System

Abstract

To provide a system capable of providing a quick and accurate evacuation instruction to a user and performing appropriate training even in normal times.SOLUTION: The system includes means for acquiring images and video feeds of a scene in real-time and pre-processing the same, means for analyzing the acquired AI using a generated AI model and evaluating the progress of a hazard, means for acquiring location information of a user and generating a safest evacuation route based on the analysis result, means for transmitting the generated evacuation route information to the user's device and providing audio and visual evacuation instructions, means for generating a virtual-training scenario and implementing an interactive Evacuation drill with the user, and means for continuously updating the generated data model based on the utilization of the system and new hazard patterns.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The technology of the present disclosure relates to a system. [Background technology]

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] Conventional evacuation systems often struggle to grasp the situation in real time and provide prompt instructions during an emergency, leading to confusion among evacuees. Furthermore, lack of training can make it difficult for evacuees to take appropriate action in an actual emergency. Furthermore, many systems do not adequately address specific user demographics, and often lack consideration for people with disabilities in particular. The present invention aims to solve these problems by providing faster and more appropriate evacuation instructions, enabling users to take appropriate evacuation actions even in an actual emergency. [Means for solving the problem]

[0005] The present invention provides a system for supporting safe evacuation in emergency situations. The system includes means for acquiring and preprocessing images and video feeds from the scene in real time. It also includes means for analyzing the acquired data using a generative AI model to assess the progress of the disaster. It also includes means for acquiring a user's location information and generating the safest evacuation route based on the analysis results. It also includes means for transmitting the generated evacuation route information to a user's device and providing audio and visual evacuation instructions. It also includes means for generating virtual training scenarios and conducting interactive evacuation drills with the user. It also includes means for analyzing past disaster data and training data to assess risks and propose preventative measures for specific areas. It also includes means for continuously updating the generative AI model based on system usage data and new disaster patterns. This allows the system to provide users with fast and accurate evacuation instructions, enabling appropriate training even during normal times.

[0006] An "emergency" is a situation that requires a prompt and appropriate response due to a disaster, accident, etc.

[0007] An "evacuation support system" is a system used to encourage people to evacuate safely in an emergency.

[0008] "Real-time" means that data is collected and processed at the moment.

[0009] "Image and video feed" refers to still images and video data obtained from cameras and sensors.

[0010] "Preprocessing" refers to the process of organizing and converting data before analyzing it.

[0011] A "generative AI model" is a model that uses machine learning algorithms to find specific patterns and trends in data.

[0012] "Analysis" is the process of examining data in detail and understanding its meaning and trends.

[0013] "Disaster Progression" is information that shows how the disaster is changing over time.

[0014] "User location information" is data indicating the user's current location.

[0015] An "evacuation route" refers to a route that a user can take to safely evacuate.

[0016] "Visual and audio evacuation instructions" refers to providing the user with visual (text and images) and audio evacuation instructions.

[0017] A "virtual training scenario" is a virtual training scenario that simulates a real-life situation.

[0018] An "interactive evacuation drill" is an evacuation drill in which users participate and interact with the system.

[0019] "Past disaster data" refers to data relating to disasters that have occurred in the past.

[0020] "Risk assessment" is the process of analyzing and evaluating the risks in a particular area or situation.

[0021] "Preventive measures" are measures and policies to prevent disasters and accidents from occurring.

[0022] "Usage Data" means data relating to the usage of the System.

[0023] "Disaster patterns" are information that indicates the characteristics and trends of when a disaster occurs.

[0024] An "update" is an update process to keep systems and software up to date. [Brief explanation of the drawings]

[0025] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0026] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0027] First, the terms used in the following description will be explained.

[0028] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0029] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0030] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0031] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0032] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0033] [First embodiment]

[0034] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0035] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0036] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0037] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0038] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0039] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0040] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0041] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0042] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0043] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0044] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0045] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0046] The AI ​​emergency evacuation support system of the present invention is designed to support safe evacuation in emergency situations. The system functions through cooperation between a server, a terminal, and a user. The program's processing is explained below with specific examples.

[0047] Basic components

[0048] Server: Responsible for real-time acquisition and pre-processing of images and video feeds, data analysis using generative AI models, generation of evacuation routes, generation of virtual training scenarios, data analysis and prevention measures proposals, and usage data updates.

[0049] Terminal: Responsible for obtaining user location information, providing real-time evacuation instructions, and conducting virtual drills.

[0050] User: Follow instructions from the device to take evacuation action and participate in virtual training.

[0051] System action

[0052] Real-time situation analysis

[0053] 1. The server receives real-time images and video feeds from cameras and sensors within the facility.

[0054] 2. The server preprocesses the acquired data and removes unnecessary noise.

[0055] 3. The server uses a generative AI model to analyze the pre-processed data and assess the spread of the fire, smoke density, and fire scale.

[0056] Examples:

[0057] The server analyzes footage from surveillance cameras in real time and detects when a fire has broken out in a specific area.

[0058] The server analyzes the spread of smoke and identifies dangerous areas.

[0059] Obtaining user location information

[0060] 1. The device (mainly a smartphone or tablet) obtains the user's current location information using the GPS function.

[0061] 2. The device sends the acquired location information to the server.

[0062] Examples:

[0063] The user launches the app and allows location sharing.

[0064] The device obtains the user's current latitude and longitude and sends them to the server.

[0065] Generate evacuation routes

[0066] 1. The server generates the optimal evacuation route based on the results of real-time situation analysis and the user's location information.

[0067] 2. The server sends the generated evacuation route information to the terminal.

[0068] Examples:

[0069] The server calculates the route from the user's current location to the nearest emergency exit, taking into account the spread of the fire and the density of the smoke.

[0070] The server transmits map data showing evacuation routes to the terminal.

[0071] Evacuation instructions to users

[0072] 1. The device provides the user with visual and audio evacuation instructions based on the received evacuation route information.

[0073] 2. The user follows the instructions on the device to evacuate safely.

[0074] Examples:

[0075] The device will then provide voice instructions such as, "Turn right, walk 30 meters, and go down the stairs."

[0076] The user follows the instructions on the device and takes evacuation action.

[0077] Virtual Training and Simulation

[0078] 1. The server generates a scenario for virtual training and sends it to the terminal.

[0079] 2. The terminal notifies the user that the virtual training has started and displays the training scenario.

[0080] 3. Users participate in virtual drills and take evacuation actions according to simulated scenarios.

[0081] Examples:

[0082] The server generates a hypothetical scenario in which a fire breaks out on the second floor of a building.

[0083] The device instructs the user to put on the VR goggles and begin training.

[0084] Users practice evacuation plans in a virtual environment.

[0085] Data analysis and preventative measures

[0086] 1. The server collects past disaster data and training data and analyzes them using a generative AI model.

[0087] 2. The server assesses the fire risk in a particular area and suggests preventative measures.

[0088] Examples:

[0089] The server analyzes past fire data to identify fire frequency and high-risk areas.

[0090] The server will suggest preventative measures to the area administrator.

[0091] Continuous learning and improvement

[0092] 1. The server collects system usage data and new disaster patterns and continuously updates the generative AI model.

[0093] 2. The server learns new algorithms and methods for optimizing evacuation routes and reflects them in the system.

[0094] Examples:

[0095] The server analyzes data from annual evacuation drills and learns that certain evacuation routes are prone to congestion.

[0096] The server will propose a new evacuation route and update the system accordingly.

[0097] In this way, the AI ​​emergency evacuation support system of the present invention supports users in safely evacuating through real-time situation assessment, analysis, instructions, training, and continuous learning.

[0098] The processing flow will be explained below.

[0099] Step 1:

[0100] The server collects real-time data from surveillance cameras and sensors installed within the facility, including still images, video feeds, and temperature sensor data.

[0101] Step 2:

[0102] The server pre-processes the collected image and video feed data to remove noise and improve image clarity, which improves the accuracy of the analysis.

[0103] Step 3:

[0104] The server uses a generative AI model to analyze the pre-processed data and evaluate the location of the fire, its spread, smoke density, and fire scale, thereby determining the level of urgency at the scene.

[0105] Step 4:

[0106] The device obtains the user's current location information using the GPS function. If the user carries a smartphone or tablet, the location information can be updated continuously through the app.

[0107] Step 5:

[0108] The device sends the acquired user location information to the server, which allows the server to determine the user's current location.

[0109] Step 6:

[0110] The server then calculates the safest evacuation route based on the analysis results and the user's location information, taking into account factors such as the spread of the fire, smoke density, and congestion on evacuation routes.

[0111] Step 7:

[0112] The server then sends the generated evacuation route information to the user's device. The sent data includes the specific evacuation route and its details (e.g., how many meters to go, which direction to turn, etc.).

[0113] Step 8:

[0114] Based on the received evacuation route information, the device provides the user with visual and audio evacuation instructions, including map display and audio guidance.

[0115] Step 9:

[0116] The user follows the instructions on the device to begin evacuation along a safe evacuation route. By following the instructions, the user can evacuate safely.

[0117] Step 10:

[0118] The server generates virtual training scenarios during normal times and transmits them to the terminals. Specific training scenarios include fire outbreak scenarios and evacuation scenarios.

[0119] Step 11:

[0120] The device notifies the user of the start of virtual training and prompts them to train in the virtual environment. Users can participate in the virtual training using VR goggles or a smartphone.

[0121] Step 12:

[0122] Users participate in virtual training and take evacuation actions according to simulated fire scenarios. After the training, they receive feedback from the system to identify areas for improvement in their evacuation actions.

[0123] Step 13:

[0124] The server collects past disaster data and training data and analyzes it using a generative AI model, which then generates a risk assessment for a specific area.

[0125] Step 14:

[0126] Based on the risk assessment results, the server will suggest preventative measures to managers of specific areas, such as installing sprinkler systems and strengthening evacuation drills.

[0127] Step 15:

[0128] The server collects system usage data and new disaster patterns to continuously update the generative AI model, providing evacuation routes and preventative measures that reflect the latest knowledge.

[0129] Example 1

[0130] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0131] While quick and safe evacuation is extremely important in an emergency, current systems have difficulty obtaining disaster information in real time and providing optimal evacuation routes. Furthermore, inaccurate location information for users and lack of evacuation drills could lead to confusion during an actual disaster. Furthermore, past disaster data and drill data are not fully utilized, preventing preventative measures from being proposed and the system from being continuously improved.

[0132] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0133] In this invention, the server includes means for acquiring and preprocessing images and video feeds of the scene in real time, means for analyzing the acquired data using a generative AI model to evaluate the progress of the disaster, means for acquiring user location information and generating the safest evacuation route based on the analysis results, means for transmitting the generated evacuation route information to the user's device and providing audio and visual evacuation instructions, means for generating virtual training scenarios and conducting interactive evacuation drills with the user, means for analyzing past disaster data and training data to assess risk and propose preventive measures for specific areas, means for continuously updating the generative AI model based on system usage data and new disaster patterns, means for acquiring location information using the GPS function of the user's device, means for analyzing the spread of fire and smoke concentration using the generative AI model, and means for notifying the user of the start of the virtual training and displaying the training scenario using VR goggles, thereby enabling real-time situation assessment and evacuation instructions to be given, thereby achieving safe evacuation.

[0134] "Images and video feeds" refers to visual information obtained in real time from surveillance cameras and sensors.

[0135] "Preprocessing" refers to the preparation of data before analysis, such as removing noise from the acquired data and correcting outliers.

[0136] A "generative AI model" is an artificial intelligence model that uses machine learning and deep learning algorithms to analyze data.

[0137] "Disaster progression" refers to changes in the situation and scope of impact in emergencies such as fires, earthquakes, and floods.

[0138] "User location information" refers to the user's current geographic coordinate information obtained using the GPS function.

[0139] An "evacuation route" is the optimal route to safety from an emergency situation.

[0140] An "evacuation instruction" is an audio and visual guide that encourages the user to take evacuation action.

[0141] A "virtual training scenario" is a scenario that allows a user to conduct an evacuation drill using virtual reality or simulation.

[0142] "Past disaster data" refers to records and information about disasters that have occurred in the past.

[0143] "Risk assessment" is the evaluation of the likelihood and impact of a disaster in a particular area or situation.

[0144] "Preventive measures" are measures proposed to prevent disasters from occurring.

[0145] "System Usage Data" means all data collected through the System.

[0146] "New disaster patterns" refer to new trends and characteristics of disasters that have never been seen before.

[0147] "GPS function" refers to a function that obtains location information using the Global Positioning System.

[0148] "Smoke density" is an indicator of the density or amount of smoke present in the air.

[0149] "VR goggles" are devices that allow users to experience a virtual reality environment.

[0150] This invention relates to an AI emergency evacuation support system that supports safe evacuation in emergency situations. The system of the invention functions through cooperation between a server, a terminal, and a user.

[0151] First, the server acquires real-time images and video feeds from cameras and sensors within the facility. Examples of hardware used include surveillance cameras and smoke detection sensors. The data obtained from these devices undergoes pre-processing, including noise removal and outlier correction. Specifically, video data is filtered and sensor data is corrected to ensure data quality.

[0152] The pre-processed data is then analyzed using generative AI models to assess fire spread, smoke density, and fire scale. For example, image analysis can detect the location and spread of a fire in real time.

[0153] The devices are primarily smartphones and tablets, which use GPS to obtain the user's current location information. The location information is then sent from the device to a server, which then uses it to calculate evacuation routes.

[0154] The server generates an optimal evacuation route based on the results of real-time situation analysis and the user's location information. The generated evacuation route information is sent to the device. Based on this information, the device provides the user with visual and audio evacuation instructions. For example, specific instructions such as "Turn right, walk 30 meters, and go down the stairs" are provided by audio guidance.

[0155] The server also generates scenarios for virtual training and sends them to the device. At this time, the user receives a notification from the device and begins the virtual training. By wearing VR goggles, evacuation training can be carried out in the virtual environment. As a specific example, a virtual scenario is generated in which "a fire breaks out on the second floor of a building," and the user conducts virtual training based on this scenario.

[0156] The server collects past disaster data and training data and analyzes it using a generative AI model. This allows it to assess the fire risk in a specific area and propose preventative measures. For example, it can analyze past fire data to identify patterns of fire occurrence in specific time periods or areas and propose measures to strengthen fire prevention equipment.

[0157] Furthermore, the server continuously updates the generative AI model based on system usage data and new disaster patterns. By learning new algorithms and evacuation route optimization methods and incorporating them into the system, the system can learn that certain evacuation routes are prone to congestion based on annual training data and suggest new evacuation routes.

[0158] As a result, the AI ​​emergency evacuation support system of the present invention supports users in safely evacuating through real-time situation assessment, analysis, instructions, training, and continuous learning.

[0159] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0160] Step 1:

[0161] The server receives real-time images and video feeds from cameras and sensors within the facility. This input data includes surveillance camera footage and smoke detection sensor data. Specifically, it collects camera stream data and converts it into a different format. The output is raw image and video data that has not yet been preprocessed.

[0162] Step 2:

[0163] The server preprocesses the acquired data and removes unwanted noise. The input data is the image and video data obtained in step 1, and a filtering algorithm is applied to remove noise. Specifically, a low-pass filter is applied to the video data, and outlier correction is performed on the sensor data. The output is high-quality data with noise removed.

[0164] Step 3:

[0165] The server uses a generative AI model to analyze the preprocessed data and evaluate the spread of the fire, smoke density, and fire scale. The input data is the output of step 2, and is fed into the generative AI model for analysis. Specifically, the deep learning model is used to identify the extent of the fire's expansion and calculate smoke density. The output is the progress of the fire and the risk assessment results.

[0166] Step 4:

[0167] The device (smartphone or tablet) obtains the user's current location information using its GPS function. Input data is obtained from the user's device and includes latitude and longitude information. Specifically, it collects real-time data from the GPS sensor to obtain location information. The output is the user's precise location information.

[0168] Step 5:

[0169] The device sends the acquired location information to the server. The input data is the location information from step 4, which is encrypted using a security protocol and sent to the server. The specific operation is to establish a secure communication channel and send the location information. The output is the location information received by the server.

[0170] Step 6:

[0171] The server generates an optimal evacuation route based on the results of real-time situation analysis and the user's location information. The input data are the risk assessment results from step 3 and the location information from step 5. Specifically, it applies a route calculation algorithm to generate a safe evacuation route. The output is evacuation route information customized for each user.

[0172] Step 7:

[0173] The server sends the generated evacuation route information to the terminal. The input data is the evacuation route information from step 6, which is sent to the user's terminal. Specifically, the server sends the evacuation route map in digital format. The output is the evacuation route information displayed on the terminal.

[0174] Step 8:

[0175] The device provides the user with visual and audio evacuation instructions based on the received evacuation route information. The input data is the evacuation route information from step 7, and this is notified to the user through a visual map display and audio guidance. Specifically, the device provides audio guidance of direction and distance information through the application. The output is the evacuation instructions received by the user.

[0176] Step 9:

[0177] The user follows the instructions on the terminal to evacuate safely. The input data is the evacuation instructions from step 8, and upon receiving this, the process moves on to actual evacuation behavior. Specifically, the user follows the instructions on the terminal and evacuates by following a safe route. The output is to reach a safe area.

[0178] (Application example 1)

[0179] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0180] Conventional evacuation support systems have difficulty responding to emergencies in real time, and have been unable to provide effective evacuation support, especially in large facilities or complex environments. Furthermore, it has been difficult to provide appropriate evacuation instructions to users with visual or hearing impairments. Furthermore, there has been a problem in that preventive measures and continuous improvements have not been proposed using past training data or disaster data.

[0181] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0182] In this invention, the server includes: means for acquiring and preprocessing images and video feeds of the scene in real time; means for analyzing the acquired data using a generative AI model to assess the progress of the disaster; means for acquiring user location information and generating the safest evacuation route based on the analysis results; means for transmitting the generated evacuation route information to the user's device and providing audio and visual evacuation instructions; means for generating virtual training scenarios and conducting interactive evacuation drills with the user; means for analyzing past disaster data and training data to assess risk and propose preventative measures for specific areas; means for continuously updating the generative AI model based on system usage data and new disaster patterns; and means installed in smart glasses for generating an optimal evacuation route based on the user's current location and providing visual and audio evacuation instructions. This enables effective evacuation support even in large facilities and complex environments, and provides appropriate evacuation instructions to visually and hearing-impaired individuals. It also enables the system to propose preventative measures and continuously improve the system by utilizing past training data and disaster data.

[0183] An "emergency" is an event that poses a threat to people's safety and lives, such as an unexpected disaster or accident.

[0184] An "evacuation support system" is a system that includes a series of technologies and devices to assist people in safely evacuating in the event of an emergency.

[0185] "On-site images and video feed" refers to real-time visual information of the location and surroundings of an emergency.

[0186] "Preprocessing" refers to the process of removing unnecessary noise from acquired image or video data and converting it into a format that is easier to analyze.

[0187] A "generative AI model" is an artificial intelligence model that learns from large amounts of data and has the ability to analyze situations and generate appropriate evacuation routes.

[0188] "Disaster Progression" refers to the process of damage expansion and current situation in emergencies such as fires and earthquakes.

[0189] "User Location Information" means an individual user's current geographic location obtained using GPS or other location-determining technology.

[0190] An "evacuation route" refers to a route that guides a user to the safest evacuation route in the event of an emergency.

[0191] "Terminal" refers to a mobile information terminal used by a user, such as a smartphone, tablet, or smart glasses.

[0192] A "virtual training scenario" refers to a scenario that uses virtual reality and simulation to allow people to realistically experience an emergency evacuation drill.

[0193] "Past disaster data" refers to records and data relating to disasters that have occurred in the past.

[0194] "Training data" refers to data collected during evacuation drills and is used to improve the system and assess risks.

[0195] "Risk assessment" is the act of assessing the possibility of a disaster occurring in a specific area and the extent of its impact.

[0196] "Preventive measures" refer to strategies and measures to reduce risk and ensure safety before a disaster occurs.

[0197] "Continuous updates" refers to the process by which systems and generative AI models learn new data and techniques, keeping them constantly up to date.

[0198] The present invention provides a system for supporting safe evacuation in an emergency situation, which functions in cooperation with a server, terminals, and users.

[0199] Basic components

[0200] 1. Server: Acquires and preprocesses images and video feeds from the scene in real time. It also analyzes the data using a generative AI model to assess the progress of the disaster. It generates the safest evacuation route based on the user's location information and analysis results and sends this to the device. It also generates virtual training scenarios and analyzes past disaster data and training data to assess risks and propose preventive measures. It is responsible for continuously updating the generative AI model based on system usage data and new disaster patterns.

[0201] 2. Terminal: This is a user device, typically a smart eyeglass or smartphone, that acquires the user's current location. It provides visual and audio evacuation instructions based on evacuation route information received from the server. It also provides training scenarios and conducts interactive training with the user during virtual training.

[0202] 3. User: Follows instructions from the terminal to take evacuation action. Participates in virtual training and acts according to the evacuation plan provided by the system.

[0203] System action

[0204] Hardware and Software

[0205] On-site cameras and sensors: Capture images and video feeds in real time.

[0206] GPS module: Obtains the user's location information.

[0207] Smart glasses: Provide visual and audio evacuation instructions.

[0208] Server: Responsible for data processing and running generative AI models.

[0209] Data processing and calculation

[0210] 1. The server acquires images and video feeds from on-site cameras and sensors in real time and performs pre-processing such as noise removal. The pre-processed data is then analyzed using a generative AI model to assess the progression of the disaster.

[0211] 2. The device acquires the user's location information using GPS and sends it to the server, which then combines this information with the analysis results to generate the optimal evacuation route.

[0212] 3. The server sends the generated evacuation route to the terminal, and the terminal provides the user with visual and audio evacuation instructions.

[0213] 4. The virtual training scenario is generated by the server and sent to the user's device. The user wears the smart glasses and conducts an interactive evacuation drill.

[0214] Examples of concrete examples and prompts

[0215] For example, if a fire breaks out in a factory, the server will analyze the footage from the factory's surveillance cameras in real time to identify the location and spread of the fire. It will then calculate the safest evacuation route based on the user's current location and provide visual and audio evacuation instructions through the smart glasses. The following prompt sentences are used:

[0216] The server analyzes the footage from the factory's surveillance cameras in real time and detects a fire in Section 4. Based on the current location of the workers (Section 3) obtained via GPS and the location of the fire, generate the optimal evacuation route. Then, send this evacuation route to the smart glasses and initiate visual and audio instructions.

[0217] In this way, the AI ​​emergency evacuation support system of the present invention can support safe and rapid evacuation in emergency situations through real-time situation analysis, generation of evacuation plans, visual and audio instructions to users, and even virtual training.

[0218] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0219] Step 1:

[0220] Acquiring and pre-processing images and video feeds from the scene

[0221] The server receives real-time image and video feeds from cameras and sensors on-site. The input is raw image and video data from the site. This data is preprocessed to remove unwanted noise, for example by smoothing the image or enhancing edges, to convert it into a format that is easier to analyze. The output is the preprocessed image and video data.

