significantly reduces movement speed. Some users are forced to wait in front of the exit, which increases evacuation time and extends the time exposed to fire hazards.

[0087] Referring to FIG. 5b, the central processing unit (700) classifies users within the same space into three groups based on their distance from the first exit and assigns a different escape route to each group. Group G1 consists of users located at the top of the space closest to the first exit, and a route to the first escape route is assigned to them. Group G1 is the proximity group, and since the distance to the first exit is the shortest, they are guided to move directly to the first escape route. Group G2 consists of users located at the bottom left, furthest from the first exit, and is classified as the distance group, so a route to the second escape route is assigned. Group G3 consists of users located at the bottom center, and since the distance from the first exit is between that of the proximity group and the distance group, they are classified as the middle group.

[0088] The path assignment for the intermediate group, Group G3, is determined through a dynamic selection process. First, an evacuation movement simulation is performed assuming that the proximity group (G1) moves to the first evacuation route and the distance group (G2) moves to the second evacuation route. Through this, the expected congestion and travel time for each of the first and second evacuation route paths are calculated. At the same time, the risk level (temperature, smoke concentration, hazardous gas concentration) of each route is evaluated from the fire spread simulation results. The central processing unit (700) synthesizes this congestion and risk information to calculate the weight of each route and assigns the route with the lower weight to Group G3.

[0089] Specifically, Group G1 moves to the right and heads toward the first evacuation route, while Groups G2 and G3 head toward either the first or second evacuation route depending on the situation. Through this distributed allocation, the number of personnel heading toward the first evacuation route is reduced, and Group G2 is dispersed toward the second evacuation route, thereby resolving bottlenecks. Although the travel distance for Group G2 increases slightly compared to going toward the first evacuation route, the actual evacuation completion time can be shortened as waiting time due to bottlenecks is eliminated. Furthermore, as personnel are dispersed toward two exits, the risk of falling or being crushed due to congestion is significantly reduced, and overall evacuation safety is improved.

[0090] The example in FIG. 5 demonstrates that bottlenecks can be effectively resolved by differentiating paths according to the user's location even within the same space and assigning an optimal path that reflects real-time conditions to intermediate groups through a dynamic selection process, and is an embodiment that clearly implements the concept of user location-based customization of the present invention.

[0091] Figure 6 shows the results of simulating the evacuation process of multiple users located in the same space. Figure 6a shows the evacuation result where a bottleneck occurs as users concentrate in the same evacuation route based on the shortest distance, and Figure 6b shows the evacuation result where evacuation stability is improved by dispersing user groups to different evacuation route directions to alleviate the bottleneck.

[0092] Figure 6a illustrates a scenario in which all users evacuate to the nearest top staircase (first escape route) based on the shortest distance. The left side shows the "Smoke Distribution - Smoke Detector Activation" screen, which visualizes the spread of smoke over time. The right side shows the "Evacuation Distribution" screen, which displays the locations of occupants and evacuation routes on the third-floor floor plan of the building. Looking at the top left of Figure 6a (8 seconds after the fire started), smoke has begun to appear around the point of origin of the fire. The building floor plan shows a staircase at the top, which corresponds to the first escape route.

[0093] Looking at the bottom left of Fig. 6a (15 seconds after the fire started), smoke fills the room at the point of origin of the fire, and users are leaving the occupied space to evacuate to the first escape route. Looking at the top right of Fig. 6a (32 seconds after the fire started), smoke has spread extensively into the occupied space where users were staying; while some users have completed evacuation, it can be seen that a bottleneck has occurred at the entrance of the first escape route. Looking at the bottom right of Fig. 6a (58 seconds after the fire started), smoke has spread beyond the occupied space to near the entrance of the first escape route, and at this point, it is confirmed that 5 out of 39 users have not completed evacuation and are remaining at the entrance of the first escape route. This suggests that passage through the stairs was delayed due to the bottleneck, and that some individuals may have had difficulty evacuating due to exposure to smoke. According to the simulation results, the evacuation completion time for the upper stairs was recorded as 60.8 seconds.

[0094] Figure 6b illustrates a scenario in which a group of users is dispersed across multiple evacuation routes. Looking at the top left of Figure 6b (8 seconds after the fire started) and the bottom left of Figure 6b (15 seconds after the fire started), users are leaving the occupied space, and up to this point, the situation is identical to that of Figure 6a. Looking at the top right of Figure 6b (32 seconds after the fire started), users located close to the first evacuation route are evacuating via the first evacuation route, while users located far from the first evacuation route are evacuating via the second evacuation route. Looking at the building floor plan, a path showing some occupants moving downwards is indicated, confirming that dispersed evacuation to the second evacuation route is taking place.