[0222] Step 2:

[0223] Data analysis using generative AI models

[0224] The server inputs preprocessed image and video data into a generative AI model for data analysis. Specifically, it evaluates the spread of the fire, smoke density, and the scale of the disaster. The input is preprocessed image and video data, and the output is various analytical results on the progression of the disaster. The generative AI model accomplishes this by generating data to determine the necessary rescue operations and evacuation routes.

[0225] Step 3:

[0226] Obtaining user location information

[0227] The device obtains the user's current location information using its GPS function. The input is location data from the GPS module, which acquires the user's latitude and longitude information. The device generates a data packet to send this location information to the server and sends it to the server. The output is the user's exact current location, which the server receives.

[0228] Step 4:

[0229] Generate evacuation routes

[0230] The server generates the optimal evacuation route based on the acquired user location information and the disaster analysis results generated by the AI ​​model. For example, it determines the safest route by taking into account the spread of fire and smoke density. The inputs are the user's current location and the disaster analysis results, and the output is the optimal route information for the user to evacuate.

[0231] Step 5:

[0232] Sending and instructing evacuation routes

[0233] The server sends the generated evacuation route information to the terminal. The terminal receives this and provides the user with visual and audio evacuation instructions. Specifically, the terminal issues specific instructions via voice, such as "Turn right, walk 30 meters, and go down the stairs." The input is the evacuation route information sent from the server, and the output is the specific evacuation instructions received by the user.

[0234] Step 6:

[0235] Generate and conduct virtual training scenarios

[0236] The server generates a scenario for the virtual training and sends it to the terminal. The terminal notifies the user of the start of the virtual training and displays the training scenario. The user takes evacuation actions in the virtual environment according to the scenario. The input is the generated virtual training scenario, and the output is the results of the virtual training performed by the user.

[0237] Step 7:

[0238] Data analysis and preventative measures

[0239] The server collects past disaster data and training data and analyzes it using a generative AI model. As a result of the analysis, it performs a risk assessment of a specific area and proposes appropriate preventive measures. For example, it may recommend installing additional fire detectors in areas where fires frequently occur. The inputs are past disaster data and training data, and the output is the risk assessment results and proposed preventive measures.

[0240] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0241] This invention provides flexible evacuation support that takes into account the user's emotional state by combining an emotion engine with a system for supporting safe evacuation in emergency situations. The following describes the system components and program processing with specific examples.

[0242] Basic components

[0243] Server: Responsible for real-time acquisition and pre-processing of images and video feeds, data analysis using generative AI models, evaluation of users' emotional state using an emotion engine, generation of evacuation routes, generation of virtual training scenarios, data analysis and prevention measures proposals, and usage data updates.

[0244] Terminal: Responsible for obtaining user location information, collecting emotion engine data, providing real-time evacuation instructions, and conducting virtual drills.

[0245] User: Follows instructions from the device to take evacuation action, provides emotional data, and participates in virtual training.

[0246] System action

[0247] Real-time situation analysis

[0248] 1. The server receives real-time images and video feeds from cameras and sensors within the facility.

[0249] 2. The server preprocesses the acquired data to remove noise and improve image clarity.

[0250] 3. The server analyzes the pre-processed data using a generative AI model to assess the location and spread of the fire, smoke density, and fire scale.

[0251] Examples:

[0252] The server analyzes footage from surveillance cameras and detects that a fire has broken out in a specific area.

[0253] The server analyzes the spread of smoke and identifies dangerous areas.

[0254] Obtaining user location information

[0255] 1. The device (smartphone or tablet) obtains the user's current location information using the GPS function.

[0256] 2. The device sends the acquired location information to the server.

[0257] Examples:

[0258] The user launches the app and allows location sharing.

[0259] The device obtains the user's current latitude and longitude and sends them to the server.

[0260] Generate evacuation routes

[0261] 1. The server generates the safest evacuation route based on the results of real-time situation analysis and the user's location information.

[0262] 2. The server sends the generated evacuation route information to the terminal.

[0263] Examples:

[0264] The server calculates the route from the user's current location to the nearest emergency exit, taking into account the spread of the fire and the density of the smoke.

[0265] The server transmits map data showing evacuation routes to the terminal.

[0266] Assessing the user's emotional state

[0267] 1. The device uses a camera and microphone to analyze the user's facial expressions and voice and collect emotional data.

[0268] 2. The device sends the collected emotion data to the server.

[0269] 3. The server uses an emotion engine to evaluate the user's emotional state and dynamically adjusts the content and method of evacuation instructions based on the results.

[0270] Examples:

[0271] The device analyzes the user's facial expressions through the camera and detects signs of stress or impatience.

[0272] The server provides more detailed and calming audio guidance to ease the user's anxiety and stress.

[0273] Evacuation instructions to users

[0274] 1. The device provides visual and audio evacuation instructions to the user based on the received evacuation route information and emotional state assessment.

[0275] 2. The user begins evacuating along a safe evacuation route following the instructions on the device.

[0276] Examples:

[0277] The device will then provide voice instructions such as, "Turn right, walk 30 meters, and go down the stairs."

[0278] The user follows the instructions on the device and evacuates along the designated evacuation route.

[0279] Virtual Training and Simulation

[0280] 1. The server generates a scenario for virtual training and sends it to the terminal.

[0281] 2. The device notifies the user that virtual training has begun and prompts them to train in the virtual environment.

[0282] 3. Users participate in virtual drills and take evacuation actions according to simulated scenarios.

[0283] Examples:

[0284] The server generates a hypothetical scenario in which a fire breaks out on the second floor of a building.

[0285] The device instructs the user to put on the VR goggles and begin training.

[0286] Users practice evacuation plans in a virtual environment.

[0287] Data analysis and preventative measures

[0288] 1. The server collects past disaster data and training data and analyzes them using a generative AI model.

[0289] 2. The server assesses the fire risk in a particular area and suggests preventative measures.

[0290] Examples:

[0291] The server analyzes past fire data to identify fire frequency and high-risk areas.

[0292] The server will suggest preventative measures to the area administrator.

[0293] Continuous learning and improvement

[0294] 1. The server collects system usage data and new disaster patterns and continuously updates the generative AI model.

[0295] 2. The server learns new algorithms and methods for optimizing evacuation routes and reflects them in the system.

[0296] Examples:

[0297] The server analyzes data from annual evacuation drills and learns that certain evacuation routes are prone to congestion.

[0298] The server will propose a new evacuation route and update the system accordingly.

[0299] In this way, the AI ​​emergency evacuation support system of the present invention supports users' safe evacuation through real-time situational awareness, data analysis, evaluation of the user's emotional state using an emotion engine, dynamic evacuation instructions, training and simulation, data analysis and prevention measures proposals, and continuous learning and improvement. In addition, the use of an emotion engine enables flexible evacuation support that takes the user's psychological state into consideration.

[0300] The processing flow will be explained below.

[0301] Step 1:

[0302] The server collects real-time images and video feeds from security cameras and sensors installed within the facility, including security camera footage, temperature sensor data, and smoke detector sensor data.

[0303] Step 2:

[0304] The server pre-processes the collected image and video feed data, removing noise, and optionally changes the image resolution or compresses the data.

[0305] Step 3:

[0306] The server analyzes the pre-processed data using a generative AI model to assess the location of the fire, its spread, smoke density, and fire scale, thereby determining the level of urgency at the scene.

[0307] Step 4:

[0308] The device (smartphone or tablet) acquires the user's current location information using the GPS function. When the user launches the emergency app, collection of location information begins.

[0309] Step 5:

[0310] The device sends the acquired user location information to the server, which then calculates an evacuation route based on the user's accurate location information at the time of the disaster.

[0311] Step 6:

[0312] The server calculates the safest evacuation route based on the results of real-time situation analysis and the user's location information, taking into account factors such as the spread of the fire, smoke density, congestion on evacuation routes, and the overall layout of the building.

[0313] Step 7:

[0314] The server then sends the generated evacuation route information to the user's terminal. The evacuation route information includes specific routes and evacuation procedures.

[0315] Step 8:

[0316] Based on the received evacuation route information, the device provides the user with visual and audio evacuation instructions, including displaying a map and providing audio guidance.

[0317] Step 9:

[0318] The user follows the instructions on the device and begins evacuating along the designated evacuation route. By taking appropriate evacuation actions, danger can be minimized.

[0319] Step 10:

[0320] The device uses a camera and microphone to analyze the user's facial expressions and voice to collect emotional data, and an emotion engine evaluates the user's stress level and anxiety state.

[0321] Step 11:

[0322] The device sends the collected emotional data to a server, which allows the user's psychological state to be grasped in real time.

[0323] Step 12:

[0324] The server uses an emotion engine to assess the user's emotional state and dynamically adjusts the evacuation instructions and methods based on the results, providing more detailed and calming guidance when stress and anxiety levels are high.

[0325] Step 13:

[0326] The device receives and provides emotion-dependent instructions to the user, and visual or audio guidance may change depending on the emotional state.

[0327] Step 14:

[0328] By receiving instructions that take emotions into consideration, users can continue their evacuation activities with peace of mind, which is expected to reduce stress.

[0329] Step 15:

[0330] The server generates virtual training scenarios during normal times and transmits them to the terminals. Specific training scenarios include fire outbreak scenarios and evacuation scenarios.

[0331] Step 16:

[0332] The device notifies the user that virtual training has started and prompts them to train in the virtual environment. The user participates in the virtual training using VR goggles or a smartphone.

[0333] Step 17:

[0334] Users participate in virtual training and take evacuation actions according to simulated fire scenarios. After the training, they receive feedback from the system to identify areas for improvement in their evacuation actions.

[0335] Step 18:

[0336] The server collects past disaster data and training data and analyzes it using a generative AI model to assess the fire risk in a specific area.

[0337] Step 19:

[0338] Based on the risk assessment results, the server will suggest preventative measures to managers of specific areas, such as installing sprinkler systems and strengthening evacuation drills.

[0339] Step 20:

[0340] The server collects system usage data and new disaster patterns to continuously update the generative AI model, providing evacuation routes and preventative measures that reflect the latest knowledge.

[0341] Example 2

[0342] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0343] The purpose of this invention is to provide evacuation support in emergency situations by grasping the situation on-site in real time, providing safe and rapid evacuation routes, and providing flexible and effective evacuation instructions that take into account the user's emotional state. It also aims to provide a more advanced evacuation support system by practicing evacuation behavior through virtual training scenarios, proposing preventive measures, and continuously improving the system.

[0344] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0345] In this invention, the server includes means for acquiring and preprocessing images and video feeds of the scene in real time, means for analyzing the acquired data using a generative AI model to evaluate the progress of the disaster, means for acquiring user location information and generating the safest evacuation route based on the analysis results, means for acquiring the user's emotional state using a camera and microphone and evaluating it using an emotion engine, means for providing audio and visual evacuation instructions based on the generated evacuation route information and emotional state evaluation, means for generating virtual training scenarios and conducting interactive evacuation drills with the user, means for analyzing past disaster data and training data to assess risks and propose preventive measures for specific areas, and means for continuously updating the generative AI model based on system usage data and new disaster patterns, thereby enabling real-time situation assessment, provision of safe evacuation routes, emotionally sensitive evacuation instructions, interactive training, and proposal of preventive measures, as well as continuous system improvement.

[0346] An "emergency situation" refers to an emergency situation that requires a rapid response, such as a natural disaster or man-made trouble.

[0347] "Evacuation" refers to the act of escaping danger and moving to a safe place.

[0348] "System" refers to a collection of devices and software in which multiple elements work together to perform a specific function.

[0349] "Image and video feed" refers to still and video data acquired through cameras and sensors.

[0350] "Preprocessing" refers to initial processing such as improving the quality of acquired data and removing noise.

[0351] A "generative AI model" refers to an algorithmic model of artificial intelligence generated using machine learning or deep learning techniques.

[0352] "Analyzing data" refers to the process of extracting information from acquired data and drawing meaningful conclusions.

[0353] "Disaster progression" refers to the extent of damage and the scope of impact in emergencies such as fires and earthquakes.

[0354] "User location information" refers to data about a user's current location obtained by GPS or other positioning systems.

[0355] An "evacuation route" refers to a route that is considered safe for escaping from a disaster.

[0356] "Emotional state" refers to the user's psychological state, such as stress or impatience.

[0357] An "emotion engine" refers to an algorithm or model for analyzing and assessing a user's emotional state.

[0358] "Evacuation instructions" refers to specific instructions or guidance that encourage users to take evacuation action.

[0359] "Virtual training scenario" refers to a virtual training environment and storyline designed around a simulated emergency situation.

[0360] "Interactive evacuation drill" refers to a training process in which users can actively participate within a virtual environment.

[0361] "Historical disaster data" refers to records and information relating to previous emergencies.

[0362] "Training data" refers to information such as behavior and responses collected during training.

[0363] "Risk assessment" refers to the process of quantitatively assessing the likelihood of a disaster occurring in a particular area.

[0364] "Preventive measures" refer to measures and actions taken to minimize the occurrence and impact of disasters.

[0365] "Usage Data" refers to all information collected in the course of using the System.

[0366] "Disaster patterns" refer to patterns or models that show how a particular disaster will unfold based on historical data and statistics.

[0367] An "update" refers to updating an existing model or system by adding new data or algorithms.

[0368] The present invention is a system for supporting safe evacuation in an emergency situation, and is implemented with the following configuration and method.

[0369] System configuration

[0370] The basic components are as follows:

[0371] Server: Responsible for real-time acquisition and pre-processing of images and video feeds, data analysis using generative AI models, evaluation of users' emotional state using an emotion engine, generation of evacuation routes, generation of virtual training scenarios, data analysis and prevention measures proposals, and usage data updates.

[0372] Terminal: Responsible for obtaining user location information, collecting emotion engine data, providing real-time evacuation instructions, and conducting virtual drills.

[0373] User: Follows instructions from the device to take evacuation action, provides emotional data, and participates in virtual training.

[0374] System action

[0375] 1. Data acquisition and preprocessing

[0376] The server receives real-time images and video feeds from cameras and sensors installed within the facility, and uses noise filtering algorithms to remove noise from the data and image enhancement techniques to improve clarity.

[0377] Example: The video stream from a surveillance camera is received by the capturing module and pre-processed by the image processing engine.

[0378] 2. Situation Analysis

[0379] The pre-processed data is analyzed on a server, and a generative AI model is used to assess the location and spread of fires, smoke density, and fire scale, identifying areas with a particularly high probability of fire outbreak.

[0380] Example: A deep learning model detects fires in a specific area and visualizes their spread with 3D mapping.

[0381] 3. Obtain location information

[0382] The device (smartphone or tablet) obtains the user's current location information using the GPS function and sends that information to the server.

[0383] Example: A user launches an app and the device sends GPS data to a server.

[0384] 4. Evacuation route generation

[0385] The server generates the safest evacuation route based on the results of real-time situation analysis and the user's location information using Dijkstra's algorithm and A algorithm. The generated route is sent to the device as map data.

[0386] Example: The server calculates an evacuation route and sends map data to the device.

[0387] 5. Emotional state assessment

[0388] The device uses a camera and microphone to analyze the user's facial expressions and voice to collect emotional data, which is then sent to a server, where an emotion engine is used to evaluate the user's emotional state and dynamically adjust evacuation instructions as needed.

[0389] Example: The device analyzes the user's facial expressions, detects signs of impatience, and the server changes the audio guidance to a calmer version.

[0390] 6. Evacuation instructions

[0391] Based on the received evacuation route information and emotional state assessment, the device provides the user with visual and audio evacuation instructions, which the user follows to begin evacuation along a safe evacuation route.

[0392] Example: The device will give voice instructions such as "Turn right, walk 30 meters, and go down the stairs," and the user will follow the instructions to evacuate.

[0393] 7. Virtual Training

[0394] The server generates a scenario for the virtual training and sends it to the terminal. The terminal notifies the user of the start of the virtual training and prompts the user to participate in the training. The user participates in the virtual training and takes evacuation actions according to the simulated scenario.

[0395] Example: The server generates a virtual scenario in which a fire breaks out on the second floor of a building, and the device instructs the user to put on VR goggles and begin training.

[0396] 8. Data analysis and preventive measures

[0397] The server collects past disaster data and training data, analyzes them using a generative AI model, and evaluates the fire risk in a specific area and suggests preventive measures.

[0398] Example: The server analyzes past fire data and presents improvement proposals for fire prevention measures to managers of high-risk areas.

[0399] 9. Continuous learning and improvement

[0400] The server collects system usage data and new disaster patterns, continually updating the generative AI model, learning new algorithms and methods for optimizing evacuation routes and incorporating them into the system.

[0401] Example: The server analyzes annual evacuation drill data, proposes new evacuation routes, and reflects them in the system.

[0402] Example prompts for generative AI models

[0403] Prompt: "Generate a virtual training scenario, analyze the user's stress level, and suggest the optimal evacuation route for the user. Please provide a detailed description of the specific training scenario and its steps, and finally include evacuation instructions for the user."

[0404] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0405] Step 1:

[0406] Data acquisition and preprocessing

[0407] The server captures real-time images and video feeds from cameras and sensors installed within the facility, then uses noise filtering algorithms to remove noise from the captured data and image enhancement techniques to improve clarity.

[0408] Input: Raw data (images and video feeds) from cameras and sensors

[0409] Output: Pre-processed, clean image and video data

[0410] Specific operation: The capturing module receives the video stream from the camera, and the image processing engine removes noise and improves clarity.

[0411] Step 2:

[0412] Analysis of the situation

[0413] The server inputs the preprocessed data into a generative AI model to assess the location and spread of the fire, smoke density, and fire scale.

[0414] Input: Preprocessed image and video data

[0415] Output: Information on fire location, spread, smoke density, and fire scale

[0416] What it does: Apply deep learning models to detect fires in specific areas and visualize their spread with 3D mapping.

[0417] Step 3:

[0418] Obtaining location information

[0419] The device (smartphone or tablet) obtains the user's current location information using the GPS function and sends that information to the server.

[0420] Input: GPS data built into the device

[0421] Output: Current location of the user (latitude and longitude)

[0422] Specific operation: The device acquires GPS data and calls an API to send it to a server via the Internet.

[0423] Step 4:

[0424] Generate evacuation routes

[0425] The server uses an evacuation route generation algorithm to generate the safest evacuation route based on the results of the situation analysis and the user's location information, and sends the generated route information to the terminal.

[0426] Input: Fire location, spread, smoke density, user's current location

[0427] Output: Optimal evacuation route information

[0428] Specific operation: Calculates an evacuation route using Dijkstra's algorithm or A algorithm, and sends the generated route to the terminal as map data.

[0429] Step 5:

[0430] Emotional state assessment

[0431] The device uses a camera and microphone to analyze the user's facial expressions and voice to collect emotional data, which is then sent to a server where an emotion engine is used to evaluate the user's emotional state.

[0432] Input: User's facial expression data and voice data

[0433] Output: User's emotional state (stress level, impatience, etc.)

[0434] Specific operation: The device analyzes the user's facial expressions, collects voice data, and sends it to the server for evaluation by the emotion engine.

[0435] Step 6:

[0436] evacuation order

[0437] The device provides the user with visual and audio evacuation instructions based on the generated evacuation route information and emotional state assessment, and the user begins evacuation in accordance with the instructions.

[0438] Input: Optimal evacuation route information, user's emotional state evaluation

[0439] Output: Visual and audio evacuation instructions

[0440] Specific operation: The device displays a map and a voice guide, and provides instructions such as "Turn right, walk 30 meters, and go down the stairs." The user follows these instructions to evacuate.

[0441] Step 7:

[0442] Generate virtual training scenarios

[0443] The server generates a scenario for the virtual training and sends it to the terminal. The terminal notifies the user of the start of the virtual training and prompts the user to participate in the virtual training.

[0444] Input: Historical scenario data and training data

[0445] Output: Virtual training scenario

[0446] Specific operation: The server generates a virtual scenario in which a fire breaks out on the second floor of a building, and the device notifies the user that training has begun. The user puts on VR goggles and begins training.

[0447] Step 8:

[0448] Data analysis and preventative measures

[0449] The server collects past disaster data and training data, analyzes them using a generative AI model, and evaluates the fire risk in a specific area and suggests preventive measures.

[0450] Input: Past disaster data, training data

[0451] Output: Risk assessment results and preventative measures recommendations

[0452] Specific operation: The server analyzes disaster data, identifies high-risk areas, and presents improvement proposals for fire prevention measures to area managers.

[0453] Step 9:

[0454] Continuous learning and improvement

[0455] The server collects system usage data and new disaster patterns, and continuously updates the generative AI model, incorporating new algorithms and methods for optimizing evacuation routes into the system.

[0456] Input: System usage data, new disaster pattern data

[0457] Output: Updated generative AI model

[0458] Specific operation: The server takes in new data, retrains the AI ​​model, and integrates it into the existing system to update it.

[0459] (Application example 2)

[0460] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0461] Evacuation assistance systems in emergency situations are required to quickly and accurately assist users in evacuating safely. However, conventional evacuation assistance systems have difficulty providing flexible evacuation instructions that take into account the user's emotional state, and they have been particularly inadequate in addressing emotions such as stress and anxiety. Furthermore, there have been no evacuation assistance systems suitable for specific environments, such as factories. Therefore, a system that can appropriately assess the user's emotional state and provide dynamic evacuation instructions based on that assessment is needed to achieve safer and more effective evacuation assistance.

[0462] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0463] In this invention, the server includes means for acquiring and preprocessing images and video feeds of the scene in real time, means for analyzing the acquired data using a generative AI model to evaluate the progress of the disaster, means for acquiring user location information and generating the safest evacuation route based on the analysis results, means for analyzing the user's facial and voice data to evaluate their emotional state, and means for dynamically adjusting the content and method of evacuation instructions based on the generated evacuation route information and the evaluation of their emotional state, thereby enabling flexible and appropriate evacuation support in emergency situations while taking the user's emotional state into consideration.

[0464] "Image and video feed" refers to image and video data used to obtain real-time visual information on the scene during an emergency.

[0465] "Pre-processing" refers to the processing of acquired images and video feeds to improve data quality, such as removing noise or increasing clarity.

[0466] A "generative AI model" is a model that uses machine learning algorithms to predict and evaluate the progression of a disaster and the user's emotional state.

[0467] "Analysis" is the process of analyzing information from acquired data using a generative AI model to assess the progression of the disaster and the user's emotional state.

[0468] "Location information" is data that indicates the user's current geographic coordinates (latitude and longitude).

[0469] An "evacuation route" is the optimal route for a user to safely evacuate in the event of a disaster.

[0470] "Facial expression and vocal data" refers to visual and audio data for assessing a user's emotional state, such as facial expressions and tone of voice.

[0471] "Emotional state assessment" is the process of analyzing the user's facial and vocal data to determine their current emotional state (e.g., stress, impatience, anxiety, etc.).

[0472] "Dynamic adjustment" refers to changing the content and method of evacuation instructions in real time based on the user's emotional state assessment.

[0473] A "virtual training scenario" is a scenario for simulating evacuation behavior in a virtual environment, and is used by users to train.

[0474] An "interactive evacuation drill" is a training exercise in which users participate in a virtual training scenario and experience real-life evacuation procedures.

[0475] "Disaster data and training data" refers to data on disasters that have occurred in the past and data generated in virtual training.

[0476] "Risk assessment" is the process of assessing the risk of disasters occurring in a particular area.

[0477] "Preventive measures" are measures and actions taken to prevent disasters from occurring.

[0478] "Continuously updating" refers to continually improving generative AI models and systems based on new disaster patterns and usage data.

[0479] New patent specification

[0480] This invention relates to a system for supporting safe evacuation in emergency situations. The system performs real-time situation assessment, data analysis, dynamic evacuation support taking into account the user's emotional state, virtual training, evacuation route generation and update, risk assessment based on past data, and proposal of preventive measures.

[0481] Hardware used

[0482] Smartphone: Used to obtain the user's location and emotional state.

[0483] In-factory sensors and cameras: Used to capture real-time images and video feeds of the scene in case of an emergency.

[0484] Server: Performs data analysis, generates evacuation routes, and manages the system.

[0485] Software used

[0486] OpenCV: A library for image processing and data preprocessing.

[0487] TensorFlow: Used to run generative AI models and the emotion engine.

[0488] Firebase: Used as a real-time database to manage location information and analysis results.

[0489] System Components and Functions

[0490] server

[0491] 1. Acquiring and Preprocessing Image and Video Feeds

[0492] The server collects data in real time from cameras and sensors within the factory and uses OpenCV for pre-processing such as noise removal and image clarity improvement.

[0493] 2. Data analysis using generative AI models

[0494] The preprocessed data is analyzed using TensorFlow to assess the progress of the disaster and the user's emotional state.

[0495] 3. Obtaining user location information and generating evacuation routes

[0496] Location information obtained from smartphones is stored in Firebase, and the safest evacuation route is generated based on the analysis results.

[0497] 4. Emotional state assessment and dynamic evacuation instructions

[0498] The system analyzes the user's facial expressions and voice to assess their emotional state, and dynamically adjusts the content and method of evacuation instructions based on the results.

[0499] 5. Creating virtual training scenarios and conducting evacuation drills

[0500] Generate virtual training scenarios and conduct interactive evacuation drills for users.

[0501] 6. Analysis of historical data and proposal of preventive measures

[0502] Past disaster data and training data are analyzed to assess risks and propose preventive measures.

[0503] 7. Continuously updating generative AI models

[0504] The generative AI model is continuously updated based on new disaster patterns and usage data.

[0505] Specific examples

[0506] The server receives images from cameras inside the factory and detects that a fire has broken out in a specific area. The server analyzes the spread of smoke and identifies dangerous areas. At the same time, the smartphone obtains the user's current location information and sends it to the server. The server calculates a route from the user's current location to the nearest emergency exit, taking into account the spread of the fire and the density of the smoke. To evaluate the user's emotional state, the smartphone analyzes facial expressions through the camera and detects signs of stress or impatience. The server then provides more detailed and calming audio guidance to relieve the user's anxiety and stress.

[0507] Prompt Sentence Examples

[0508] Example of an input prompt for a generative AI model: "Identify the location of the fire and the spread of smoke from camera footage inside the factory, and generate an evacuation route."

[0509] Example prompt for the emotion engine: "Please assess employee stress levels based on data from their smartphone cameras and microphones and adjust evacuation instructions accordingly."

[0510] Through the above processing, flexible and appropriate evacuation assistance in emergency situations is possible while taking into consideration the emotional state of the user.

[0511] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0512] Step 1:

[0513] The server acquires images and video feeds in real time from cameras and sensors in the factory. The input of data acquisition is the video data from the cameras and sensors, and the output is the acquired raw data.

[0514] Step 2:

[0515] The server preprocesses the acquired video data using OpenCV. The input for the preprocessing is raw data, and data processing such as noise removal and image clarity improvement is performed, and the output is preprocessed video data.

[0516] Step 3:

[0517] The server analyzes the preprocessed video data using TensorFlow to assess the progression of the disaster. The input for the analysis is the preprocessed video data, and the generative AI model performs data calculations. The output is the disaster detection results and an assessment of the progression of the disaster.

[0518] Step 4:

[0519] The device acquires the user's location information using the GPS function. The input for acquiring location information is the user's current location, and the output is the user's latitude and longitude information.

[0520] Step 5:

[0521] The device sends the acquired location information to Firebase and shares it with the server in real time. The input of the transmission is the user's location information, and the output is the location data stored in Firebase.

[0522] Step 6:

[0523] The server generates the safest evacuation route based on the analysis results and the user's location information stored in Firebase. The inputs for evacuation route generation are the disaster progression assessment and the user's location information, and the output is the optimal evacuation route information.

[0524] Step 7:

[0525] The device acquires facial and voice data of the user using the smartphone's camera and microphone. The input for emotion data acquisition is the user's face and voice, and the output is the acquired facial and voice data.

[0526] Step 8:

[0527] The device analyzes the acquired facial and voice data using TensorFlow to evaluate the user's emotional state. The input for the analysis is facial and voice data, and the data is calculated using a generative AI model, and the output is the evaluated emotional state.

[0528] Step 9:

[0529] The server dynamically adjusts the content and method of evacuation instructions based on the user's emotional state evaluation result, where the input of the dynamic adjustment is the emotional state evaluation and the output is the adjusted evacuation instructions.

[0530] Step 10:

[0531] The terminal provides the user with coordinated evacuation instructions visually and audibly, where the input for providing evacuation instructions is the coordinated evacuation instructions and the output is the visual and audio evacuation instruction information.

[0532] Step 11:

[0533] The server generates a virtual training scenario and sends it to the terminal. The input of the virtual training scenario generation is the past data and the current scenario requirements, and the output is the generated virtual training scenario.

[0534] Step 12:

[0535] The terminal conducts an interactive evacuation drill for the user based on the virtual training scenario, where the input for the training is the virtual training scenario and the output is the training result.

[0536] Step 13:

[0537] The server analyzes past disaster data and training data to assess the risk of a specific area and propose preventive measures. The inputs for risk assessment are past data and training data, and data analysis is performed. The output is the risk assessment results and preventive measures.

[0538] Step 14:

[0539] The server continuously updates the generative AI model based on system usage data and new disaster patterns. The input for the update is usage data and new disaster patterns, and the server performs data calculations, outputting the latest generative AI model.

[0540] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0541] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0542] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0543] [Second embodiment]

[0544] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0545] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0546] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0547] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0548] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0549] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0550] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0551] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0552] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0553] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0554] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0555] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0556] The AI ​​emergency evacuation support system of the present invention is designed to support safe evacuation in emergency situations. The system functions through cooperation between a server, a terminal, and a user. The program's processing is explained below with specific examples.

[0557] Basic components

[0558] Server: Responsible for real-time acquisition and pre-processing of images and video feeds, data analysis using generative AI models, generation of evacuation routes, generation of virtual training scenarios, data analysis and prevention measures proposals, and usage data updates.

[0559] Terminal: Responsible for obtaining user location information, providing real-time evacuation instructions, and conducting virtual drills.

[0560] User: Follow instructions from the device to take evacuation action and participate in virtual training.

[0561] System action

[0562] Real-time situation analysis

[0563] 1. The server receives real-time images and video feeds from cameras and sensors within the facility.

[0564] 2. The server preprocesses the acquired data and removes unnecessary noise.

[0565] 3. The server uses a generative AI model to analyze the pre-processed data and assess the spread of the fire, smoke density, and fire scale.

[0566] Examples:

[0567] The server analyzes footage from surveillance cameras in real time and detects when a fire has broken out in a specific area.

[0568] The server analyzes the spread of smoke and identifies dangerous areas.

[0569] Obtaining user location information

[0570] 1. The device (mainly a smartphone or tablet) obtains the user's current location information using the GPS function.

[0571] 2. The device sends the acquired location information to the server.

[0572] Examples:

[0573] The user launches the app and allows location sharing.

[0574] The device obtains the user's current latitude and longitude and sends them to the server.

[0575] Generate evacuation routes

[0576] 1. The server generates the optimal evacuation route based on the results of real-time situation analysis and the user's location information.

[0577] 2. The server sends the generated evacuation route information to the terminal.

[0578] Examples:

[0579] The server calculates the route from the user's current location to the nearest emergency exit, taking into account the spread of the fire and the density of the smoke.

[0580] The server transmits map data showing evacuation routes to the terminal.

[0581] Evacuation instructions to users

[0582] 1. The device provides the user with visual and audio evacuation instructions based on the received evacuation route information.

[0583] 2. The user follows the instructions on the device to evacuate safely.

[0584] Examples:

[0585] The device will then provide voice instructions such as, "Turn right, walk 30 meters, and go down the stairs."

[0586] The user follows the instructions on the device and takes evacuation action.

[0587] Virtual Training and Simulation

[0588] 1. The server generates a scenario for virtual training and sends it to the terminal.

[0589] 2. The terminal notifies the user that the virtual training has started and displays the training scenario.

[0590] 3. Users participate in virtual drills and take evacuation actions according to simulated scenarios.

[0591] Examples:

[0592] The server generates a hypothetical scenario in which a fire breaks out on the second floor of a building.

[0593] The device instructs the user to put on the VR goggles and begin training.

[0594] Users practice evacuation plans in a virtual environment.

[0595] Data analysis and preventative measures

[0596] 1. The server collects past disaster data and training data and analyzes them using a generative AI model.

[0597] 2. The server assesses the fire risk in a particular area and suggests preventative measures.

[0598] Examples:

[0599] The server analyzes past fire data to identify fire frequency and high-risk areas.

[0600] The server will suggest preventative measures to the area administrator.

[0601] Continuous learning and improvement

[0602] 1. The server collects system usage data and new disaster patterns and continuously updates the generative AI model.

[0603] 2. The server learns new algorithms and methods for optimizing evacuation routes and reflects them in the system.

[0604] Examples:

[0605] The server analyzes data from annual evacuation drills and learns that certain evacuation routes are prone to congestion.

[0606] The server will propose a new evacuation route and update the system accordingly.

[0607] In this way, the AI ​​emergency evacuation support system of the present invention supports users in safely evacuating through real-time situation assessment, analysis, instructions, training, and continuous learning.

[0608] The processing flow will be explained below.

[0609] Step 1:

[0610] The server collects real-time data from surveillance cameras and sensors installed within the facility, including still images, video feeds, and temperature sensor data.

[0611] Step 2:

[0612] The server pre-processes the collected image and video feed data to remove noise and improve image clarity, which improves the accuracy of the analysis.

[0613] Step 3:

[0614] The server uses a generative AI model to analyze the pre-processed data and evaluate the location of the fire, its spread, smoke density, and fire scale, thereby determining the level of urgency at the scene.

[0615] Step 4:

[0616] The device obtains the user's current location information using the GPS function. If the user carries a smartphone or tablet, the location information can be updated continuously through the app.

[0617] Step 5:

[0618] The device sends the acquired user location information to the server, which allows the server to determine the user's current location.

[0619] Step 6:

[0620] The server then calculates the safest evacuation route based on the analysis results and the user's location information, taking into account factors such as the spread of the fire, smoke density, and congestion on evacuation routes.

[0621] Step 7:

[0622] The server then sends the generated evacuation route information to the user's device. The sent data includes the specific evacuation route and its details (e.g., how many meters to go, which direction to turn, etc.).

[0623] Step 8:

[0624] Based on the received evacuation route information, the device provides the user with visual and audio evacuation instructions, including map display and audio guidance.

[0625] Step 9:

[0626] The user follows the instructions on the device to begin evacuation along a safe evacuation route. By following the instructions, the user can evacuate safely.

[0627] Step 10:

[0628] The server generates virtual training scenarios during normal times and transmits them to the terminals. Specific training scenarios include fire outbreak scenarios and evacuation scenarios.

[0629] Step 11:

[0630] The device notifies the user of the start of virtual training and prompts them to train in the virtual environment. Users can participate in the virtual training using VR goggles or a smartphone.

[0631] Step 12:

[0632] Users participate in virtual training and take evacuation actions according to simulated fire scenarios. After the training, they receive feedback from the system to identify areas for improvement in their evacuation actions.

[0633] Step 13:

[0634] The server collects past disaster data and training data and analyzes it using a generative AI model, which then generates a risk assessment for a specific area.

[0635] Step 14:

[0636] Based on the risk assessment results, the server will suggest preventative measures to managers of specific areas, such as installing sprinkler systems and strengthening evacuation drills.

[0637] Step 15:

[0638] The server collects system usage data and new disaster patterns to continuously update the generative AI model, providing evacuation routes and preventative measures that reflect the latest knowledge.

[0639] Example 1

[0640] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0641] While quick and safe evacuation is extremely important in an emergency, current systems have difficulty obtaining disaster information in real time and providing optimal evacuation routes. Furthermore, inaccurate location information for users and lack of evacuation drills could lead to confusion during an actual disaster. Furthermore, past disaster data and drill data are not fully utilized, preventing preventative measures from being proposed and the system from being continuously improved.

[0642] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0643] In this invention, the server includes means for acquiring and preprocessing images and video feeds of the scene in real time, means for analyzing the acquired data using a generative AI model to evaluate the progress of the disaster, means for acquiring user location information and generating the safest evacuation route based on the analysis results, means for transmitting the generated evacuation route information to the user's device and providing audio and visual evacuation instructions, means for generating virtual training scenarios and conducting interactive evacuation drills with the user, means for analyzing past disaster data and training data to assess risk and propose preventive measures for specific areas, means for continuously updating the generative AI model based on system usage data and new disaster patterns, means for acquiring location information using the GPS function of the user's device, means for analyzing the spread of fire and smoke concentration using the generative AI model, and means for notifying the user of the start of the virtual training and displaying the training scenario using VR goggles, thereby enabling real-time situation assessment and evacuation instructions to be given, thereby achieving safe evacuation.

[0644] "Images and video feeds" refers to visual information obtained in real time from surveillance cameras and sensors.

[0645] "Preprocessing" refers to the preparation of data before analysis, such as removing noise from the acquired data and correcting outliers.

[0646] A "generative AI model" is an artificial intelligence model that uses machine learning and deep learning algorithms to analyze data.

[0647] "Disaster progression" refers to changes in the situation and scope of impact in emergencies such as fires, earthquakes, and floods.

[0648] "User location information" refers to the user's current geographic coordinate information obtained using the GPS function.

[0649] An "evacuation route" is the optimal route to safety from an emergency situation.

[0650] An "evacuation instruction" is an audio and visual guide that encourages the user to take evacuation action.

[0651] A "virtual training scenario" is a scenario that allows a user to conduct an evacuation drill using virtual reality or simulation.

[0652] "Past disaster data" refers to records and information about disasters that have occurred in the past.

[0653] "Risk assessment" is the evaluation of the likelihood and impact of a disaster in a particular area or situation.

[0654] "Preventive measures" are measures proposed to prevent disasters from occurring.

[0655] "System Usage Data" means all data collected through the System.

[0656] "New disaster patterns" refer to new trends and characteristics of disasters that have never been seen before.

[0657] "GPS function" refers to a function that obtains location information using the Global Positioning System.

[0658] "Smoke density" is an indicator of the density or amount of smoke present in the air.

[0659] "VR goggles" are devices that allow users to experience a virtual reality environment.

[0660] This invention relates to an AI emergency evacuation support system that supports safe evacuation in emergency situations. The system of the invention functions through cooperation between a server, a terminal, and a user.

[0661] First, the server acquires real-time images and video feeds from cameras and sensors within the facility. Examples of hardware used include surveillance cameras and smoke detection sensors. The data obtained from these devices undergoes pre-processing, including noise removal and outlier correction. Specifically, video data is filtered and sensor data is corrected to ensure data quality.

[0662] The pre-processed data is then analyzed using generative AI models to assess fire spread, smoke density, and fire scale. For example, image analysis can detect the location and spread of a fire in real time.

[0663] The devices are primarily smartphones and tablets, which use GPS to obtain the user's current location information. The location information is then sent from the device to a server, which then uses it to calculate evacuation routes.

[0664] The server generates an optimal evacuation route based on the results of real-time situation analysis and the user's location information. The generated evacuation route information is sent to the device. Based on this information, the device provides the user with visual and audio evacuation instructions. For example, specific instructions such as "Turn right, walk 30 meters, and go down the stairs" are provided by audio guidance.

[0665] The server also generates scenarios for virtual training and sends them to the device. At this time, the user receives a notification from the device and begins the virtual training. By wearing VR goggles, evacuation training can be carried out in the virtual environment. As a specific example, a virtual scenario is generated in which "a fire breaks out on the second floor of a building," and the user conducts virtual training based on this scenario.

[0666] The server collects past disaster data and training data and analyzes it using a generative AI model. This allows it to assess the fire risk in a specific area and propose preventative measures. For example, it can analyze past fire data to identify patterns of fire occurrence in specific time periods or areas and propose measures to strengthen fire prevention equipment.

[0667] Furthermore, the server continuously updates the generative AI model based on system usage data and new disaster patterns. By learning new algorithms and evacuation route optimization methods and incorporating them into the system, the system can learn that certain evacuation routes are prone to congestion based on annual training data and suggest new evacuation routes.

[0668] As a result, the AI ​​emergency evacuation support system of the present invention supports users in safely evacuating through real-time situation assessment, analysis, instructions, training, and continuous learning.

[0669] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0670] Step 1:

[0671] The server receives real-time images and video feeds from cameras and sensors within the facility. This input data includes surveillance camera footage and smoke detection sensor data. Specifically, it collects camera stream data and converts it into a different format. The output is raw image and video data that has not yet been preprocessed.

[0672] Step 2:

[0673] The server preprocesses the acquired data and removes unwanted noise. The input data is the image and video data obtained in step 1, and a filtering algorithm is applied to remove noise. Specifically, a low-pass filter is applied to the video data, and outlier correction is performed on the sensor data. The output is high-quality data with noise removed.

[0674] Step 3:

[0675] The server uses a generative AI model to analyze the preprocessed data and evaluate the spread of the fire, smoke density, and fire scale. The input data is the output of step 2, and is fed into the generative AI model for analysis. Specifically, the deep learning model is used to identify the extent of the fire's expansion and calculate smoke density. The output is the progress of the fire and the risk assessment results.

[0676] Step 4:

[0677] The device (smartphone or tablet) obtains the user's current location information using its GPS function. Input data is obtained from the user's device and includes latitude and longitude information. Specifically, it collects real-time data from the GPS sensor to obtain location information. The output is the user's precise location information.

[0678] Step 5:

[0679] The device sends the acquired location information to the server. The input data is the location information from step 4, which is encrypted using a security protocol and sent to the server. The specific operation is to establish a secure communication channel and send the location information. The output is the location information received by the server.

[0680] Step 6:

[0681] The server generates an optimal evacuation route based on the results of real-time situation analysis and the user's location information. The input data are the risk assessment results from step 3 and the location information from step 5. Specifically, it applies a route calculation algorithm to generate a safe evacuation route. The output is evacuation route information customized for each user.

[0682] Step 7:

[0683] The server sends the generated evacuation route information to the terminal. The input data is the evacuation route information from step 6, which is sent to the user's terminal. Specifically, the server sends the evacuation route map in digital format. The output is the evacuation route information displayed on the terminal.

[0684] Step 8:

[0685] The device provides the user with visual and audio evacuation instructions based on the received evacuation route information. The input data is the evacuation route information from step 7, and this is notified to the user through a visual map display and audio guidance. Specifically, the device provides audio guidance of direction and distance information through the application. The output is the evacuation instructions received by the user.

[0686] Step 9:

[0687] The user follows the instructions on the terminal to evacuate safely. The input data is the evacuation instructions from step 8, and upon receiving this, the process moves on to actual evacuation behavior. Specifically, the user follows the instructions on the terminal and evacuates by following a safe route. The output is to reach a safe area.

[0688] (Application example 1)

[0689] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0690] Conventional evacuation support systems have difficulty responding to emergencies in real time, and have been unable to provide effective evacuation support, especially in large facilities or complex environments. Furthermore, it has been difficult to provide appropriate evacuation instructions to users with visual or hearing impairments. Furthermore, there has been a problem in that preventive measures and continuous improvements have not been proposed using past training data or disaster data.

[0691] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0692] In this invention, the server includes: means for acquiring and preprocessing images and video feeds of the scene in real time; means for analyzing the acquired data using a generative AI model to assess the progress of the disaster; means for acquiring user location information and generating the safest evacuation route based on the analysis results; means for transmitting the generated evacuation route information to the user's device and providing audio and visual evacuation instructions; means for generating virtual training scenarios and conducting interactive evacuation drills with the user; means for analyzing past disaster data and training data to assess risk and propose preventative measures for specific areas; means for continuously updating the generative AI model based on system usage data and new disaster patterns; and means installed in smart glasses for generating an optimal evacuation route based on the user's current location and providing visual and audio evacuation instructions. This enables effective evacuation support even in large facilities and complex environments, and provides appropriate evacuation instructions to visually and hearing-impaired individuals. It also enables the system to propose preventative measures and continuously improve the system by utilizing past training data and disaster data.

[0693] An "emergency" is an event that poses a threat to people's safety and lives, such as an unexpected disaster or accident.

[0694] An "evacuation support system" is a system that includes a series of technologies and devices to assist people in safely evacuating in the event of an emergency.

[0695] "On-site images and video feed" refers to real-time visual information of the location and surroundings of an emergency.

[0696] "Preprocessing" refers to the process of removing unnecessary noise from acquired image or video data and converting it into a format that is easier to analyze.

[0697] A "generative AI model" is an artificial intelligence model that learns from large amounts of data and has the ability to analyze situations and generate appropriate evacuation routes.

[0698] "Disaster Progression" refers to the process of damage expansion and current situation in emergencies such as fires and earthquakes.

[0699] "User Location Information" means an individual user's current geographic location obtained using GPS or other location-determining technology.

[0700] An "evacuation route" refers to a route that guides a user to the safest evacuation route in the event of an emergency.

[0701] "Terminal" refers to a mobile information terminal used by a user, such as a smartphone, tablet, or smart glasses.

[0702] A "virtual training scenario" refers to a scenario that uses virtual reality and simulation to allow people to realistically experience an emergency evacuation drill.

[0703] "Past disaster data" refers to records and data relating to disasters that have occurred in the past.

[0704] "Training data" refers to data collected during evacuation drills and is used to improve the system and assess risks.