[0095] Looking at the bottom right of Fig. 6b (58 seconds after the fire started), it can be seen that all users heading toward the first evacuation route have completed evacuation, and users heading toward the second evacuation route have arrived near the second evacuation route. At this time, compared to the evacuation distribution in the bottom right of Fig. 6a where 34 out of 39 people completed evacuation, it is confirmed that only 29 out of 39 people completed evacuation in the evacuation distribution in the bottom right of Fig. 6b. Although the number of evacuees actually decreased during the same period, it can be confirmed that all users evacuating toward the first evacuation route completed evacuation, meaning there were no users exposed to smoke or in dangerous situations. This implies that the bottleneck in the first evacuation route was resolved, allowing evacuation along that route to be completed quickly.

[0096] According to the simulation results in Fig. 6b, as a result of dispersing users for evacuation, the total evacuation completion time increased by 9.7 seconds to 70.5 seconds compared to 60.8 seconds in Fig. 6a; however, the time for complete evacuation of personnel at the first evacuation route was reduced by 15.8 seconds to 45.0 seconds compared to 60.8 seconds in Fig. 6a. In other words, although the total evacuation time increased slightly due to the longer evacuation routes of some users, it means that the safety of personnel evacuating to the first evacuation route was significantly improved by avoiding bottlenecks. Users evacuating to the first evacuation route had their exposure time to smoke reduced, and the risk of being crushed or falling due to congestion was also reduced. Additionally, users guided to the second evacuation route were able to evacuate safely through an uncongested route, even though their travel distance increased. These simulation results empirically demonstrate that route assignment that comprehensively considers risk and congestion, rather than the shortest physical distance, improves overall evacuation safety.

[0097] The central processing unit (700) can receive user attribute information including at least one of the user's age, disability status, and movement speed from the user terminal (1), or determine user attribute information through at least one of video analysis using an IoT sensor (300) or a camera, and calculate an optimal evacuation path by differentiating path generation criteria according to the user attribute information. The user attribute information is an indicator representing each user's evacuation ability, reflecting that even in the same fire situation, the time required for evacuation and preferred path may differ for each user.

[0098] The central processing unit (700) can collect user attribute information in various ways. There are methods of receiving it directly from the user terminal (1) and methods through IoT sensors (300). For example, if the user is wearing a smartwatch or fitness band, information such as heart rate, number of steps, and average movement speed can be collected from the device. By analyzing this information, the user's physical activity level and mobility ability can be estimated. In addition, sensors installed at the entrance gate or elevator of a building can detect the passage of a wheelchair or stroller, thereby identifying that the user is using a walking aid. A video analysis method using a camera can also be used. By analyzing the video from CCTV cameras installed in hallways, lobbies, entrances, etc., within the building in real time, the user's physical characteristics and movement patterns can be identified. By utilizing deep learning-based object recognition technology, assistive devices such as wheelchairs, canes, and strollers can be automatically identified in the video.

[0099] The central processing unit (700) can differentiate path generation criteria based on the collected user attribute information. Differentiating path generation criteria means that, rather than applying the same path selection algorithm to all users, the elements prioritized when selecting a path are set differently according to each user's attributes. For example, for a group of general adults with fast movement speeds, the goal of minimizing arrival time is set as the priority, and a path with a low risk can be selected even if the physical distance is somewhat long. This is because, due to their fast movement speeds, they can evacuate within a reasonable time even if the path is somewhat long. On the other hand, for a group of the elderly or children with slow movement speeds, a path with a short physical distance is selected as a priority, but the safety of the path is verified through a dynamic risk map. Since their slow movement speeds take a long time to travel long distances, and thus the time exposed to fire risk increases, it is advantageous for them to evacuate quickly via the shortest possible path. However, if even a short path is dangerous due to smoke or flames, an alternative path is provided.

[0100] For users with disabilities, route generation criteria are differentiated more specifically. Wheelchair users must select only routes that do not include stairs, and routes utilizing elevators or ramps are provided. Generally, the use of elevators is prohibited during a fire, but they may be utilized in buildings equipped with specially designed evacuation elevators. In the absence of evacuation elevators, wheelchair users can move between floors via ramps, or, if evacuation is not possible, be guided to a safe refuge area to wait for rescue teams.