[0705] "Risk assessment" is the act of assessing the possibility of a disaster occurring in a specific area and the extent of its impact.

[0706] "Preventive measures" refer to strategies and measures to reduce risk and ensure safety before a disaster occurs.

[0707] "Continuous updates" refers to the process by which systems and generative AI models learn new data and techniques, keeping them constantly up to date.

[0708] The present invention provides a system for supporting safe evacuation in an emergency situation, which functions in cooperation with a server, terminals, and users.

[0709] Basic components

[0710] 1. Server: Acquires and preprocesses images and video feeds from the scene in real time. It also analyzes the data using a generative AI model to assess the progress of the disaster. It generates the safest evacuation route based on the user's location information and analysis results and sends this to the device. It also generates virtual training scenarios and analyzes past disaster data and training data to assess risks and propose preventive measures. It is responsible for continuously updating the generative AI model based on system usage data and new disaster patterns.

[0711] 2. Terminal: This is a user device, typically a smart eyeglass or smartphone, that acquires the user's current location. It provides visual and audio evacuation instructions based on evacuation route information received from the server. It also provides training scenarios and conducts interactive training with the user during virtual training.

[0712] 3. User: Follows instructions from the terminal to take evacuation action. Participates in virtual training and acts according to the evacuation plan provided by the system.

[0713] System action

[0714] Hardware and Software

[0715] On-site cameras and sensors: Capture images and video feeds in real time.

[0716] GPS module: Obtains the user's location information.

[0717] Smart glasses: Provide visual and audio evacuation instructions.

[0718] Server: Responsible for data processing and running generative AI models.

[0719] Data processing and calculation

[0720] 1. The server acquires images and video feeds from on-site cameras and sensors in real time and performs pre-processing such as noise removal. The pre-processed data is then analyzed using a generative AI model to assess the progression of the disaster.

[0721] 2. The device acquires the user's location information using GPS and sends it to the server, which then combines this information with the analysis results to generate the optimal evacuation route.

[0722] 3. The server sends the generated evacuation route to the terminal, and the terminal provides the user with visual and audio evacuation instructions.

[0723] 4. The virtual training scenario is generated by the server and sent to the user's device. The user wears the smart glasses and conducts an interactive evacuation drill.

[0724] Examples of concrete examples and prompts

[0725] For example, if a fire breaks out in a factory, the server will analyze the footage from the factory's surveillance cameras in real time to identify the location and spread of the fire. It will then calculate the safest evacuation route based on the user's current location and provide visual and audio evacuation instructions through the smart glasses. The following prompt sentences are used:

[0726] The server analyzes the footage from the factory's surveillance cameras in real time and detects a fire in Section 4. Based on the current location of the workers (Section 3) obtained via GPS and the location of the fire, generate the optimal evacuation route. Then, send this evacuation route to the smart glasses and initiate visual and audio instructions.

[0727] In this way, the AI ​​emergency evacuation support system of the present invention can support safe and rapid evacuation in emergency situations through real-time situation analysis, generation of evacuation plans, visual and audio instructions to users, and even virtual training.

[0728] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0729] Step 1:

[0730] Acquiring and pre-processing images and video feeds from the scene

[0731] The server receives real-time image and video feeds from cameras and sensors on-site. The input is raw image and video data from the site. This data is preprocessed to remove unwanted noise, for example by smoothing the image or enhancing edges, to convert it into a format that is easier to analyze. The output is the preprocessed image and video data.

[0732] Step 2:

[0733] Data analysis using generative AI models

[0734] The server inputs preprocessed image and video data into a generative AI model for data analysis. Specifically, it evaluates the spread of the fire, smoke density, and the scale of the disaster. The input is preprocessed image and video data, and the output is various analytical results on the progression of the disaster. The generative AI model accomplishes this by generating data to determine the necessary rescue operations and evacuation routes.

[0735] Step 3:

[0736] Obtaining user location information

[0737] The device obtains the user's current location information using its GPS function. The input is location data from the GPS module, which acquires the user's latitude and longitude information. The device generates a data packet to send this location information to the server and sends it to the server. The output is the user's exact current location, which the server receives.

[0738] Step 4:

[0739] Generate evacuation routes

[0740] The server generates the optimal evacuation route based on the acquired user location information and the disaster analysis results generated by the AI ​​model. For example, it determines the safest route by taking into account the spread of fire and smoke density. The inputs are the user's current location and the disaster analysis results, and the output is the optimal route information for the user to evacuate.

[0741] Step 5:

[0742] Sending and instructing evacuation routes

[0743] The server sends the generated evacuation route information to the terminal. The terminal receives this and provides the user with visual and audio evacuation instructions. Specifically, the terminal issues specific instructions via voice, such as "Turn right, walk 30 meters, and go down the stairs." The input is the evacuation route information sent from the server, and the output is the specific evacuation instructions received by the user.

[0744] Step 6:

[0745] Generate and conduct virtual training scenarios

[0746] The server generates a scenario for the virtual training and sends it to the terminal. The terminal notifies the user of the start of the virtual training and displays the training scenario. The user takes evacuation actions in the virtual environment according to the scenario. The input is the generated virtual training scenario, and the output is the results of the virtual training performed by the user.

[0747] Step 7:

[0748] Data analysis and preventative measures

[0749] The server collects past disaster data and training data and analyzes it using a generative AI model. As a result of the analysis, it performs a risk assessment of a specific area and proposes appropriate preventive measures. For example, it may recommend installing additional fire detectors in areas where fires frequently occur. The inputs are past disaster data and training data, and the output is the risk assessment results and proposed preventive measures.

[0750] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0751] This invention provides flexible evacuation support that takes into account the user's emotional state by combining an emotion engine with a system for supporting safe evacuation in emergency situations. The following describes the system components and program processing with specific examples.

[0752] Basic components

[0753] Server: Responsible for real-time acquisition and pre-processing of images and video feeds, data analysis using generative AI models, evaluation of users' emotional state using an emotion engine, generation of evacuation routes, generation of virtual training scenarios, data analysis and prevention measures proposals, and usage data updates.

[0754] Terminal: Responsible for obtaining user location information, collecting emotion engine data, providing real-time evacuation instructions, and conducting virtual drills.

[0755] User: Follows instructions from the device to take evacuation action, provides emotional data, and participates in virtual training.

[0756] System action

[0757] Real-time situation analysis

[0758] 1. The server receives real-time images and video feeds from cameras and sensors within the facility.

[0759] 2. The server preprocesses the acquired data to remove noise and improve image clarity.

[0760] 3. The server analyzes the pre-processed data using a generative AI model to assess the location and spread of the fire, smoke density, and fire scale.

[0761] Examples:

[0762] The server analyzes footage from surveillance cameras and detects that a fire has broken out in a specific area.

[0763] The server analyzes the spread of smoke and identifies dangerous areas.

[0764] Obtaining user location information

[0765] 1. The device (smartphone or tablet) obtains the user's current location information using the GPS function.

[0766] 2. The device sends the acquired location information to the server.

[0767] Examples:

[0768] The user launches the app and allows location sharing.

[0769] The device obtains the user's current latitude and longitude and sends them to the server.

[0770] Generate evacuation routes

[0771] 1. The server generates the safest evacuation route based on the results of real-time situation analysis and the user's location information.

[0772] 2. The server sends the generated evacuation route information to the terminal.

[0773] Examples:

[0774] The server calculates the route from the user's current location to the nearest emergency exit, taking into account the spread of the fire and the density of the smoke.

[0775] The server transmits map data showing evacuation routes to the terminal.

[0776] Evaluating the user's emotional state

[0777] 1. The device uses a camera and microphone to analyze the user's facial expressions and voice and collect emotional data.

[0778] 2. The device sends the collected emotion data to the server.

[0779] 3. The server uses an emotion engine to evaluate the user's emotional state and dynamically adjusts the content and method of evacuation instructions based on the results.

[0780] Examples:

[0781] The device analyzes the user's facial expressions through the camera and detects signs of stress or impatience.

[0782] The server provides more detailed and calming audio guidance to ease the user's anxiety and stress.

[0783] Evacuation instructions to users

[0784] 1. The device provides visual and audio evacuation instructions to the user based on the received evacuation route information and emotional state assessment.

[0785] 2. The user begins evacuating along a safe evacuation route following the instructions on the device.

[0786] Examples:

[0787] The device will then provide voice instructions such as, "Turn right, walk 30 meters, and go down the stairs."

[0788] The user follows the instructions on the device and evacuates along the designated evacuation route.

[0789] Virtual Training and Simulation

[0790] 1. The server generates a scenario for virtual training and sends it to the terminal.

[0791] 2. The device notifies the user that virtual training has begun and prompts them to train in the virtual environment.

[0792] 3. Users participate in virtual drills and take evacuation actions according to simulated scenarios.

[0793] Examples:

[0794] The server generates a hypothetical scenario in which a fire breaks out on the second floor of a building.

[0795] The device instructs the user to put on the VR goggles and begin training.

[0796] Users practice evacuation plans in a virtual environment.

[0797] Data analysis and preventative measures

[0798] 1. The server collects past disaster data and training data and analyzes them using a generative AI model.

[0799] 2. The server assesses the fire risk in a particular area and suggests preventative measures.

[0800] Examples:

[0801] The server analyzes past fire data to identify fire frequency and high-risk areas.

[0802] The server will suggest preventative measures to the area administrator.

[0803] Continuous learning and improvement

[0804] 1. The server collects system usage data and new disaster patterns and continuously updates the generative AI model.

[0805] 2. The server learns new algorithms and methods for optimizing evacuation routes and reflects them in the system.

[0806] Examples:

[0807] The server analyzes data from annual evacuation drills and learns that certain evacuation routes are prone to congestion.

[0808] The server will propose a new evacuation route and update the system accordingly.

[0809] In this way, the AI ​​emergency evacuation support system of the present invention supports users' safe evacuation through real-time situational awareness, data analysis, evaluation of the user's emotional state using an emotion engine, dynamic evacuation instructions, training and simulation, data analysis and prevention measures proposals, and continuous learning and improvement. In addition, the use of an emotion engine enables flexible evacuation support that takes the user's psychological state into consideration.

[0810] The processing flow will be explained below.

[0811] Step 1:

[0812] The server collects real-time images and video feeds from security cameras and sensors installed within the facility, including security camera footage, temperature sensor data, and smoke detector sensor data.

[0813] Step 2:

[0814] The server pre-processes the collected image and video feed data, removing noise, and optionally changes the image resolution or compresses the data.

[0815] Step 3:

[0816] The server analyzes the pre-processed data using a generative AI model to assess the location of the fire, its spread, smoke density, and fire scale, thereby determining the level of urgency at the scene.

[0817] Step 4:

[0818] The device (smartphone or tablet) acquires the user's current location information using the GPS function. When the user launches the emergency app, collection of location information begins.

[0819] Step 5:

[0820] The device sends the acquired user location information to the server, which then calculates an evacuation route based on the user's accurate location information at the time of the disaster.

[0821] Step 6:

[0822] The server calculates the safest evacuation route based on the results of real-time situation analysis and the user's location information, taking into account factors such as the spread of the fire, smoke density, congestion on evacuation routes, and the overall layout of the building.

[0823] Step 7:

[0824] The server then sends the generated evacuation route information to the user's terminal. The evacuation route information includes specific routes and evacuation procedures.

[0825] Step 8:

[0826] Based on the received evacuation route information, the device provides the user with visual and audio evacuation instructions, including displaying a map and providing audio guidance.

[0827] Step 9:

[0828] The user follows the instructions on the device and begins evacuating along the designated evacuation route. By taking appropriate evacuation actions, danger can be minimized.

[0829] Step 10:

[0830] The device uses a camera and microphone to analyze the user's facial expressions and voice to collect emotional data, and an emotion engine evaluates the user's stress level and anxiety state.

[0831] Step 11:

[0832] The device sends the collected emotional data to a server, which allows the user's psychological state to be grasped in real time.

[0833] Step 12:

[0834] The server uses an emotion engine to assess the user's emotional state and dynamically adjusts the evacuation instructions and methods based on the results, providing more detailed and calming guidance when stress and anxiety levels are high.

[0835] Step 13:

[0836] The device receives and provides emotion-dependent instructions to the user, and visual or audio guidance may change depending on the emotional state.

[0837] Step 14:

[0838] By receiving instructions that take emotions into consideration, users can continue their evacuation activities with peace of mind, which is expected to reduce stress.

[0839] Step 15:

[0840] The server generates virtual training scenarios during normal times and transmits them to the terminals. Specific training scenarios include fire outbreak scenarios and evacuation scenarios.

[0841] Step 16:

[0842] The device notifies the user that virtual training has started and prompts them to train in the virtual environment. The user participates in the virtual training using VR goggles or a smartphone.

[0843] Step 17:

[0844] Users participate in virtual training and take evacuation actions according to simulated fire scenarios. After the training, they receive feedback from the system to identify areas for improvement in their evacuation actions.

[0845] Step 18:

[0846] The server collects past disaster data and training data and analyzes it using a generative AI model to assess the fire risk in a specific area.

[0847] Step 19:

[0848] Based on the risk assessment results, the server will suggest preventative measures to managers of specific areas, such as installing sprinkler systems and strengthening evacuation drills.

[0849] Step 20:

[0850] The server collects system usage data and new disaster patterns to continuously update the generative AI model, providing evacuation routes and preventative measures that reflect the latest knowledge.

[0851] Example 2

[0852] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0853] The purpose of this invention is to provide evacuation support in emergency situations by grasping the situation on-site in real time, providing safe and rapid evacuation routes, and providing flexible and effective evacuation instructions that take into account the user's emotional state. It also aims to provide a more advanced evacuation support system by practicing evacuation behavior through virtual training scenarios, proposing preventive measures, and continuously improving the system.

[0854] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0855] In this invention, the server includes means for acquiring and preprocessing images and video feeds of the scene in real time, means for analyzing the acquired data using a generative AI model to evaluate the progress of the disaster, means for acquiring user location information and generating the safest evacuation route based on the analysis results, means for acquiring the user's emotional state using a camera and microphone and evaluating it using an emotion engine, means for providing audio and visual evacuation instructions based on the generated evacuation route information and emotional state evaluation, means for generating virtual training scenarios and conducting interactive evacuation drills with the user, means for analyzing past disaster data and training data to assess risks and propose preventive measures for specific areas, and means for continuously updating the generative AI model based on system usage data and new disaster patterns, thereby enabling real-time situation assessment, provision of safe evacuation routes, emotionally sensitive evacuation instructions, interactive training, and proposal of preventive measures, as well as continuous system improvement.

[0856] An "emergency situation" refers to an emergency situation that requires a rapid response, such as a natural disaster or man-made trouble.

[0857] "Evacuation" refers to the act of escaping danger and moving to a safe place.

[0858] "System" refers to a collection of devices and software in which multiple elements work together to perform a specific function.

[0859] "Image and video feed" refers to still and video data acquired through cameras and sensors.

[0860] "Preprocessing" refers to initial processing such as improving the quality of acquired data and removing noise.

[0861] A "generative AI model" refers to an algorithmic model of artificial intelligence generated using machine learning or deep learning techniques.

[0862] "Analyzing data" refers to the process of extracting information from acquired data and drawing meaningful conclusions.

[0863] "Disaster progression" refers to the extent of damage and the scope of impact in emergencies such as fires and earthquakes.

[0864] "User location information" refers to data about a user's current location obtained by GPS or other positioning systems.

[0865] An "evacuation route" refers to a route that is considered safe for escaping from a disaster.

[0866] "Emotional state" refers to the user's psychological state, such as stress or impatience.

[0867] An "emotion engine" refers to an algorithm or model for analyzing and assessing a user's emotional state.

[0868] "Evacuation instructions" refers to specific instructions or guidance that encourage users to take evacuation action.

[0869] "Virtual training scenario" refers to a virtual training environment and storyline designed around a simulated emergency situation.

[0870] "Interactive evacuation drill" refers to a training process in which users can actively participate within a virtual environment.

[0871] "Historical disaster data" refers to records and information relating to previous emergencies.

[0872] "Training data" refers to information such as behavior and responses collected during training.

[0873] "Risk assessment" refers to the process of quantitatively assessing the likelihood of a disaster occurring in a particular area.

[0874] "Preventive measures" refer to measures and actions taken to minimize the occurrence and impact of disasters.

[0875] "Usage Data" refers to all information collected in the course of using the System.

[0876] "Disaster patterns" refer to patterns or models that show how a particular disaster will unfold based on historical data and statistics.

[0877] An "update" refers to updating an existing model or system by adding new data or algorithms.

[0878] The present invention is a system for supporting safe evacuation in an emergency situation, and is implemented with the following configuration and method.

[0879] System configuration

[0880] The basic components are as follows:

[0881] Server: Responsible for real-time acquisition and pre-processing of images and video feeds, data analysis using generative AI models, evaluation of users' emotional state using an emotion engine, generation of evacuation routes, generation of virtual training scenarios, data analysis and prevention measures proposals, and usage data updates.

[0882] Terminal: Responsible for obtaining user location information, collecting emotion engine data, providing real-time evacuation instructions, and conducting virtual drills.

[0883] User: Follows instructions from the device to take evacuation action, provides emotional data, and participates in virtual training.

[0884] System action

[0885] 1. Data Acquisition and Preprocessing

[0886] The server receives real-time images and video feeds from cameras and sensors installed within the facility, and uses noise filtering algorithms to remove noise from the data and image enhancement techniques to improve clarity.

[0887] Example: The video stream from a surveillance camera is received by the capturing module and pre-processed by the image processing engine.

[0888] 2. Situation Analysis

[0889] The pre-processed data is analyzed on a server, and a generative AI model is used to assess the location and spread of fires, smoke density, and fire scale, identifying areas with a particularly high probability of fire outbreak.

[0890] Example: A deep learning model detects fires in a specific area and visualizes their spread with 3D mapping.

[0891] 3. Obtain location information

[0892] The device (smartphone or tablet) obtains the user's current location information using the GPS function and sends that information to the server.

[0893] Example: A user launches an app and the device sends GPS data to a server.

[0894] 4. Evacuation route generation

[0895] The server generates the safest evacuation route based on the results of real-time situation analysis and the user's location information using Dijkstra's algorithm and A algorithm. The generated route is sent to the device as map data.

[0896] Example: The server calculates an evacuation route and sends map data to the device.

[0897] 5. Emotional state assessment

[0898] The device uses a camera and microphone to analyze the user's facial expressions and voice to collect emotional data, which is then sent to a server, where an emotion engine is used to evaluate the user's emotional state and dynamically adjust evacuation instructions as needed.

[0899] Example: The device analyzes the user's facial expressions, detects signs of impatience, and the server changes the audio guidance to a calmer version.

[0900] 6. Evacuation instructions

[0901] Based on the received evacuation route information and emotional state assessment, the device provides the user with visual and audio evacuation instructions, which the user follows to begin evacuation along a safe evacuation route.

[0902] Example: The device will give voice instructions such as "Turn right, walk 30 meters, and go down the stairs," and the user will follow the instructions to evacuate.

[0903] 7. Virtual Training

[0904] The server generates a scenario for the virtual training and sends it to the terminal. The terminal notifies the user of the start of the virtual training and prompts the user to participate in the training. The user participates in the virtual training and takes evacuation actions according to the simulated scenario.

[0905] Example: The server generates a virtual scenario in which a fire breaks out on the second floor of a building, and the device instructs the user to put on VR goggles and begin training.

[0906] 8. Data analysis and preventive measures

[0907] The server collects past disaster data and training data, analyzes them using a generative AI model, and evaluates the fire risk in a specific area and suggests preventive measures.

[0908] Example: The server analyzes past fire data and presents improvement proposals for fire prevention measures to managers of high-risk areas.

[0909] 9. Continuous learning and improvement

[0910] The server collects system usage data and new disaster patterns, continually updating the generative AI model, learning new algorithms and methods for optimizing evacuation routes and incorporating them into the system.

[0911] Example: The server analyzes annual evacuation drill data, proposes new evacuation routes, and reflects them in the system.

[0912] Example prompts for generative AI models

[0913] Prompt: "Generate a virtual training scenario, analyze the user's stress level, and suggest the optimal evacuation route for the user. Please provide a detailed description of the specific training scenario and its steps, and finally include evacuation instructions for the user."

[0914] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0915] Step 1:

[0916] Data acquisition and preprocessing

[0917] The server captures real-time images and video feeds from cameras and sensors installed within the facility, then uses noise filtering algorithms to remove noise from the captured data and image enhancement techniques to improve clarity.

[0918] Input: Raw data (images and video feeds) from cameras and sensors

[0919] Output: Pre-processed, clean image and video data

[0920] Specific operation: The capturing module receives the video stream from the camera, and the image processing engine removes noise and improves clarity.

[0921] Step 2:

[0922] Analysis of the situation

[0923] The server inputs the preprocessed data into a generative AI model to assess the location and spread of the fire, smoke density, and fire scale.

[0924] Input: Preprocessed image and video data

[0925] Output: Information on fire location, spread, smoke density, and fire scale

[0926] What it does: Apply deep learning models to detect fires in specific areas and visualize their spread with 3D mapping.

[0927] Step 3:

[0928] Obtaining location information

[0929] The device (smartphone or tablet) obtains the user's current location information using the GPS function and sends that information to the server.

[0930] Input: GPS data built into the device

[0931] Output: Current location of the user (latitude and longitude)

[0932] Specific operation: The device acquires GPS data and calls an API to send it to a server via the Internet.

[0933] Step 4:

[0934] Generate evacuation routes

[0935] The server uses an evacuation route generation algorithm to generate the safest evacuation route based on the results of the situation analysis and the user's location information, and sends the generated route information to the terminal.

[0936] Input: Fire location, spread, smoke density, user's current location

[0937] Output: Optimal evacuation route information

[0938] Specific operation: Calculates an evacuation route using Dijkstra's algorithm or A algorithm, and sends the generated route to the terminal as map data.

[0939] Step 5:

[0940] Emotional state assessment

[0941] The device uses a camera and microphone to analyze the user's facial expressions and voice to collect emotional data, which is then sent to a server where an emotion engine is used to evaluate the user's emotional state.

[0942] Input: User's facial expression data and voice data

[0943] Output: User's emotional state (stress level, impatience, etc.)

[0944] Specific operation: The device analyzes the user's facial expressions, collects voice data, and sends it to the server for evaluation by the emotion engine.

[0945] Step 6:

[0946] evacuation order

[0947] The device provides the user with visual and audio evacuation instructions based on the generated evacuation route information and emotional state assessment, and the user begins evacuation in accordance with the instructions.

[0948] Input: Optimal evacuation route information, user's emotional state evaluation

[0949] Output: Visual and audio evacuation instructions

[0950] Specific operation: The device displays a map and a voice guide, and provides instructions such as "Turn right, walk 30 meters, and go down the stairs." The user follows these instructions to evacuate.

[0951] Step 7:

[0952] Generate virtual training scenarios

[0953] The server generates a scenario for the virtual training and sends it to the terminal. The terminal notifies the user of the start of the virtual training and prompts the user to participate in the virtual training.

[0954] Input: Historical scenario data and training data

[0955] Output: Virtual training scenario

[0956] Specific operation: The server generates a virtual scenario in which a fire breaks out on the second floor of a building, and the device notifies the user that training has begun. The user puts on VR goggles and begins training.

[0957] Step 8:

[0958] Data analysis and preventative measures

[0959] The server collects past disaster data and training data, analyzes them using a generative AI model, and evaluates the fire risk in a specific area and suggests preventive measures.

[0960] Input: Past disaster data, training data

[0961] Output: Risk assessment results and preventative measures recommendations

[0962] Specific operation: The server analyzes disaster data, identifies high-risk areas, and presents improvement proposals for fire prevention measures to area managers.

[0963] Step 9:

[0964] Continuous learning and improvement

[0965] The server collects system usage data and new disaster patterns, and continuously updates the generative AI model, incorporating new algorithms and methods for optimizing evacuation routes into the system.

[0966] Input: System usage data, new disaster pattern data

[0967] Output: Updated generative AI model

[0968] Specific operation: The server takes in new data, retrains the AI ​​model, and integrates it into the existing system to update it.

[0969] (Application example 2)

[0970] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0971] Evacuation assistance systems in emergency situations are required to quickly and accurately assist users in evacuating safely. However, conventional evacuation assistance systems have difficulty providing flexible evacuation instructions that take into account the user's emotional state, and they have been particularly inadequate in addressing emotions such as stress and anxiety. Furthermore, there have been no evacuation assistance systems suitable for specific environments, such as factories. Therefore, a system that can appropriately assess the user's emotional state and provide dynamic evacuation instructions based on that assessment is needed to achieve safer and more effective evacuation assistance.

[0972] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0973] In this invention, the server includes means for acquiring and preprocessing images and video feeds of the scene in real time, means for analyzing the acquired data using a generative AI model to evaluate the progress of the disaster, means for acquiring user location information and generating the safest evacuation route based on the analysis results, means for analyzing the user's facial and voice data to evaluate their emotional state, and means for dynamically adjusting the content and method of evacuation instructions based on the generated evacuation route information and the evaluation of their emotional state, thereby enabling flexible and appropriate evacuation support in emergency situations while taking the user's emotional state into consideration.

[0974] "Image and video feed" refers to image and video data used to obtain real-time visual information on the scene during an emergency.

[0975] "Pre-processing" refers to the processing of acquired images and video feeds to improve data quality, such as removing noise or increasing clarity.

[0976] A "generative AI model" is a model that uses machine learning algorithms to predict and evaluate the progression of a disaster and the user's emotional state.

[0977] "Analysis" is the process of analyzing information from acquired data using a generative AI model to assess the progression of the disaster and the user's emotional state.

[0978] "Location information" is data that indicates the user's current geographic coordinates (latitude and longitude).

[0979] An "evacuation route" is the optimal route for a user to safely evacuate in the event of a disaster.

[0980] "Facial expression and vocal data" refers to visual and audio data for assessing a user's emotional state, such as facial expressions and tone of voice.

[0981] "Emotional state assessment" is the process of analyzing the user's facial and vocal data to determine their current emotional state (e.g., stress, impatience, anxiety, etc.).

[0982] "Dynamic adjustment" refers to changing the content and method of evacuation instructions in real time based on the user's emotional state assessment.

[0983] A "virtual training scenario" is a scenario for simulating evacuation behavior in a virtual environment, and is used by users to train.

[0984] An "interactive evacuation drill" is a training exercise in which users participate in a virtual training scenario and experience real-life evacuation procedures.

[0985] "Disaster data and training data" refers to data on disasters that have occurred in the past and data generated in virtual training.

[0986] "Risk assessment" is the process of assessing the risk of disasters occurring in a particular area.

[0987] "Preventive measures" are measures and actions taken to prevent disasters from occurring.

[0988] "Continuously updating" refers to continually improving generative AI models and systems based on new disaster patterns and usage data.

[0989] New patent specification

[0990] This invention relates to a system for supporting safe evacuation in emergency situations. The system performs real-time situation assessment, data analysis, dynamic evacuation support taking into account the user's emotional state, virtual training, evacuation route generation and update, risk assessment based on past data, and proposal of preventive measures.

[0991] Hardware used

[0992] Smartphone: Used to obtain the user's location and emotional state.

[0993] In-factory sensors and cameras: Used to capture real-time images and video feeds of the scene in case of an emergency.

[0994] Server: Performs data analysis, generates evacuation routes, and manages the system.

[0995] Software used

[0996] OpenCV: A library for image processing and data preprocessing.

[0997] TensorFlow: Used to run generative AI models and the emotion engine.

[0998] Firebase: Used as a real-time database to manage location information and analysis results.

[0999] System Components and Functions

[1000] server

[1001] 1. Acquiring and Preprocessing Image and Video Feeds

[1002] The server collects data in real time from cameras and sensors within the factory and uses OpenCV for pre-processing such as noise removal and image clarity improvement.

[1003] 2. Data analysis using generative AI models

[1004] The preprocessed data is analyzed using TensorFlow to assess the progress of the disaster and the user's emotional state.

[1005] 3. Obtaining user location information and generating evacuation routes

[1006] Location information obtained from smartphones is stored in Firebase, and the safest evacuation route is generated based on the analysis results.

[1007] 4. Emotional state assessment and dynamic evacuation instructions

[1008] The system analyzes the user's facial expressions and voice to assess their emotional state, and dynamically adjusts the content and method of evacuation instructions based on the results.

[1009] 5. Creating virtual training scenarios and conducting evacuation drills

[1010] Generate virtual training scenarios and conduct interactive evacuation drills for users.

[1011] 6. Analysis of historical data and proposal of preventive measures

[1012] Past disaster data and training data are analyzed to assess risks and propose preventive measures.

[1013] 7. Continuously updating generative AI models

[1014] The generative AI model is continuously updated based on new disaster patterns and usage data.

[1015] Specific examples

[1016] The server receives images from cameras inside the factory and detects that a fire has broken out in a specific area. The server analyzes the spread of smoke and identifies dangerous areas. At the same time, the smartphone obtains the user's current location information and sends it to the server. The server calculates a route from the user's current location to the nearest emergency exit, taking into account the spread of the fire and the density of the smoke. To evaluate the user's emotional state, the smartphone analyzes facial expressions through the camera and detects signs of stress or impatience. The server then provides more detailed and calming audio guidance to relieve the user's anxiety and stress.

[1017] Prompt Sentence Examples

[1018] Example of an input prompt for a generative AI model: "Identify the location of the fire and the spread of smoke from camera footage inside the factory, and generate an evacuation route."

[1019] Example prompt for the emotion engine: "Please assess employee stress levels based on data from their smartphone cameras and microphones and adjust evacuation instructions accordingly."

[1020] Through the above processing, flexible and appropriate evacuation assistance in emergency situations is possible while taking into consideration the emotional state of the user.

[1021] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1022] Step 1:

[1023] The server acquires images and video feeds in real time from cameras and sensors in the factory. The input of data acquisition is the video data from the cameras and sensors, and the output is the acquired raw data.

[1024] Step 2:

[1025] The server preprocesses the acquired video data using OpenCV. The input for the preprocessing is raw data, and data processing such as noise removal and image clarity improvement is performed, and the output is preprocessed video data.

[1026] Step 3:

[1027] The server analyzes the preprocessed video data using TensorFlow to assess the progression of the disaster. The input for the analysis is the preprocessed video data, and the generative AI model performs data calculations. The output is the disaster detection results and an assessment of the progression of the disaster.

[1028] Step 4:

[1029] The device acquires the user's location information using the GPS function. The input for acquiring location information is the user's current location, and the output is the user's latitude and longitude information.

[1030] Step 5:

[1031] The device sends the acquired location information to Firebase and shares it with the server in real time. The input of the transmission is the user's location information, and the output is the location data stored in Firebase.

[1032] Step 6:

[1033] The server generates the safest evacuation route based on the analysis results and the user's location information stored in Firebase. The inputs for evacuation route generation are the disaster progression assessment and the user's location information, and the output is the optimal evacuation route information.

[1034] Step 7:

[1035] The device acquires facial and voice data of the user using the smartphone's camera and microphone. The input for emotion data acquisition is the user's face and voice, and the output is the acquired facial and voice data.

[1036] Step 8:

[1037] The device analyzes the acquired facial and voice data using TensorFlow to evaluate the user's emotional state. The input for the analysis is facial and voice data, and the data is calculated using a generative AI model, and the output is the evaluated emotional state.

[1038] Step 9:

[1039] The server dynamically adjusts the content and method of evacuation instructions based on the user's emotional state evaluation result, where the input of the dynamic adjustment is the emotional state evaluation and the output is the adjusted evacuation instructions.

[1040] Step 10:

[1041] The terminal provides the user with coordinated evacuation instructions visually and audibly, where the input for providing evacuation instructions is the coordinated evacuation instructions and the output is the visual and audio evacuation instruction information.

[1042] Step 11:

[1043] The server generates a virtual training scenario and sends it to the terminal. The input of the virtual training scenario generation is the past data and the current scenario requirements, and the output is the generated virtual training scenario.

[1044] Step 12:

[1045] The terminal conducts an interactive evacuation drill for the user based on the virtual training scenario, where the input for the training is the virtual training scenario and the output is the training result.

[1046] Step 13:

[1047] The server analyzes past disaster data and training data to assess the risk of a specific area and propose preventive measures. The inputs for risk assessment are past data and training data, and data analysis is performed. The output is the risk assessment results and preventive measures.

[1048] Step 14:

[1049] The server continuously updates the generative AI model based on system usage data and new disaster patterns. The input for the update is usage data and new disaster patterns, and the server performs data calculations, outputting the latest generative AI model.

[1050] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1051] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1052] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[1053] [Third embodiment]

[1054] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1055] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[1056] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1057] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[1058] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1059] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1060] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1061] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1062] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1063] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1064] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1065] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[1066] The AI ​​emergency evacuation support system of the present invention is designed to support safe evacuation in emergency situations. The system functions through cooperation between a server, a terminal, and a user. The program's processing is explained below with specific examples.

[1067] Basic components

[1068] Server: Responsible for real-time acquisition and pre-processing of images and video feeds, data analysis using generative AI models, generation of evacuation routes, generation of virtual training scenarios, data analysis and prevention measures proposals, and usage data updates.

[1069] Terminal: Responsible for obtaining user location information, providing real-time evacuation instructions, and conducting virtual drills.

[1070] User: Follow instructions from the device to take evacuation action and participate in virtual training.

[1071] System action

[1072] Real-time situation analysis

[1073] 1. The server receives real-time images and video feeds from cameras and sensors within the facility.

[1074] 2. The server preprocesses the acquired data and removes unnecessary noise.

[1075] 3. The server uses a generative AI model to analyze the pre-processed data and assess the spread of the fire, smoke density, and fire scale.

[1076] Examples:

[1077] The server analyzes footage from surveillance cameras in real time and detects when a fire has broken out in a specific area.

[1078] The server analyzes the spread of smoke and identifies dangerous areas.

[1079] Obtaining user location information

[1080] 1. The device (mainly a smartphone or tablet) obtains the user's current location information using the GPS function.

[1081] 2. The device sends the acquired location information to the server.

[1082] Examples:

[1083] The user launches the app and allows location sharing.

[1084] The device obtains the user's current latitude and longitude and sends them to the server.

[1085] Generate evacuation routes

[1086] 1. The server generates the optimal evacuation route based on the results of real-time situation analysis and the user's location information.

[1087] 2. The server sends the generated evacuation route information to the terminal.

[1088] Examples:

[1089] The server calculates the route from the user's current location to the nearest emergency exit, taking into account the spread of the fire and the density of the smoke.

[1090] The server transmits map data showing evacuation routes to the terminal.

[1091] Evacuation instructions to users

[1092] 1. The device provides the user with visual and audio evacuation instructions based on the received evacuation route information.

[1093] 2. The user follows the instructions on the device to evacuate safely.

[1094] Examples:

[1095] The device will then provide voice instructions such as, "Turn right, walk 30 meters, and go down the stairs."

[1096] The user follows the instructions on the device and takes evacuation action.

[1097] Virtual Training and Simulation

[1098] 1. The server generates a scenario for virtual training and sends it to the terminal.

[1099] 2. The terminal notifies the user that the virtual training has started and displays the training scenario.

[1100] 3. Users participate in virtual drills and take evacuation actions according to simulated scenarios.

[1101] Examples:

[1102] The server generates a hypothetical scenario in which a fire breaks out on the second floor of a building.

[1103] The device instructs the user to put on the VR goggles and begin training.

[1104] Users practice evacuation plans in a virtual environment.

[1105] Data analysis and preventative measures

[1106] 1. The server collects past disaster data and training data and analyzes them using a generative AI model.

[1107] 2. The server assesses the fire risk in a particular area and suggests preventative measures.

[1108] Examples:

[1109] The server analyzes past fire data to identify fire frequency and high-risk areas.

[1110] The server will suggest preventative measures to the area administrator.

[1111] Continuous learning and improvement

[1112] 1. The server collects system usage data and new disaster patterns and continuously updates the generative AI model.

[1113] 2. The server learns new algorithms and methods for optimizing evacuation routes and reflects them in the system.

[1114] Examples:

[1115] The server analyzes data from annual evacuation drills and learns that certain evacuation routes are prone to congestion.

[1116] The server will propose a new evacuation route and update the system accordingly.

[1117] In this way, the AI ​​emergency evacuation support system of the present invention supports users in safely evacuating through real-time situation assessment, analysis, instructions, training, and continuous learning.

[1118] The processing flow will be explained below.

[1119] Step 1:

[1120] The server collects real-time data from surveillance cameras and sensors installed within the facility, including still images, video feeds, and temperature sensor data.

[1121] Step 2:

[1122] The server pre-processes the collected image and video feed data to remove noise and improve image clarity, which improves the accuracy of the analysis.

[1123] Step 3:

[1124] The server uses a generative AI model to analyze the pre-processed data and evaluate the location of the fire, its spread, smoke density, and fire scale, thereby determining the level of urgency at the scene.

[1125] Step 4:

[1126] The device obtains the user's current location information using the GPS function. If the user carries a smartphone or tablet, the location information can be updated continuously through the app.

[1127] Step 5:

[1128] The device sends the acquired user location information to the server, which allows the server to determine the user's current location.

[1129] Step 6:

[1130] The server then calculates the safest evacuation route based on the analysis results and the user's location information, taking into account factors such as the spread of the fire, smoke density, and congestion on evacuation routes.

[1131] Step 7:

[1132] The server then sends the generated evacuation route information to the user's device. The sent data includes the specific evacuation route and its details (e.g., how many meters to go, which direction to turn, etc.).

[1133] Step 8:

[1134] Based on the received evacuation route information, the device provides the user with visual and audio evacuation instructions, including map display and audio guidance.

[1135] Step 9:

[1136] The user follows the instructions on the device to begin evacuation along a safe evacuation route. By following the instructions, the user can evacuate safely.

[1137] Step 10:

[1138] The server generates virtual training scenarios during normal times and transmits them to the terminals. Specific training scenarios include fire outbreak scenarios and evacuation scenarios.

[1139] Step 11:

[1140] The device notifies the user of the start of virtual training and prompts them to train in the virtual environment. Users can participate in the virtual training using VR goggles or a smartphone.

[1141] Step 12:

[1142] Users participate in virtual training and take evacuation actions according to simulated fire scenarios. After the training, they receive feedback from the system to identify areas for improvement in their evacuation actions.

[1143] Step 13:

[1144] The server collects past disaster data and training data and analyzes it using a generative AI model, which then generates a risk assessment for a specific area.

[1145] Step 14:

[1146] Based on the risk assessment results, the server will suggest preventative measures to managers of specific areas, such as installing sprinkler systems and strengthening evacuation drills.

[1147] Step 15:

[1148] The server collects system usage data and new disaster patterns to continuously update the generative AI model, providing evacuation routes and preventative measures that reflect the latest knowledge.

[1149] Example 1

[1150] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1151] While quick and safe evacuation is extremely important in an emergency, current systems have difficulty obtaining disaster information in real time and providing optimal evacuation routes. Furthermore, inaccurate location information for users and lack of evacuation drills could lead to confusion during an actual disaster. Furthermore, past disaster data and drill data are not fully utilized, preventing preventative measures from being proposed and the system from being continuously improved.

[1152] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1153] In this invention, the server includes means for acquiring and preprocessing images and video feeds of the scene in real time, means for analyzing the acquired data using a generative AI model to evaluate the progress of the disaster, means for acquiring user location information and generating the safest evacuation route based on the analysis results, means for transmitting the generated evacuation route information to the user's device and providing audio and visual evacuation instructions, means for generating virtual training scenarios and conducting interactive evacuation drills with the user, means for analyzing past disaster data and training data to assess risk and propose preventive measures for specific areas, means for continuously updating the generative AI model based on system usage data and new disaster patterns, means for acquiring location information using the GPS function of the user's device, means for analyzing the spread of fire and smoke concentration using the generative AI model, and means for notifying the user of the start of the virtual training and displaying the training scenario using VR goggles, thereby enabling real-time situation assessment and evacuation instructions to be given, thereby achieving safe evacuation.

[1154] "Images and video feeds" refers to visual information obtained in real time from surveillance cameras and sensors.

[1155] "Preprocessing" refers to the preparation of data before analysis, such as removing noise from the acquired data and correcting outliers.

[1156] A "generative AI model" is an artificial intelligence model that uses machine learning and deep learning algorithms to analyze data.

[1157] "Disaster progression" refers to changes in the situation and scope of impact in emergencies such as fires, earthquakes, and floods.

[1158] "User location information" refers to the user's current geographic coordinate information obtained using the GPS function.

[1159] An "evacuation route" is the optimal route to safety from an emergency situation.

[1160] An "evacuation instruction" is an audio and visual guide that encourages the user to take evacuation action.

[1161] A "virtual training scenario" is a scenario that allows a user to conduct an evacuation drill using virtual reality or simulation.

[1162] "Past disaster data" refers to records and information about disasters that have occurred in the past.

[1163] "Risk assessment" is the evaluation of the likelihood and impact of a disaster in a particular area or situation.

[1164] "Preventive measures" are measures proposed to prevent disasters from occurring.

[1165] "System Usage Data" means all data collected through the System.

[1166] "New disaster patterns" refer to new trends and characteristics of disasters that have never been seen before.

[1167] "GPS function" refers to a function that obtains location information using the Global Positioning System.

[1168] "Smoke density" is an indicator of the density or amount of smoke present in the air.

[1169] "VR goggles" are devices that allow users to experience a virtual reality environment.

[1170] This invention relates to an AI emergency evacuation support system that supports safe evacuation in emergency situations. The system of the invention functions through cooperation between a server, a terminal, and a user.

[1171] First, the server acquires real-time images and video feeds from cameras and sensors within the facility. Examples of hardware used include surveillance cameras and smoke detection sensors. The data obtained from these devices undergoes pre-processing, including noise removal and outlier correction. Specifically, video data is filtered and sensor data is corrected to ensure data quality.

[1172] The pre-processed data is then analyzed using generative AI models to assess fire spread, smoke density, and fire scale. For example, image analysis can detect the location and spread of a fire in real time.

[1173] The devices are primarily smartphones and tablets, which use GPS to obtain the user's current location information. The location information is then sent from the device to a server, which then uses it to calculate evacuation routes.

[1174] The server generates an optimal evacuation route based on the results of real-time situation analysis and the user's location information. The generated evacuation route information is sent to the device. Based on this information, the device provides the user with visual and audio evacuation instructions. For example, specific instructions such as "Turn right, walk 30 meters, and go down the stairs" are provided by audio guidance.

[1175] The server also generates scenarios for virtual training and sends them to the device. At this time, the user receives a notification from the device and begins the virtual training. By wearing VR goggles, evacuation training can be carried out in the virtual environment. As a specific example, a virtual scenario is generated in which "a fire breaks out on the second floor of a building," and the user conducts virtual training based on this scenario.

[1176] The server collects past disaster data and training data and analyzes it using a generative AI model. This allows it to assess the fire risk in a specific area and propose preventative measures. For example, it can analyze past fire data to identify patterns of fire occurrence in specific time periods or areas and propose measures to strengthen fire prevention equipment.

[1177] Furthermore, the server continuously updates the generative AI model based on system usage data and new disaster patterns. By learning new algorithms and evacuation route optimization methods and incorporating them into the system, the system can learn that certain evacuation routes are prone to congestion based on annual training data and suggest new evacuation routes.

[1178] As a result, the AI ​​emergency evacuation support system of the present invention supports users in safely evacuating through real-time situation assessment, analysis, instructions, training, and continuous learning.

[1179] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1180] Step 1:

[1181] The server receives real-time images and video feeds from cameras and sensors within the facility. This input data includes surveillance camera footage and smoke detection sensor data. Specifically, it collects camera stream data and converts it into a different format. The output is raw image and video data that has not yet been preprocessed.

[1182] Step 2:

[1183] The server preprocesses the acquired data and removes unwanted noise. The input data is the image and video data obtained in step 1, and a filtering algorithm is applied to remove noise. Specifically, a low-pass filter is applied to the video data, and outlier correction is performed on the sensor data. The output is high-quality data with noise removed.

[1184] Step 3:

[1185] The server uses a generative AI model to analyze the preprocessed data and evaluate the spread of the fire, smoke density, and fire scale. The input data is the output of step 2, and is fed into the generative AI model for analysis. Specifically, the deep learning model is used to identify the extent of the fire's expansion and calculate smoke density. The output is the progress of the fire and the risk assessment results.

[1186] Step 4:

[1187] The device (smartphone or tablet) obtains the user's current location information using its GPS function. Input data is obtained from the user's device and includes latitude and longitude information. Specifically, it collects real-time data from the GPS sensor to obtain location information. The output is the user's precise location information.

[1188] Step 5:

[1189] The device sends the acquired location information to the server. The input data is the location information from step 4, which is encrypted using a security protocol and sent to the server. The specific operation is to establish a secure communication channel and send the location information. The output is the location information received by the server.

[1190] Step 6:

[1191] The server generates an optimal evacuation route based on the results of real-time situation analysis and the user's location information. The input data are the risk assessment results from step 3 and the location information from step 5. Specifically, it applies a route calculation algorithm to generate a safe evacuation route. The output is evacuation route information customized for each user.

[1192] Step 7:

[1193] The server sends the generated evacuation route information to the terminal. The input data is the evacuation route information from step 6, which is sent to the user's terminal. Specifically, the server sends the evacuation route map in digital format. The output is the evacuation route information displayed on the terminal.