[0101] The central processing unit (700) can classify users into multiple groups based on their mobility. For example, the first group (high-speed group) consists of users with a mobility speed of 1.2 m / s or more, and they are assigned a route that can be reached in the shortest time. The second group (standard group) consists of users with a mobility speed of 0.8 m / s to 1.2 m / s, and they are assigned a balanced route that prioritizes safety while considering congestion. The third group (low-speed group) consists of the elderly or children with a mobility speed of less than 0.8 m / s, and they are assigned a route that minimizes stair sections or has a low gradient. Since stairs are particularly difficult obstacles for the elderly, if possible, a ramp or a route that allows evacuation within a single floor is prioritized. The fourth group (special group) consists of people with mobility impairments, i.e., wheelchair users or people with severe disabilities, and they are assigned a route that includes only elevators or ramps, or if evacuation is impossible, they are guided to a safe evacuation space and their location is reported to the rescue team.

[0102] This user attribute-based path differentiation enables all users to be guided to the optimal path suited to their physical abilities and limitations, thereby implementing an inclusive evacuation system that allows everyone to evacuate safely without anyone being left out in the event of a fire.

[0103] FIG. 7 shows a flowchart of a dynamic evacuation route provision method, which is another embodiment of the present invention. Referring to FIG. 7, the dynamic evacuation route provision method may include the steps of receiving environmental data in real time (S100), loading building information model data (S200), performing a simulation (S300), generating a dynamic risk map in real time (S400), calculating an individualized optimal evacuation route (S500), and transmitting to a user terminal (S600). The dynamic evacuation route provision method corresponds to the operation method of the intelligent fire response system (10) described above, and each step may be performed by a central processing unit (700).

[0104] The step of receiving environmental data in real time (S100) can receive at least one environmental data among temperature, smoke, and harmful gas in real time from a plurality of IoT sensors installed inside the building. The configuration of the IoT sensors (300) and the types of environmental data are as described above.

[0105] The step of loading building information model data (S200) can load building information model data when a fire is detected by an IoT sensor. The loaded BIM data includes 3D shape information, spatial attribute information, and evacuation path information, as described above.

[0106] The step of performing the simulation (S300) can perform a fire spread simulation that predicts the three-dimensional spatial spread of fire and hazardous substances over time using building information model data and environmental data, and an evacuation movement simulation that predicts the congestion level of each path at different times according to the evacuation paths of multiple users. The specific method of performing the simulation is as described in detail above.

[0107] The step of generating a dynamic risk map in real time (S400) can generate a dynamic risk map in real time by integrating fire spread simulation results and evacuation movement simulation results to quantify the risk level of each unit space within the building. The dynamic risk map is formed by dividing each floor of the building into a two-dimensional grid and assigning a risk index to each grid cell, as previously described.

[0108] The step of calculating an individualized optimal evacuation route (S500) can calculate an individualized optimal evacuation route for each user based on a dynamic risk map and each user's current location. The step of transmitting to a user terminal (S600) can transmit the calculated optimal evacuation route to the corresponding user terminal.

[0109] Although the present invention has been described in detail above through representative embodiments, those skilled in the art will understand that various modifications can be made to the above-described embodiments within the scope of the present invention. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined by the claims set forth below as well as all modifications or variations derived from the claims and equivalent concepts. Explanation of the symbols

[0111] 1: User terminal 10: Intelligent Fire Response System 100: BIM Database 300: IoT Sensor 500: Data Transmitter / Receiver 700: Central Processing Unit