[1194] Step 8:

[1195] The device provides the user with visual and audio evacuation instructions based on the received evacuation route information. The input data is the evacuation route information from step 7, and this is notified to the user through a visual map display and audio guidance. Specifically, the device provides audio guidance of direction and distance information through the application. The output is the evacuation instructions received by the user.

[1196] Step 9:

[1197] The user follows the instructions on the terminal to evacuate safely. The input data is the evacuation instructions from step 8, and upon receiving this, the process moves on to actual evacuation behavior. Specifically, the user follows the instructions on the terminal and evacuates by following a safe route. The output is to reach a safe area.

[1198] (Application example 1)

[1199] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1200] Conventional evacuation support systems have difficulty responding to emergencies in real time, and have been unable to provide effective evacuation support, especially in large facilities or complex environments. Furthermore, it has been difficult to provide appropriate evacuation instructions to users with visual or hearing impairments. Furthermore, there has been a problem in that preventive measures and continuous improvements have not been proposed using past training data or disaster data.

[1201] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1202] In this invention, the server includes: means for acquiring and preprocessing images and video feeds of the scene in real time; means for analyzing the acquired data using a generative AI model to assess the progress of the disaster; means for acquiring user location information and generating the safest evacuation route based on the analysis results; means for transmitting the generated evacuation route information to the user's device and providing audio and visual evacuation instructions; means for generating virtual training scenarios and conducting interactive evacuation drills with the user; means for analyzing past disaster data and training data to assess risk and propose preventative measures for specific areas; means for continuously updating the generative AI model based on system usage data and new disaster patterns; and means installed in smart glasses for generating an optimal evacuation route based on the user's current location and providing visual and audio evacuation instructions. This enables effective evacuation support even in large facilities and complex environments, and provides appropriate evacuation instructions to visually and hearing-impaired individuals. It also enables the system to propose preventative measures and continuously improve the system by utilizing past training data and disaster data.

[1203] An "emergency" is an event that poses a threat to people's safety and lives, such as an unexpected disaster or accident.

[1204] An "evacuation support system" is a system that includes a series of technologies and devices to assist people in safely evacuating in the event of an emergency.

[1205] "On-site images and video feed" refers to real-time visual information of the location and surroundings of an emergency.

[1206] "Preprocessing" refers to the process of removing unnecessary noise from acquired image or video data and converting it into a format that is easier to analyze.

[1207] A "generative AI model" is an artificial intelligence model that learns from large amounts of data and has the ability to analyze situations and generate appropriate evacuation routes.

[1208] "Disaster Progression" refers to the process of damage expansion and current situation in emergencies such as fires and earthquakes.

[1209] "User Location Information" means an individual user's current geographic location obtained using GPS or other location-determining technology.

[1210] An "evacuation route" refers to a route that guides a user to the safest evacuation route in the event of an emergency.

[1211] "Terminal" refers to a mobile information terminal used by a user, such as a smartphone, tablet, or smart glasses.

[1212] A "virtual training scenario" refers to a scenario that uses virtual reality and simulation to allow people to realistically experience an emergency evacuation drill.

[1213] "Past disaster data" refers to records and data relating to disasters that have occurred in the past.

[1214] "Training data" refers to data collected during evacuation drills and is used to improve the system and assess risks.

[1215] "Risk assessment" is the act of assessing the possibility of a disaster occurring in a specific area and the extent of its impact.

[1216] "Preventive measures" refer to strategies and measures to reduce risk and ensure safety before a disaster occurs.

[1217] "Continuous updates" refers to the process by which systems and generative AI models learn new data and techniques, keeping them constantly up to date.

[1218] The present invention provides a system for supporting safe evacuation in an emergency situation, which functions in cooperation with a server, terminals, and users.

[1219] Basic components

[1220] 1. Server: Acquires and preprocesses images and video feeds from the scene in real time. It also analyzes the data using a generative AI model to assess the progress of the disaster. It generates the safest evacuation route based on the user's location information and analysis results and sends this to the device. It also generates virtual training scenarios and analyzes past disaster data and training data to assess risks and propose preventive measures. It is responsible for continuously updating the generative AI model based on system usage data and new disaster patterns.

[1221] 2. Terminal: This is a user device, typically a smart eyeglass or smartphone, that acquires the user's current location. It provides visual and audio evacuation instructions based on evacuation route information received from the server. It also provides training scenarios and conducts interactive training with the user during virtual training.

[1222] 3. User: Follows instructions from the terminal to take evacuation action. Participates in virtual training and acts according to the evacuation plan provided by the system.

[1223] System action

[1224] Hardware and Software

[1225] On-site cameras and sensors: Capture images and video feeds in real time.

[1226] GPS module: Obtains the user's location information.

[1227] Smart glasses: Provide visual and audio evacuation instructions.

[1228] Server: Responsible for data processing and running generative AI models.

[1229] Data processing and calculation

[1230] 1. The server acquires images and video feeds from on-site cameras and sensors in real time and performs pre-processing such as noise removal. The pre-processed data is then analyzed using a generative AI model to assess the progression of the disaster.

[1231] 2. The device acquires the user's location information using GPS and sends it to the server, which then combines this information with the analysis results to generate the optimal evacuation route.

[1232] 3. The server sends the generated evacuation route to the terminal, and the terminal provides the user with visual and audio evacuation instructions.

[1233] 4. The virtual training scenario is generated by the server and sent to the user's device. The user wears the smart glasses and conducts an interactive evacuation drill.

[1234] Examples of concrete examples and prompts

[1235] For example, if a fire breaks out in a factory, the server will analyze the footage from the factory's surveillance cameras in real time to identify the location and spread of the fire. It will then calculate the safest evacuation route based on the user's current location and provide visual and audio evacuation instructions through the smart glasses. The following prompt sentences are used:

[1236] The server analyzes the footage from the factory's surveillance cameras in real time and detects a fire in Section 4. Based on the current location of the workers (Section 3) obtained via GPS and the location of the fire, generate the optimal evacuation route. Then, send this evacuation route to the smart glasses and initiate visual and audio instructions.

[1237] In this way, the AI ​​emergency evacuation support system of the present invention can support safe and rapid evacuation in emergency situations through real-time situation analysis, generation of evacuation plans, visual and audio instructions to users, and even virtual training.

[1238] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1239] Step 1:

[1240] Acquiring and pre-processing images and video feeds from the scene

[1241] The server receives real-time image and video feeds from cameras and sensors on-site. The input is raw image and video data from the site. This data is preprocessed to remove unwanted noise, for example by smoothing the image or enhancing edges, to convert it into a format that is easier to analyze. The output is the preprocessed image and video data.

[1242] Step 2:

[1243] Data analysis using generative AI models

[1244] The server inputs preprocessed image and video data into a generative AI model for data analysis. Specifically, it evaluates the spread of the fire, smoke density, and the scale of the disaster. The input is preprocessed image and video data, and the output is various analytical results on the progression of the disaster. The generative AI model accomplishes this by generating data to determine the necessary rescue operations and evacuation routes.

[1245] Step 3:

[1246] Obtaining user location information

[1247] The device obtains the user's current location information using its GPS function. The input is location data from the GPS module, which acquires the user's latitude and longitude information. The device generates a data packet to send this location information to the server and sends it to the server. The output is the user's exact current location, which the server receives.

[1248] Step 4:

[1249] Generate evacuation routes

[1250] The server generates the optimal evacuation route based on the acquired user location information and the disaster analysis results generated by the AI ​​model. For example, it determines the safest route by taking into account the spread of fire and smoke density. The inputs are the user's current location and the disaster analysis results, and the output is the optimal route information for the user to evacuate.

[1251] Step 5:

[1252] Sending and instructing evacuation routes

[1253] The server sends the generated evacuation route information to the terminal. The terminal receives this and provides the user with visual and audio evacuation instructions. Specifically, the terminal issues specific instructions via voice, such as "Turn right, walk 30 meters, and go down the stairs." The input is the evacuation route information sent from the server, and the output is the specific evacuation instructions received by the user.

[1254] Step 6:

[1255] Generate and conduct virtual training scenarios

[1256] The server generates a scenario for the virtual training and sends it to the terminal. The terminal notifies the user of the start of the virtual training and displays the training scenario. The user takes evacuation actions in the virtual environment according to the scenario. The input is the generated virtual training scenario, and the output is the results of the virtual training performed by the user.

[1257] Step 7:

[1258] Data analysis and preventative measures

[1259] The server collects past disaster data and training data and analyzes it using a generative AI model. As a result of the analysis, it performs a risk assessment of a specific area and proposes appropriate preventive measures. For example, it recommends installing additional fire detectors in areas where fires frequently occur. The inputs are past disaster data and training data, and the output is the risk assessment results and proposed preventive measures.

[1260] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1261] This invention provides flexible evacuation support that takes into account the user's emotional state by combining an emotion engine with a system for supporting safe evacuation in emergency situations. The following describes the system components and program processing with specific examples.

[1262] Basic components

[1263] Server: Responsible for real-time acquisition and pre-processing of images and video feeds, data analysis using generative AI models, evaluation of users' emotional state using an emotion engine, generation of evacuation routes, generation of virtual training scenarios, data analysis and prevention measures proposals, and usage data updates.

[1264] Terminal: Responsible for obtaining user location information, collecting emotion engine data, providing real-time evacuation instructions, and conducting virtual drills.

[1265] User: Follows instructions from the device to take evacuation action, provides emotional data, and participates in virtual training.

[1266] System action

[1267] Real-time situation analysis

[1268] 1. The server receives real-time images and video feeds from cameras and sensors within the facility.

[1269] 2. The server preprocesses the acquired data to remove noise and improve image clarity.

[1270] 3. The server analyzes the pre-processed data using a generative AI model to assess the location and spread of the fire, smoke density, and fire scale.

[1271] Examples:

[1272] The server analyzes footage from surveillance cameras and detects that a fire has broken out in a specific area.

[1273] The server analyzes the spread of smoke and identifies dangerous areas.

[1274] Obtaining user location information

[1275] 1. The device (smartphone or tablet) obtains the user's current location information using the GPS function.

[1276] 2. The device sends the acquired location information to the server.

[1277] Examples:

[1278] The user launches the app and allows location sharing.

[1279] The device obtains the user's current latitude and longitude and sends them to the server.

[1280] Generate evacuation routes

[1281] 1. The server generates the safest evacuation route based on the results of real-time situation analysis and the user's location information.

[1282] 2. The server sends the generated evacuation route information to the terminal.

[1283] Examples:

[1284] The server calculates the route from the user's current location to the nearest emergency exit, taking into account the spread of the fire and the density of the smoke.

[1285] The server transmits map data showing evacuation routes to the terminal.

[1286] Assessing the user's emotional state

[1287] 1. The device uses a camera and microphone to analyze the user's facial expressions and voice and collect emotional data.

[1288] 2. The device sends the collected emotion data to the server.

[1289] 3. The server uses an emotion engine to evaluate the user's emotional state and dynamically adjusts the content and method of evacuation instructions based on the results.

[1290] Examples:

[1291] The device analyzes the user's facial expressions through the camera and detects signs of stress or impatience.

[1292] The server provides more detailed and calming audio guidance to ease the user's anxiety and stress.

[1293] Evacuation instructions to users

[1294] 1. The device provides visual and audio evacuation instructions to the user based on the received evacuation route information and emotional state assessment.

[1295] 2. The user begins evacuating along a safe evacuation route following the instructions on the device.

[1296] Examples:

[1297] The device will then provide voice instructions such as, "Turn right, walk 30 meters, and go down the stairs."

[1298] The user follows the instructions on the device and evacuates along the designated evacuation route.

[1299] Virtual Training and Simulation

[1300] 1. The server generates a scenario for virtual training and sends it to the terminal.

[1301] 2. The device notifies the user that virtual training has begun and prompts them to train in the virtual environment.

[1302] 3. Users participate in virtual drills and take evacuation actions according to simulated scenarios.

[1303] Examples:

[1304] The server generates a hypothetical scenario in which a fire breaks out on the second floor of a building.

[1305] The device instructs the user to put on the VR goggles and begin training.

[1306] Users practice evacuation plans in a virtual environment.

[1307] Data analysis and preventative measures

[1308] 1. The server collects past disaster data and training data and analyzes them using a generative AI model.

[1309] 2. The server assesses the fire risk in a particular area and suggests preventative measures.

[1310] Examples:

[1311] The server analyzes past fire data to identify fire frequency and high-risk areas.

[1312] The server will suggest preventative measures to the area administrator.

[1313] Continuous learning and improvement

[1314] 1. The server collects system usage data and new disaster patterns and continuously updates the generative AI model.

[1315] 2. The server learns new algorithms and methods for optimizing evacuation routes and reflects them in the system.

[1316] Examples:

[1317] The server analyzes data from annual evacuation drills and learns that certain evacuation routes are prone to congestion.

[1318] The server will propose a new evacuation route and update the system accordingly.

[1319] In this way, the AI ​​emergency evacuation support system of the present invention supports users' safe evacuation through real-time situational awareness, data analysis, evaluation of the user's emotional state using an emotion engine, dynamic evacuation instructions, training and simulation, data analysis and prevention measures proposals, and continuous learning and improvement. In addition, the use of an emotion engine enables flexible evacuation support that takes the user's psychological state into consideration.

[1320] The processing flow will be explained below.

[1321] Step 1:

[1322] The server collects real-time images and video feeds from security cameras and sensors installed within the facility, including security camera footage, temperature sensor data, and smoke detector sensor data.

[1323] Step 2:

[1324] The server pre-processes the collected image and video feed data, removing noise, and optionally changes the image resolution or compresses the data.

[1325] Step 3:

[1326] The server analyzes the pre-processed data using a generative AI model to assess the location of the fire, its spread, smoke density, and fire scale, thereby determining the level of urgency at the scene.

[1327] Step 4:

[1328] The device (smartphone or tablet) acquires the user's current location information using the GPS function. When the user launches the emergency app, collection of location information begins.

[1329] Step 5:

[1330] The device sends the acquired user location information to the server, which then calculates an evacuation route based on the user's accurate location information at the time of the disaster.

[1331] Step 6:

[1332] The server calculates the safest evacuation route based on the results of real-time situation analysis and the user's location information, taking into account factors such as the spread of the fire, smoke density, congestion on evacuation routes, and the overall layout of the building.

[1333] Step 7:

[1334] The server then sends the generated evacuation route information to the user's terminal. The evacuation route information includes specific routes and evacuation procedures.

[1335] Step 8:

[1336] Based on the received evacuation route information, the device provides the user with visual and audio evacuation instructions, including displaying a map and providing audio guidance.

[1337] Step 9:

[1338] The user follows the instructions on the device and begins evacuating along the designated evacuation route. By taking appropriate evacuation actions, danger can be minimized.

[1339] Step 10:

[1340] The device uses a camera and microphone to analyze the user's facial expressions and voice to collect emotional data, and an emotion engine evaluates the user's stress level and anxiety state.

[1341] Step 11:

[1342] The device sends the collected emotional data to a server, which allows the user's psychological state to be grasped in real time.

[1343] Step 12:

[1344] The server uses an emotion engine to assess the user's emotional state and dynamically adjusts the evacuation instructions and methods based on the results, providing more detailed and calming guidance when stress and anxiety levels are high.

[1345] Step 13:

[1346] The device receives and provides emotion-dependent instructions to the user, and visual or audio guidance may change depending on the emotional state.

[1347] Step 14:

[1348] By receiving instructions that take emotions into consideration, users can continue their evacuation activities with peace of mind, which is expected to reduce stress.

[1349] Step 15:

[1350] The server generates virtual training scenarios during normal times and transmits them to the terminals. Specific training scenarios include fire outbreak scenarios and evacuation scenarios.

[1351] Step 16:

[1352] The device notifies the user that virtual training has started and prompts them to train in the virtual environment. The user participates in the virtual training using VR goggles or a smartphone.

[1353] Step 17:

[1354] Users participate in virtual training and take evacuation actions according to simulated fire scenarios. After the training, they receive feedback from the system to identify areas for improvement in their evacuation actions.

[1355] Step 18:

[1356] The server collects past disaster data and training data and analyzes it using a generative AI model to assess the fire risk in a specific area.

[1357] Step 19:

[1358] Based on the risk assessment results, the server will suggest preventative measures to managers of specific areas, such as installing sprinkler systems and strengthening evacuation drills.

[1359] Step 20:

[1360] The server collects system usage data and new disaster patterns to continuously update the generative AI model, providing evacuation routes and preventative measures that reflect the latest knowledge.

[1361] Example 2

[1362] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1363] The purpose of this invention is to provide evacuation support in emergency situations by grasping the situation on-site in real time, providing safe and rapid evacuation routes, and providing flexible and effective evacuation instructions that take into account the user's emotional state. It also aims to provide a more advanced evacuation support system by practicing evacuation behavior through virtual training scenarios, proposing preventive measures, and continuously improving the system.

[1364] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1365] In this invention, the server includes means for acquiring and preprocessing images and video feeds of the scene in real time, means for analyzing the acquired data using a generative AI model to evaluate the progress of the disaster, means for acquiring user location information and generating the safest evacuation route based on the analysis results, means for acquiring the user's emotional state using a camera and microphone and evaluating it using an emotion engine, means for providing audio and visual evacuation instructions based on the generated evacuation route information and emotional state evaluation, means for generating virtual training scenarios and conducting interactive evacuation drills with the user, means for analyzing past disaster data and training data to assess risks and propose preventive measures for specific areas, and means for continuously updating the generative AI model based on system usage data and new disaster patterns, thereby enabling real-time situation assessment, provision of safe evacuation routes, emotionally sensitive evacuation instructions, interactive training, and proposal of preventive measures, as well as continuous system improvement.

[1366] An "emergency situation" refers to an emergency situation that requires a rapid response, such as a natural disaster or man-made trouble.

[1367] "Evacuation" refers to the act of escaping danger and moving to a safe place.

[1368] "System" refers to a collection of devices and software in which multiple elements work together to perform a specific function.

[1369] "Image and video feed" refers to still and video data acquired through cameras and sensors.

[1370] "Preprocessing" refers to initial processing such as improving the quality of acquired data and removing noise.

[1371] A "generative AI model" refers to an algorithmic model of artificial intelligence generated using machine learning or deep learning techniques.

[1372] "Analyzing data" refers to the process of extracting information from acquired data and drawing meaningful conclusions.

[1373] "Disaster progression" refers to the extent of damage and the scope of impact in emergencies such as fires and earthquakes.

[1374] "User location information" refers to data about a user's current location obtained by GPS or other positioning systems.

[1375] An "evacuation route" refers to a route that is considered safe for escaping from a disaster.

[1376] "Emotional state" refers to the user's psychological state, such as stress or impatience.

[1377] An "emotion engine" refers to an algorithm or model for analyzing and assessing a user's emotional state.

[1378] "Evacuation instructions" refers to specific instructions or guidance that encourage users to take evacuation action.

[1379] "Virtual training scenario" refers to a virtual training environment and storyline designed around a simulated emergency situation.

[1380] "Interactive evacuation drill" refers to a training process in which users can actively participate within a virtual environment.

[1381] "Historical disaster data" refers to records and information relating to previous emergencies.

[1382] "Training data" refers to information such as behavior and responses collected during training.

[1383] "Risk assessment" refers to the process of quantitatively assessing the likelihood of a disaster occurring in a particular area.

[1384] "Preventive measures" refer to measures and actions taken to minimize the occurrence and impact of disasters.

[1385] "Usage Data" refers to all information collected in the course of using the System.

[1386] "Disaster patterns" refer to patterns or models that show how a particular disaster will unfold based on historical data and statistics.

[1387] An "update" refers to updating an existing model or system by adding new data or algorithms.

[1388] The present invention is a system for supporting safe evacuation in an emergency situation, and is implemented with the following configuration and method.

[1389] System configuration

[1390] The basic components are as follows:

[1391] Server: Responsible for real-time acquisition and pre-processing of images and video feeds, data analysis using generative AI models, evaluation of users' emotional state using an emotion engine, generation of evacuation routes, generation of virtual training scenarios, data analysis and prevention measures proposals, and usage data updates.

[1392] Terminal: Responsible for obtaining user location information, collecting emotion engine data, providing real-time evacuation instructions, and conducting virtual drills.

[1393] User: Follows instructions from the device to take evacuation action, provides emotional data, and participates in virtual training.

[1394] System action

[1395] 1. Data acquisition and preprocessing

[1396] The server receives real-time images and video feeds from cameras and sensors installed within the facility, and uses noise filtering algorithms to remove noise from the data and image enhancement techniques to improve clarity.

[1397] Example: The video stream from a surveillance camera is received by the capturing module and pre-processed by the image processing engine.

[1398] 2. Situation Analysis

[1399] The pre-processed data is analyzed on a server, and a generative AI model is used to assess the location and spread of fires, smoke density, and fire scale, identifying areas with a particularly high probability of fire outbreak.

[1400] Example: A deep learning model detects fires in a specific area and visualizes their spread with 3D mapping.

[1401] 3. Obtain location information

[1402] The device (smartphone or tablet) obtains the user's current location information using the GPS function and sends that information to the server.

[1403] Example: A user launches an app and the device sends GPS data to a server.

[1404] 4. Evacuation route generation

[1405] The server generates the safest evacuation route based on the results of real-time situation analysis and the user's location information using Dijkstra's algorithm and A algorithm. The generated route is sent to the device as map data.

[1406] Example: The server calculates an evacuation route and sends map data to the device.

[1407] 5. Emotional state assessment

[1408] The device uses a camera and microphone to analyze the user's facial expressions and voice to collect emotional data, which is then sent to a server, where an emotion engine is used to evaluate the user's emotional state and dynamically adjust evacuation instructions as needed.

[1409] Example: The device analyzes the user's facial expressions, detects signs of impatience, and the server changes the audio guidance to a calmer version.

[1410] 6. Evacuation instructions

[1411] Based on the received evacuation route information and emotional state assessment, the device provides the user with visual and audio evacuation instructions, which the user follows to begin evacuation along a safe evacuation route.

[1412] Example: The device will give voice instructions such as "Turn right, walk 30 meters, and go down the stairs," and the user will follow the instructions to evacuate.

[1413] 7. Virtual Training

[1414] The server generates a scenario for the virtual training and sends it to the terminal. The terminal notifies the user of the start of the virtual training and prompts the user to participate in the training. The user participates in the virtual training and takes evacuation actions according to the simulated scenario.

[1415] Example: The server generates a virtual scenario in which a fire breaks out on the second floor of a building, and the device instructs the user to put on VR goggles and begin training.

[1416] 8. Data analysis and preventive measures

[1417] The server collects past disaster data and training data, analyzes them using a generative AI model, and evaluates the fire risk in a specific area and suggests preventive measures.

[1418] Example: The server analyzes past fire data and presents improvement proposals for fire prevention measures to managers of high-risk areas.

[1419] 9. Continuous learning and improvement

[1420] The server collects system usage data and new disaster patterns, continually updating the generative AI model, learning new algorithms and methods for optimizing evacuation routes and incorporating them into the system.

[1421] Example: The server analyzes annual evacuation drill data, proposes new evacuation routes, and reflects them in the system.

[1422] Example prompts for generative AI models

[1423] Prompt: "Generate a virtual training scenario, analyze the user's stress level, and suggest the optimal evacuation route for the user. Please provide a detailed description of the specific training scenario and its steps, and finally include evacuation instructions for the user."

[1424] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1425] Step 1:

[1426] Data acquisition and preprocessing

[1427] The server captures real-time images and video feeds from cameras and sensors installed within the facility, then uses noise filtering algorithms to remove noise from the captured data and image enhancement techniques to improve clarity.

[1428] Input: Raw data (images and video feeds) from cameras and sensors

[1429] Output: Pre-processed, clean image and video data

[1430] Specific operation: The capturing module receives the video stream from the camera, and the image processing engine removes noise and improves clarity.

[1431] Step 2:

[1432] Analysis of the situation

[1433] The server inputs the preprocessed data into a generative AI model to assess the location and spread of the fire, smoke density, and fire scale.

[1434] Input: Preprocessed image and video data

[1435] Output: Information on fire location, spread, smoke density, and fire scale

[1436] What it does: Apply deep learning models to detect fires in specific areas and visualize their spread with 3D mapping.

[1437] Step 3:

[1438] Obtaining location information

[1439] The device (smartphone or tablet) obtains the user's current location information using the GPS function and sends that information to the server.

[1440] Input: GPS data built into the device

[1441] Output: Current location of the user (latitude and longitude)

[1442] Specific operation: The device acquires GPS data and calls an API to send it to a server via the Internet.

[1443] Step 4:

[1444] Generate evacuation routes

[1445] The server uses an evacuation route generation algorithm to generate the safest evacuation route based on the results of the situation analysis and the user's location information, and sends the generated route information to the terminal.

[1446] Input: Fire location, spread, smoke density, user's current location

[1447] Output: Optimal evacuation route information

[1448] Specific operation: Calculates an evacuation route using Dijkstra's algorithm or A algorithm, and sends the generated route to the terminal as map data.

[1449] Step 5:

[1450] Emotional state assessment

[1451] The device uses a camera and microphone to analyze the user's facial expressions and voice to collect emotional data, which is then sent to a server where an emotion engine is used to evaluate the user's emotional state.

[1452] Input: User's facial expression data and voice data

[1453] Output: User's emotional state (stress level, impatience, etc.)

[1454] Specific operation: The device analyzes the user's facial expressions, collects voice data, and sends it to the server for evaluation by the emotion engine.

[1455] Step 6:

[1456] evacuation order

[1457] The device provides the user with visual and audio evacuation instructions based on the generated evacuation route information and emotional state assessment, and the user begins evacuation in accordance with the instructions.

[1458] Input: Optimal evacuation route information, user's emotional state evaluation

[1459] Output: Visual and audio evacuation instructions

[1460] Specific operation: The device displays a map and a voice guide, and provides instructions such as "Turn right, walk 30 meters, and go down the stairs." The user follows these instructions to evacuate.

[1461] Step 7:

[1462] Generate virtual training scenarios

[1463] The server generates a scenario for the virtual training and sends it to the terminal. The terminal notifies the user of the start of the virtual training and prompts the user to participate in the virtual training.

[1464] Input: Historical scenario data and training data

[1465] Output: Virtual training scenario

[1466] Specific operation: The server generates a virtual scenario in which a fire breaks out on the second floor of a building, and the device notifies the user that training has begun. The user puts on VR goggles and begins training.

[1467] Step 8:

[1468] Data analysis and preventative measures

[1469] The server collects past disaster data and training data, analyzes them using a generative AI model, and evaluates the fire risk in a specific area and suggests preventive measures.

[1470] Input: Past disaster data, training data

[1471] Output: Risk assessment results and preventative measures recommendations

[1472] Specific operation: The server analyzes disaster data, identifies high-risk areas, and presents improvement proposals for fire prevention measures to area managers.

[1473] Step 9:

[1474] Continuous learning and improvement

[1475] The server collects system usage data and new disaster patterns, and continuously updates the generative AI model, incorporating new algorithms and methods for optimizing evacuation routes into the system.

[1476] Input: System usage data, new disaster pattern data

[1477] Output: Updated generative AI model

[1478] Specific operation: The server takes in new data, retrains the AI ​​model, and integrates it into the existing system to update it.

[1479] (Application example 2)

[1480] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1481] Evacuation assistance systems in emergency situations are required to quickly and accurately assist users in evacuating safely. However, conventional evacuation assistance systems have difficulty providing flexible evacuation instructions that take into account the user's emotional state, and they have been particularly inadequate in addressing emotions such as stress and anxiety. Furthermore, there have been no evacuation assistance systems suitable for specific environments, such as factories. Therefore, a system that can appropriately assess the user's emotional state and provide dynamic evacuation instructions based on that assessment is needed to achieve safer and more effective evacuation assistance.

[1482] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1483] In this invention, the server includes means for acquiring and preprocessing images and video feeds of the scene in real time, means for analyzing the acquired data using a generative AI model to evaluate the progress of the disaster, means for acquiring user location information and generating the safest evacuation route based on the analysis results, means for analyzing the user's facial and voice data to evaluate their emotional state, and means for dynamically adjusting the content and method of evacuation instructions based on the generated evacuation route information and the evaluation of their emotional state, thereby enabling flexible and appropriate evacuation support in emergency situations while taking the user's emotional state into consideration.

[1484] "Image and video feed" refers to image and video data used to obtain real-time visual information on the scene during an emergency.

[1485] "Pre-processing" refers to the processing of acquired images and video feeds to improve data quality, such as removing noise or increasing clarity.

[1486] A "generative AI model" is a model that uses machine learning algorithms to predict and evaluate the progression of a disaster and the user's emotional state.

[1487] "Analysis" is the process of analyzing information from acquired data using a generative AI model to assess the progression of the disaster and the user's emotional state.

[1488] "Location information" is data that indicates the user's current geographic coordinates (latitude and longitude).

[1489] An "evacuation route" is the optimal route for a user to safely evacuate in the event of a disaster.

[1490] "Facial expression and vocal data" refers to visual and audio data for assessing a user's emotional state, such as facial expressions and tone of voice.

[1491] "Emotional state assessment" is the process of analyzing the user's facial and vocal data to determine their current emotional state (e.g., stress, impatience, anxiety, etc.).

[1492] "Dynamic adjustment" refers to changing the content and method of evacuation instructions in real time based on the user's emotional state assessment.

[1493] A "virtual training scenario" is a scenario for simulating evacuation behavior in a virtual environment, and is used by users to train.

[1494] An "interactive evacuation drill" is a training exercise in which users participate in a virtual training scenario and experience real-life evacuation procedures.

[1495] "Disaster data and training data" refers to data on disasters that have occurred in the past and data generated in virtual training.

[1496] "Risk assessment" is the process of assessing the risk of disasters occurring in a particular area.

[1497] "Preventive measures" are measures and actions taken to prevent disasters from occurring.

[1498] "Continuously updating" refers to continually improving generative AI models and systems based on new disaster patterns and usage data.

[1499] New patent specification

[1500] This invention relates to a system for supporting safe evacuation in emergency situations. The system performs real-time situation assessment, data analysis, dynamic evacuation support taking into account the user's emotional state, virtual training, evacuation route generation and update, risk assessment based on past data, and proposal of preventive measures.

[1501] Hardware used

[1502] Smartphone: Used to obtain the user's location and emotional state.

[1503] In-factory sensors and cameras: Used to capture real-time images and video feeds of the scene in case of an emergency.

[1504] Server: Performs data analysis, generates evacuation routes, and manages the system.

[1505] Software used

[1506] OpenCV: A library for image processing and data preprocessing.

[1507] TensorFlow: Used to run generative AI models and the emotion engine.

[1508] Firebase: Used as a real-time database to manage location information and analysis results.

[1509] System Components and Functions

[1510] server

[1511] 1. Acquiring and Preprocessing Image and Video Feeds

[1512] The server collects data in real time from cameras and sensors within the factory and uses OpenCV for pre-processing such as noise removal and image clarity improvement.

[1513] 2. Data analysis using generative AI models

[1514] The preprocessed data is analyzed using TensorFlow to assess the progress of the disaster and the user's emotional state.

[1515] 3. Obtaining user location information and generating evacuation routes

[1516] Location information obtained from smartphones is stored in Firebase, and the safest evacuation route is generated based on the analysis results.

[1517] 4. Emotional state assessment and dynamic evacuation instructions

[1518] The system analyzes the user's facial expressions and voice to assess their emotional state, and dynamically adjusts the content and method of evacuation instructions based on the results.

[1519] 5. Creating virtual training scenarios and conducting evacuation drills

[1520] Generate virtual training scenarios and conduct interactive evacuation drills for users.

[1521] 6. Analysis of historical data and proposal of preventive measures

[1522] Past disaster data and training data are analyzed to assess risks and propose preventive measures.

[1523] 7. Continuously updating generative AI models

[1524] The generative AI model is continuously updated based on new disaster patterns and usage data.

[1525] Specific examples

[1526] The server receives images from cameras inside the factory and detects that a fire has broken out in a specific area. The server analyzes the spread of smoke and identifies dangerous areas. At the same time, the smartphone obtains the user's current location information and sends it to the server. The server calculates a route from the user's current location to the nearest emergency exit, taking into account the spread of the fire and the density of the smoke. To evaluate the user's emotional state, the smartphone analyzes facial expressions through the camera and detects signs of stress or impatience. The server then provides more detailed and calming audio guidance to relieve the user's anxiety and stress.

[1527] Prompt Sentence Examples

[1528] Example of an input prompt for a generative AI model: "Identify the location of the fire and the spread of smoke from camera footage inside the factory, and generate an evacuation route."

[1529] Example prompt for the emotion engine: "Please assess employee stress levels based on data from their smartphone cameras and microphones and adjust evacuation instructions accordingly."

[1530] Through the above processing, flexible and appropriate evacuation assistance in emergency situations is possible while taking into consideration the emotional state of the user.

[1531] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1532] Step 1:

[1533] The server acquires images and video feeds in real time from cameras and sensors in the factory. The input of data acquisition is the video data from the cameras and sensors, and the output is the acquired raw data.

[1534] Step 2:

[1535] The server preprocesses the acquired video data using OpenCV. The input for the preprocessing is raw data, and data processing such as noise removal and image clarity improvement is performed, and the output is preprocessed video data.

[1536] Step 3:

[1537] The server analyzes the preprocessed video data using TensorFlow to assess the progression of the disaster. The input for the analysis is the preprocessed video data, and the generative AI model performs data calculations. The output is the disaster detection results and an assessment of the progression of the disaster.

[1538] Step 4:

[1539] The device acquires the user's location information using the GPS function. The input for acquiring location information is the user's current location, and the output is the user's latitude and longitude information.

[1540] Step 5:

[1541] The device sends the acquired location information to Firebase and shares it with the server in real time. The input of the transmission is the user's location information, and the output is the location data stored in Firebase.

[1542] Step 6:

[1543] The server generates the safest evacuation route based on the analysis results and the user's location information stored in Firebase. The inputs for evacuation route generation are the disaster progression assessment and the user's location information, and the output is the optimal evacuation route information.

[1544] Step 7:

[1545] The device acquires facial and voice data of the user using the smartphone's camera and microphone. The input for emotion data acquisition is the user's face and voice, and the output is the acquired facial and voice data.

[1546] Step 8:

[1547] The device analyzes the acquired facial and voice data using TensorFlow to evaluate the user's emotional state. The input for the analysis is facial and voice data, and the data is calculated using a generative AI model, and the output is the evaluated emotional state.

[1548] Step 9:

[1549] The server dynamically adjusts the content and method of evacuation instructions based on the user's emotional state evaluation result, where the input of the dynamic adjustment is the emotional state evaluation and the output is the adjusted evacuation instructions.

[1550] Step 10:

[1551] The terminal provides the user with coordinated evacuation instructions visually and audibly, where the input for providing evacuation instructions is the coordinated evacuation instructions and the output is the visual and audio evacuation instruction information.

[1552] Step 11:

[1553] The server generates a virtual training scenario and sends it to the terminal. The input of the virtual training scenario generation is the past data and the current scenario requirements, and the output is the generated virtual training scenario.

[1554] Step 12:

[1555] The terminal conducts an interactive evacuation drill for the user based on the virtual training scenario, where the input for the training is the virtual training scenario and the output is the training result.

[1556] Step 13:

[1557] The server analyzes past disaster data and training data to assess the risk of a specific area and propose preventive measures. The inputs for risk assessment are past data and training data, and data analysis is performed. The output is the risk assessment results and preventive measures.

[1558] Step 14:

[1559] The server continuously updates the generative AI model based on system usage data and new disaster patterns. The input for the update is usage data and new disaster patterns, and the server performs data calculations, outputting the latest generative AI model.

[1560] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1561] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1562] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1563] [Fourth embodiment]

[1564] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1565] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1566] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1567] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1568] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1569] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1570] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1571] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1572] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1573] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1574] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1575] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1576] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1577] The AI ​​emergency evacuation support system of the present invention is designed to support safe evacuation in emergency situations. The system functions through cooperation between a server, a terminal, and a user. The program's processing is explained below with specific examples.

[1578] Basic components

[1579] Server: Responsible for real-time acquisition and pre-processing of images and video feeds, data analysis using generative AI models, generation of evacuation routes, generation of virtual training scenarios, data analysis and prevention measures proposals, and usage data updates.

[1580] Terminal: Responsible for obtaining user location information, providing real-time evacuation instructions, and conducting virtual drills.

[1581] User: Follow instructions from the device to take evacuation action and participate in virtual training.

[1582] System action

[1583] Real-time situation analysis

[1584] 1. The server receives real-time images and video feeds from cameras and sensors within the facility.

[1585] 2. The server preprocesses the acquired data and removes unnecessary noise.

[1586] 3. The server uses a generative AI model to analyze the pre-processed data and assess the spread of the fire, smoke density, and fire scale.

[1587] Examples:

[1588] The server analyzes footage from surveillance cameras in real time and detects when a fire has broken out in a specific area.

[1589] The server analyzes the spread of smoke and identifies dangerous areas.

[1590] Obtaining user location information

[1591] 1. The device (mainly a smartphone or tablet) obtains the user's current location information using the GPS function.

[1592] 2. The device sends the acquired location information to the server.

[1593] Examples:

[1594] The user launches the app and allows location sharing.

[1595] The device obtains the user's current latitude and longitude and sends them to the server.

[1596] Generate evacuation routes

[1597] 1. The server generates the optimal evacuation route based on the results of real-time situation analysis and the user's location information.

[1598] 2. The server sends the generated evacuation route information to the terminal.

[1599] Examples:

[1600] The server calculates the route from the user's current location to the nearest emergency exit, taking into account the spread of the fire and the density of the smoke.

[1601] The server transmits map data showing evacuation routes to the terminal.

[1602] Evacuation instructions to users

[1603] 1. The device provides the user with visual and audio evacuation instructions based on the received evacuation route information.

[1604] 2. The user follows the instructions on the device to evacuate safely.

[1605] Examples:

[1606] The device will then provide voice instructions such as, "Turn right, walk 30 meters, and go down the stairs."

[1607] The user follows the instructions on the device and takes evacuation action.

[1608] Virtual Training and Simulation

[1609] 1. The server generates a scenario for virtual training and sends it to the terminal.

[1610] 2. The terminal notifies the user that the virtual training has started and displays the training scenario.

[1611] 3. Users participate in virtual drills and take evacuation actions according to simulated scenarios.

[1612] Examples:

[1613] The server generates a hypothetical scenario in which a fire breaks out on the second floor of a building.

[1614] The device instructs the user to put on the VR goggles and begin training.

[1615] Users practice evacuation plans in a virtual environment.

[1616] Data analysis and preventative measures

[1617] 1. The server collects past disaster data and training data and analyzes them using a generative AI model.

[1618] 2. The server assesses the fire risk in a particular area and suggests preventative measures.

[1619] Examples:

[1620] The server analyzes past fire data to identify fire frequency and high-risk areas.

[1621] The server will suggest preventative measures to the area administrator.

[1622] Continuous learning and improvement

[1623] 1. The server collects system usage data and new disaster patterns and continuously updates the generative AI model.

[1624] 2. The server learns new algorithms and methods for optimizing evacuation routes and reflects them in the system.

[1625] Examples:

[1626] The server analyzes data from annual evacuation drills and learns that certain evacuation routes are prone to congestion.

[1627] The server will propose a new evacuation route and update the system accordingly.

[1628] In this way, the AI ​​emergency evacuation support system of the present invention supports users in safely evacuating through real-time situation assessment, analysis, instructions, training, and continuous learning.

[1629] The processing flow will be explained below.

[1630] Step 1:

[1631] The server collects real-time data from surveillance cameras and sensors installed within the facility, including still images, video feeds, and temperature sensor data.

[1632] Step 2:

[1633] The server pre-processes the collected image and video feed data to remove noise and improve image clarity, which improves the accuracy of the analysis.

[1634] Step 3:

[1635] The server uses a generative AI model to analyze the pre-processed data and evaluate the location of the fire, its spread, smoke density, and fire scale, thereby determining the level of urgency at the scene.

[1636] Step 4:

[1637] The device obtains the user's current location information using the GPS function. If the user carries a smartphone or tablet, the location information can be updated continuously through the app.

[1638] Step 5:

[1639] The device sends the acquired user location information to the server, which allows the server to determine the user's current location.

[1640] Step 6:

[1641] The server then calculates the safest evacuation route based on the analysis results and the user's location information, taking into account factors such as the spread of the fire, smoke density, and congestion on evacuation routes.

[1642] Step 7:

[1643] The server then sends the generated evacuation route information to the user's device. The sent data includes the specific evacuation route and its details (e.g., how many meters to go, which direction to turn, etc.).

[1644] Step 8:

[1645] Based on the received evacuation route information, the device provides the user with visual and audio evacuation instructions, including map display and audio guidance.

[1646] Step 9:

[1647] The user follows the instructions on the device to begin evacuation along a safe evacuation route. By following the instructions, the user can evacuate safely.

[1648] Step 10:

[1649] The server generates virtual training scenarios during normal times and transmits them to the terminals. Specific training scenarios include fire outbreak scenarios and evacuation scenarios.

[1650] Step 11:

[1651] The device notifies the user of the start of virtual training and prompts them to train in the virtual environment. Users can participate in the virtual training using VR goggles or a smartphone.

[1652] Step 12:

[1653] Users participate in virtual training and take evacuation actions according to simulated fire scenarios. After the training, they receive feedback from the system to identify areas for improvement in their evacuation actions.

[1654] Step 13:

[1655] The server collects past disaster data and training data and analyzes it using a generative AI model, which then generates a risk assessment for a specific area.

[1656] Step 14:

[1657] Based on the risk assessment results, the server will suggest preventative measures to managers of specific areas, such as installing sprinkler systems and strengthening evacuation drills.

[1658] Step 15:

[1659] The server collects system usage data and new disaster patterns to continuously update the generative AI model, providing evacuation routes and preventative measures that reflect the latest knowledge.

[1660] Example 1

[1661] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1662] While quick and safe evacuation is extremely important in an emergency, current systems have difficulty obtaining disaster information in real time and providing optimal evacuation routes. Furthermore, inaccurate location information for users and lack of evacuation drills could lead to confusion during an actual disaster. Furthermore, past disaster data and drill data are not fully utilized, preventing preventative measures from being proposed and the system from being continuously improved.

[1663] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1664] In this invention, the server includes means for acquiring and preprocessing images and video feeds of the scene in real time, means for analyzing the acquired data using a generative AI model to evaluate the progress of the disaster, means for acquiring user location information and generating the safest evacuation route based on the analysis results, means for transmitting the generated evacuation route information to the user's device and providing audio and visual evacuation instructions, means for generating virtual training scenarios and conducting interactive evacuation drills with the user, means for analyzing past disaster data and training data to assess risk and propose preventive measures for specific areas, means for continuously updating the generative AI model based on system usage data and new disaster patterns, means for acquiring location information using the GPS function of the user's device, means for analyzing the spread of fire and smoke concentration using the generative AI model, and means for notifying the user of the start of the virtual training and displaying the training scenario using VR goggles, thereby enabling real-time situation assessment and evacuation instructions to be given, thereby achieving safe evacuation.

[1665] "Images and video feeds" refers to visual information obtained in real time from surveillance cameras and sensors.

[1666] "Preprocessing" refers to the preparation of data before analysis, such as removing noise from the acquired data and correcting outliers.

[1667] A "generative AI model" is an artificial intelligence model that uses machine learning and deep learning algorithms to analyze data.

[1668] "Disaster progression" refers to changes in the situation and scope of impact in emergencies such as fires, earthquakes, and floods.

[1669] "User location information" refers to the user's current geographic coordinate information obtained using the GPS function.

[1670] An "evacuation route" is the optimal route to safety from an emergency situation.

[1671] An "evacuation instruction" is an audio and visual guide that encourages the user to take evacuation action.

[1672] A "virtual training scenario" is a scenario that allows a user to conduct an evacuation drill using virtual reality or simulation.

[1673] "Past disaster data" refers to records and information about disasters that have occurred in the past.

[1674] "Risk assessment" is the evaluation of the likelihood and impact of a disaster in a particular area or situation.

[1675] "Preventive measures" are measures proposed to prevent disasters from occurring.

[1676] "System Usage Data" means all data collected through the System.

[1677] "New disaster patterns" refer to new trends and characteristics of disasters that have never been seen before.

[1678] "GPS function" refers to a function that obtains location information using the Global Positioning System.

[1679] "Smoke density" is an indicator of the density or amount of smoke present in the air.

[1680] "VR goggles" are devices that allow users to experience a virtual reality environment.

[1681] This invention relates to an AI emergency evacuation support system that supports safe evacuation in emergency situations. The system of the invention functions through cooperation between a server, a terminal, and a user.

[1682] First, the server acquires real-time images and video feeds from cameras and sensors within the facility. Examples of hardware used include surveillance cameras and smoke detection sensors. The data obtained from these devices undergoes pre-processing, including noise removal and outlier correction. Specifically, video data is filtered and sensor data is corrected to ensure data quality.

[1683] The pre-processed data is then analyzed using generative AI models to assess fire spread, smoke density, and fire scale. For example, image analysis can detect the location and spread of a fire in real time.

[1684] The devices are primarily smartphones and tablets, which use GPS to obtain the user's current location information. The location information is then sent from the device to a server, which then uses it to calculate evacuation routes.