Claims

Claim 1 A BIM database storing building information model data including three-dimensional shape information, spatial attribute information, and possible evacuation routes of a building; a plurality of IoT sensors installed inside the building to detect at least one environmental data among temperature, smoke, and harmful gases in real time; and a data transmission and reception unit receiving user information including user location information from a user terminal;and includes a central processing unit that operates in conjunction with the above BIM database, IoT sensor, and data transmission / reception unit. When a fire is detected by the IoT sensor, the central processing unit performs in real time a fire spread simulation that predicts the spread of fire and hazardous substances using a computational fluid dynamics (CFD)-based model utilizing the building information model data and the environment data, and an agent-based evacuation movement simulation that models each of multiple users as an independent agent and calculates the congestion level of each path by time using a velocity-density relationship based on congestion. It generates a dynamic risk map that quantifies the risk level of each unit space within the building by integrating the fire risk derived from the fire spread simulation results and the congestion level derived from the evacuation movement simulation results. It calculates an optimal evacuation path individualized for each user based on the dynamic risk map and each user's current location, classifying multiple users located within the same space into a proximity group, an intermediate group, and a distance group based on their distance from the first exit. It assigns a path to the first exit to the proximity group and a path to the second exit, which is further than the first exit, to the distance group. For the intermediate group, the first An intelligent fire response system characterized by evaluating the congestion based on the evacuation movement simulation and the risk based on the fire spread simulation for each of the path to the exit and the path to the second exit, assigning the path with the lower weight, updating the fire spread simulation and the evacuation movement simulation in real time and updating the dynamic risk map by reflecting environmental data continuously received from the IoT sensor and user location information continuously received from the data transmission / reception unit during the fire, and re-selecting the path of the intermediate group based on the updated dynamic risk map. Claim 2 An intelligent fire response system according to claim 1, wherein the central processing unit updates the fire spread simulation and the evacuation movement simulation by reflecting user location information continuously received from the data transmission and reception unit and environmental data continuously received from the IoT sensor during the progress of the fire, updates the dynamic risk map by integrating the updated simulation results, and recalculates and updates the optimal evacuation path based on the updated dynamic risk map. Claim 3 An intelligent fire response system according to claim 1, wherein the central processing unit converts the possible evacuation path extracted from the building information model data into a graph structure composed of nodes and edges, dynamically updates the cost of each edge according to the dynamic risk map, and calculates the optimal evacuation path for each user using a shortest path search algorithm based on the updated cost. Claim 4 An intelligent fire response system according to paragraph 3, wherein the central processing unit sets the weight (Cost) by reflecting the risk level of the corresponding path, and calculates the path with the minimum sum of weights as the optimal evacuation path using the shortest path search algorithm. Claim 5 An intelligent fire response system according to claim 4, wherein the central processing unit uses the Dijkstra algorithm or the A* (A-star) algorithm as the shortest path search algorithm. Claim 6 An intelligent fire response system according to claim 1, wherein the central processing unit calculates the personnel density of each path by time period from the evacuation movement simulation results, identifies a section where the personnel density exceeds a preset threshold density as a bottleneck section, and reassigns some of the users heading toward the bottleneck section to another path. Claim 7 An intelligent fire response system according to claim 1, wherein the central processing unit receives user attribute information including at least one of the user's age, disability status, and movement speed from a user terminal, or determines the user attribute information through at least one of image analysis using an IoT sensor or a camera, and calculates an optimal evacuation path by differentiating path generation criteria according to the user attribute information. Claim 8 delete Claim 9 delete Claim 10 A method for providing a dynamic evacuation route using an intelligent fire response system comprises: receiving in real time at least one environmental data among temperature, smoke, and hazardous gas from a plurality of IoT sensors installed inside a building, and receiving user information including user location information from a user terminal; when a fire is detected by the IoT sensors, loading building information model data including three-dimensional shape information of the building, spatial attribute information, and possible evacuation routes; using the building information model data and the environmental data, performing in real time a fire spread simulation that predicts the three-dimensional spatial spread of fire and hazardous substances over time using a computational fluid dynamics (CFD)-based model, and an agent-based evacuation movement simulation that models each of a plurality of users as independent agents and calculates the congestion level of each route by time using a velocity-density relationship based on congestion; generating in real time a dynamic risk map that quantifies the risk level of each unit space within the building by integrating the fire risk level derived from the fire spread simulation results and the congestion level derived from the evacuation movement simulation results; and calculating an optimal evacuation route individualized for each user based on the dynamic risk map and each user's current location. A step of transmitting the calculated optimal evacuation route to the corresponding user terminal;A method for providing a dynamic evacuation route, comprising: a step of calculating the optimal evacuation route, wherein a plurality of users located within the same space are classified into a proximity group, an intermediate group, and a distance group based on their distance from a first exit; a route to the first exit is assigned to the proximity group, and a route to a second exit further than the first exit is assigned to the distance group; for the intermediate group, the congestion level according to the evacuation movement simulation and the risk level according to the fire spread simulation are evaluated for each of the route to the first exit and the route to the second exit, and the route with the lower weight is assigned; the fire spread simulation and the evacuation movement simulation are updated in real time and the dynamic risk map is updated by reflecting environmental data continuously received from the IoT sensor and user location information continuously received from the user terminal during the fire; and the route of the intermediate group is re-selected based on the updated dynamic risk map.

Images