[1685] The server generates an optimal evacuation route based on the results of real-time situation analysis and the user's location information. The generated evacuation route information is sent to the device. Based on this information, the device provides the user with visual and audio evacuation instructions. For example, specific instructions such as "Turn right, walk 30 meters, and go down the stairs" are provided by audio guidance.

[1686] The server also generates scenarios for virtual training and sends them to the device. At this time, the user receives a notification from the device and begins the virtual training. By wearing VR goggles, evacuation training can be carried out in the virtual environment. As a specific example, a virtual scenario is generated in which "a fire breaks out on the second floor of a building," and the user conducts virtual training based on this scenario.

[1687] The server collects past disaster data and training data and analyzes it using a generative AI model. This allows it to assess the fire risk in a specific area and propose preventative measures. For example, it can analyze past fire data to identify patterns of fire occurrence in specific time periods or areas and propose measures to strengthen fire prevention equipment.

[1688] Furthermore, the server continuously updates the generative AI model based on system usage data and new disaster patterns. By learning new algorithms and evacuation route optimization methods and incorporating them into the system, the system can learn that certain evacuation routes are prone to congestion based on annual training data and suggest new evacuation routes.

[1689] As a result, the AI ​​emergency evacuation support system of the present invention supports users in safely evacuating through real-time situation assessment, analysis, instructions, training, and continuous learning.

[1690] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1691] Step 1:

[1692] The server receives real-time images and video feeds from cameras and sensors within the facility. This input data includes surveillance camera footage and smoke detection sensor data. Specifically, it collects camera stream data and converts it into a different format. The output is raw image and video data that has not yet been preprocessed.

[1693] Step 2:

[1694] The server preprocesses the acquired data and removes unwanted noise. The input data is the image and video data obtained in step 1, and a filtering algorithm is applied to remove noise. Specifically, a low-pass filter is applied to the video data, and outlier correction is performed on the sensor data. The output is high-quality data with noise removed.

[1695] Step 3:

[1696] The server uses a generative AI model to analyze the preprocessed data and evaluate the spread of the fire, smoke density, and fire scale. The input data is the output of step 2, and is fed into the generative AI model for analysis. Specifically, the deep learning model is used to identify the extent of the fire's expansion and calculate smoke density. The output is the progress of the fire and the risk assessment results.

[1697] Step 4:

[1698] The device (smartphone or tablet) obtains the user's current location information using its GPS function. Input data is obtained from the user's device and includes latitude and longitude information. Specifically, it collects real-time data from the GPS sensor to obtain location information. The output is the user's precise location information.

[1699] Step 5:

[1700] The device sends the acquired location information to the server. The input data is the location information from step 4, which is encrypted using a security protocol and sent to the server. The specific operation is to establish a secure communication channel and send the location information. The output is the location information received by the server.

[1701] Step 6:

[1702] The server generates an optimal evacuation route based on the results of real-time situation analysis and the user's location information. The input data are the risk assessment results from step 3 and the location information from step 5. Specifically, it applies a route calculation algorithm to generate a safe evacuation route. The output is evacuation route information customized for each user.

[1703] Step 7:

[1704] The server sends the generated evacuation route information to the terminal. The input data is the evacuation route information from step 6, which is sent to the user's terminal. Specifically, the server sends the evacuation route map in digital format. The output is the evacuation route information displayed on the terminal.

[1705] Step 8:

[1706] The device provides the user with visual and audio evacuation instructions based on the received evacuation route information. The input data is the evacuation route information from step 7, and this is notified to the user through a visual map display and audio guidance. Specifically, the device provides audio guidance of direction and distance information through the application. The output is the evacuation instructions received by the user.

[1707] Step 9:

[1708] The user follows the instructions on the terminal to evacuate safely. The input data is the evacuation instructions from step 8, and upon receiving this, the process moves on to actual evacuation behavior. Specifically, the user follows the instructions on the terminal and evacuates by following a safe route. The output is to reach a safe area.

[1709] (Application example 1)

[1710] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1711] Conventional evacuation support systems have difficulty responding to emergencies in real time, and have been unable to provide effective evacuation support, especially in large facilities or complex environments. Furthermore, it has been difficult to provide appropriate evacuation instructions to users with visual or hearing impairments. Furthermore, there has been a problem in that preventive measures and continuous improvements have not been proposed using past training data or disaster data.

[1712] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1713] In this invention, the server includes: means for acquiring and preprocessing images and video feeds of the scene in real time; means for analyzing the acquired data using a generative AI model to assess the progress of the disaster; means for acquiring user location information and generating the safest evacuation route based on the analysis results; means for transmitting the generated evacuation route information to the user's device and providing audio and visual evacuation instructions; means for generating virtual training scenarios and conducting interactive evacuation drills with the user; means for analyzing past disaster data and training data to assess risk and propose preventative measures for specific areas; means for continuously updating the generative AI model based on system usage data and new disaster patterns; and means installed in smart glasses for generating an optimal evacuation route based on the user's current location and providing visual and audio evacuation instructions. This enables effective evacuation support even in large facilities and complex environments, and provides appropriate evacuation instructions to visually and hearing-impaired individuals. It also enables the system to propose preventative measures and continuously improve the system by utilizing past training data and disaster data.

[1714] An "emergency" is an event that poses a threat to people's safety and lives, such as an unexpected disaster or accident.

[1715] An "evacuation support system" is a system that includes a series of technologies and devices to assist people in safely evacuating in the event of an emergency.

[1716] "On-site images and video feed" refers to real-time visual information of the location and surroundings of an emergency.

[1717] "Preprocessing" refers to the process of removing unnecessary noise from acquired image or video data and converting it into a format that is easier to analyze.

[1718] A "generative AI model" is an artificial intelligence model that learns from large amounts of data and has the ability to analyze situations and generate appropriate evacuation routes.

[1719] "Disaster Progression" refers to the process of damage expansion and current situation in emergencies such as fires and earthquakes.

[1720] "User Location Information" means an individual user's current geographic location obtained using GPS or other location-determining technology.

[1721] An "evacuation route" refers to a route that guides a user to the safest evacuation route in the event of an emergency.

[1722] "Terminal" refers to a mobile information terminal used by a user, such as a smartphone, tablet, or smart glasses.

[1723] A "virtual training scenario" refers to a scenario that uses virtual reality and simulation to allow people to realistically experience an emergency evacuation drill.

[1724] "Past disaster data" refers to records and data relating to disasters that have occurred in the past.

[1725] "Training data" refers to data collected during evacuation drills and is used to improve the system and assess risks.

[1726] "Risk assessment" is the act of assessing the possibility of a disaster occurring in a specific area and the extent of its impact.

[1727] "Preventive measures" refer to strategies and measures to reduce risk and ensure safety before a disaster occurs.

[1728] "Continuous updates" refers to the process by which systems and generative AI models learn new data and techniques, keeping them constantly up to date.

[1729] The present invention provides a system for supporting safe evacuation in an emergency situation, which functions in cooperation with a server, terminals, and users.

[1730] Basic components

[1731] 1. Server: Acquires and preprocesses images and video feeds from the scene in real time. It also analyzes the data using a generative AI model to assess the progress of the disaster. It generates the safest evacuation route based on the user's location information and analysis results and sends this to the device. It also generates virtual training scenarios and analyzes past disaster data and training data to assess risks and propose preventive measures. It is responsible for continuously updating the generative AI model based on system usage data and new disaster patterns.

[1732] 2. Terminal: This is a user device, typically a smart eyeglass or smartphone, that acquires the user's current location. It provides visual and audio evacuation instructions based on evacuation route information received from the server. It also provides training scenarios and conducts interactive training with the user during virtual training.

[1733] 3. User: Follows instructions from the terminal to take evacuation action. Participates in virtual training and acts according to the evacuation plan provided by the system.

[1734] System action

[1735] Hardware and Software

[1736] On-site cameras and sensors: Capture images and video feeds in real time.

[1737] GPS module: Obtains the user's location information.

[1738] Smart glasses: Provide visual and audio evacuation instructions.

[1739] Server: Responsible for data processing and running generative AI models.

[1740] Data processing and calculation

[1741] 1. The server acquires images and video feeds from on-site cameras and sensors in real time and performs pre-processing such as noise removal. The pre-processed data is then analyzed using a generative AI model to assess the progression of the disaster.

[1742] 2. The device acquires the user's location information using GPS and sends it to the server, which then combines this information with the analysis results to generate the optimal evacuation route.

[1743] 3. The server sends the generated evacuation route to the terminal, and the terminal provides the user with visual and audio evacuation instructions.

[1744] 4. The virtual training scenario is generated by the server and sent to the user's device. The user wears the smart glasses and conducts an interactive evacuation drill.

[1745] Examples of concrete examples and prompts

[1746] For example, if a fire breaks out in a factory, the server will analyze the footage from the factory's surveillance cameras in real time to identify the location and spread of the fire. It will then calculate the safest evacuation route based on the user's current location and provide visual and audio evacuation instructions through the smart glasses. The following prompt sentences are used:

[1747] The server analyzes the footage from the factory's surveillance cameras in real time and detects a fire in Section 4. Based on the current location of the workers (Section 3) obtained via GPS and the location of the fire, generate the optimal evacuation route. Then, send this evacuation route to the smart glasses and initiate visual and audio instructions.

[1748] In this way, the AI ​​emergency evacuation support system of the present invention can support safe and rapid evacuation in emergency situations through real-time situation analysis, generation of evacuation plans, visual and audio instructions to users, and even virtual training.

[1749] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1750] Step 1:

[1751] Acquiring and pre-processing images and video feeds from the scene

[1752] The server receives real-time image and video feeds from cameras and sensors on-site. The input is raw image and video data from the site. This data is preprocessed to remove unwanted noise, for example by smoothing the image or enhancing edges, to convert it into a format that is easier to analyze. The output is the preprocessed image and video data.

[1753] Step 2:

[1754] Data analysis using generative AI models

[1755] The server inputs preprocessed image and video data into a generative AI model for data analysis. Specifically, it evaluates the spread of the fire, smoke density, and the scale of the disaster. The input is preprocessed image and video data, and the output is various analytical results on the progression of the disaster. The generative AI model accomplishes this by generating data to determine the necessary rescue operations and evacuation routes.

[1756] Step 3:

[1757] Obtaining user location information

[1758] The device obtains the user's current location information using its GPS function. The input is location data from the GPS module, which acquires the user's latitude and longitude information. The device generates a data packet to send this location information to the server and sends it to the server. The output is the user's exact current location, which the server receives.

[1759] Step 4:

[1760] Generate evacuation routes

[1761] The server generates the optimal evacuation route based on the acquired user location information and the disaster analysis results generated by the AI ​​model. For example, it determines the safest route by taking into account the spread of fire and smoke density. The inputs are the user's current location and the disaster analysis results, and the output is the optimal route information for the user to evacuate.

[1762] Step 5:

[1763] Sending and instructing evacuation routes

[1764] The server sends the generated evacuation route information to the terminal. The terminal receives this and provides the user with visual and audio evacuation instructions. Specifically, the terminal issues specific instructions via voice, such as "Turn right, walk 30 meters, and go down the stairs." The input is the evacuation route information sent from the server, and the output is the specific evacuation instructions received by the user.

[1765] Step 6:

[1766] Generate and conduct virtual training scenarios

[1767] The server generates a scenario for the virtual training and sends it to the terminal. The terminal notifies the user of the start of the virtual training and displays the training scenario. The user takes evacuation actions in the virtual environment according to the scenario. The input is the generated virtual training scenario, and the output is the results of the virtual training performed by the user.

[1768] Step 7:

[1769] Data analysis and preventative measures

[1770] The server collects past disaster data and training data and analyzes it using a generative AI model. As a result of the analysis, it performs a risk assessment of a specific area and proposes appropriate preventive measures. For example, it recommends installing additional fire detectors in areas where fires frequently occur. The inputs are past disaster data and training data, and the output is the risk assessment results and proposed preventive measures.

[1771] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1772] This invention provides flexible evacuation support that takes into account the user's emotional state by combining an emotion engine with a system for supporting safe evacuation in emergency situations. The following describes the system components and program processing with specific examples.

[1773] Basic components

[1774] Server: Responsible for real-time acquisition and pre-processing of images and video feeds, data analysis using generative AI models, evaluation of users' emotional state using an emotion engine, generation of evacuation routes, generation of virtual training scenarios, data analysis and prevention measures proposals, and usage data updates.

[1775] Terminal: Responsible for obtaining user location information, collecting emotion engine data, providing real-time evacuation instructions, and conducting virtual drills.

[1776] User: Follows instructions from the device to take evacuation action, provides emotional data, and participates in virtual training.

[1777] System action

[1778] Real-time situation analysis

[1779] 1. The server receives real-time images and video feeds from cameras and sensors within the facility.

[1780] 2. The server preprocesses the acquired data to remove noise and improve image clarity.

[1781] 3. The server analyzes the pre-processed data using a generative AI model to assess the location and spread of the fire, smoke density, and fire scale.

[1782] Examples:

[1783] The server analyzes footage from surveillance cameras and detects that a fire has broken out in a specific area.

[1784] The server analyzes the spread of smoke and identifies dangerous areas.

[1785] Obtaining user location information

[1786] 1. The device (smartphone or tablet) obtains the user's current location information using the GPS function.

[1787] 2. The device sends the acquired location information to the server.

[1788] Examples:

[1789] The user launches the app and allows location sharing.

[1790] The device obtains the user's current latitude and longitude and sends them to the server.

[1791] Generate evacuation routes

[1792] 1. The server generates the safest evacuation route based on the results of real-time situation analysis and the user's location information.

[1793] 2. The server sends the generated evacuation route information to the terminal.

[1794] Examples:

[1795] The server calculates the route from the user's current location to the nearest emergency exit, taking into account the spread of the fire and the density of the smoke.

[1796] The server transmits map data showing evacuation routes to the terminal.

[1797] Assessing the user's emotional state

[1798] 1. The device uses a camera and microphone to analyze the user's facial expressions and voice and collect emotional data.

[1799] 2. The device sends the collected emotion data to the server.

[1800] 3. The server uses an emotion engine to evaluate the user's emotional state and dynamically adjusts the content and method of evacuation instructions based on the results.

[1801] Examples:

[1802] The device analyzes the user's facial expressions through the camera and detects signs of stress or impatience.

[1803] The server provides more detailed and calming audio guidance to ease the user's anxiety and stress.

[1804] Evacuation instructions to users

[1805] 1. The device provides visual and audio evacuation instructions to the user based on the received evacuation route information and emotional state assessment.

[1806] 2. The user begins evacuating along a safe evacuation route following the instructions on the device.

[1807] Examples:

[1808] The device will then provide voice instructions such as, "Turn right, walk 30 meters, and go down the stairs."

[1809] The user follows the instructions on the device and evacuates along the designated evacuation route.

[1810] Virtual Training and Simulation

[1811] 1. The server generates a scenario for virtual training and sends it to the terminal.

[1812] 2. The device notifies the user that virtual training has begun and prompts them to train in the virtual environment.

[1813] 3. Users participate in virtual drills and take evacuation actions according to simulated scenarios.

[1814] Examples:

[1815] The server generates a hypothetical scenario in which a fire breaks out on the second floor of a building.

[1816] The device instructs the user to put on the VR goggles and begin training.

[1817] Users practice evacuation plans in a virtual environment.

[1818] Data analysis and preventative measures

[1819] 1. The server collects past disaster data and training data and analyzes them using a generative AI model.

[1820] 2. The server assesses the fire risk in a particular area and suggests preventative measures.

[1821] Examples:

[1822] The server analyzes past fire data to identify fire frequency and high-risk areas.

[1823] The server will suggest preventative measures to the area administrator.

[1824] Continuous learning and improvement

[1825] 1. The server collects system usage data and new disaster patterns and continuously updates the generative AI model.

[1826] 2. The server learns new algorithms and methods for optimizing evacuation routes and reflects them in the system.

[1827] Examples:

[1828] The server analyzes data from annual evacuation drills and learns that certain evacuation routes are prone to congestion.

[1829] The server will propose a new evacuation route and update the system accordingly.

[1830] In this way, the AI ​​emergency evacuation support system of the present invention supports users' safe evacuation through real-time situational awareness, data analysis, evaluation of the user's emotional state using an emotion engine, dynamic evacuation instructions, training and simulation, data analysis and prevention measures proposals, and continuous learning and improvement. In addition, the use of an emotion engine enables flexible evacuation support that takes the user's psychological state into consideration.

[1831] The processing flow will be explained below.

[1832] Step 1:

[1833] The server collects real-time images and video feeds from security cameras and sensors installed within the facility, including security camera footage, temperature sensor data, and smoke detector sensor data.

[1834] Step 2:

[1835] The server pre-processes the collected image and video feed data, removing noise, and optionally changes the image resolution or compresses the data.

[1836] Step 3:

[1837] The server analyzes the pre-processed data using a generative AI model to assess the location of the fire, its spread, smoke density, and fire scale, thereby determining the level of urgency at the scene.

[1838] Step 4:

[1839] The device (smartphone or tablet) acquires the user's current location information using the GPS function. When the user launches the emergency app, collection of location information begins.

[1840] Step 5:

[1841] The device sends the acquired user location information to the server, which then calculates an evacuation route based on the user's accurate location information at the time of the disaster.

[1842] Step 6:

[1843] The server calculates the safest evacuation route based on the results of real-time situation analysis and the user's location information, taking into account factors such as the spread of the fire, smoke density, congestion on evacuation routes, and the overall layout of the building.

[1844] Step 7:

[1845] The server then sends the generated evacuation route information to the user's terminal. The evacuation route information includes specific routes and evacuation procedures.

[1846] Step 8:

[1847] Based on the received evacuation route information, the device provides the user with visual and audio evacuation instructions, including displaying a map and providing audio guidance.

[1848] Step 9:

[1849] The user follows the instructions on the device and begins evacuating along the designated evacuation route. By taking appropriate evacuation actions, danger can be minimized.

[1850] Step 10:

[1851] The device uses a camera and microphone to analyze the user's facial expressions and voice to collect emotional data, and an emotion engine evaluates the user's stress level and anxiety state.

[1852] Step 11:

[1853] The device sends the collected emotional data to a server, which allows the user's psychological state to be grasped in real time.

[1854] Step 12:

[1855] The server uses an emotion engine to assess the user's emotional state and dynamically adjusts the evacuation instructions and methods based on the results, providing more detailed and calming guidance when stress and anxiety levels are high.

[1856] Step 13:

[1857] The device receives and provides emotion-dependent instructions to the user, and visual or audio guidance may change depending on the emotional state.

[1858] Step 14:

[1859] By receiving instructions that take emotions into consideration, users can continue their evacuation activities with peace of mind, which is expected to reduce stress.

[1860] Step 15:

[1861] The server generates virtual training scenarios during normal times and transmits them to the terminals. Specific training scenarios include fire outbreak scenarios and evacuation scenarios.

[1862] Step 16:

[1863] The device notifies the user that virtual training has started and prompts them to train in the virtual environment. The user participates in the virtual training using VR goggles or a smartphone.

[1864] Step 17:

[1865] Users participate in virtual training and take evacuation actions according to simulated fire scenarios. After the training, they receive feedback from the system to identify areas for improvement in their evacuation actions.

[1866] Step 18:

[1867] The server collects past disaster data and training data and analyzes it using a generative AI model to assess the fire risk in a specific area.

[1868] Step 19:

[1869] Based on the risk assessment results, the server will suggest preventative measures to managers of specific areas, such as installing sprinkler systems and strengthening evacuation drills.

[1870] Step 20:

[1871] The server collects system usage data and new disaster patterns to continuously update the generative AI model, providing evacuation routes and preventative measures that reflect the latest knowledge.

[1872] Example 2

[1873] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1874] The purpose of this invention is to provide evacuation support in emergency situations by grasping the situation on-site in real time, providing safe and rapid evacuation routes, and providing flexible and effective evacuation instructions that take into account the user's emotional state. It also aims to provide a more advanced evacuation support system by practicing evacuation behavior through virtual training scenarios, proposing preventive measures, and continuously improving the system.

[1875] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1876] In this invention, the server includes means for acquiring and preprocessing images and video feeds of the scene in real time, means for analyzing the acquired data using a generative AI model to evaluate the progress of the disaster, means for acquiring user location information and generating the safest evacuation route based on the analysis results, means for acquiring the user's emotional state using a camera and microphone and evaluating it using an emotion engine, means for providing audio and visual evacuation instructions based on the generated evacuation route information and emotional state evaluation, means for generating virtual training scenarios and conducting interactive evacuation drills with the user, means for analyzing past disaster data and training data to assess risks and propose preventive measures for specific areas, and means for continuously updating the generative AI model based on system usage data and new disaster patterns, thereby enabling real-time situation assessment, provision of safe evacuation routes, emotionally sensitive evacuation instructions, interactive training, and proposal of preventive measures, as well as continuous system improvement.

[1877] An "emergency situation" refers to an emergency situation that requires a rapid response, such as a natural disaster or man-made trouble.

[1878] "Evacuation" refers to the act of escaping danger and moving to a safe place.

[1879] "System" refers to a collection of devices and software in which multiple elements work together to perform a specific function.

[1880] "Image and video feed" refers to still and video data acquired through cameras and sensors.

[1881] "Preprocessing" refers to initial processing such as improving the quality of acquired data and removing noise.

[1882] A "generative AI model" refers to an algorithmic model of artificial intelligence generated using machine learning or deep learning techniques.

[1883] "Analyzing data" refers to the process of extracting information from acquired data and drawing meaningful conclusions.

[1884] "Disaster progression" refers to the extent of damage and the scope of impact in emergencies such as fires and earthquakes.

[1885] "User location information" refers to data about a user's current location obtained by GPS or other positioning systems.

[1886] An "evacuation route" refers to a route that is considered safe for escaping from a disaster.

[1887] "Emotional state" refers to the user's psychological state, such as stress or impatience.

[1888] An "emotion engine" refers to an algorithm or model for analyzing and assessing a user's emotional state.

[1889] "Evacuation instructions" refers to specific instructions or guidance that encourage users to take evacuation action.

[1890] "Virtual training scenario" refers to a virtual training environment and storyline designed around a simulated emergency situation.

[1891] "Interactive evacuation drill" refers to a training process in which users can actively participate within a virtual environment.

[1892] "Historical disaster data" refers to records and information relating to previous emergencies.

[1893] "Training data" refers to information such as behavior and responses collected during training.

[1894] "Risk assessment" refers to the process of quantitatively assessing the likelihood of a disaster occurring in a particular area.

[1895] "Preventive measures" refer to measures and actions taken to minimize the occurrence and impact of disasters.

[1896] "Usage Data" refers to all information collected in the course of using the System.

[1897] "Disaster patterns" refer to patterns or models that show how a particular disaster will unfold based on historical data and statistics.

[1898] An "update" refers to updating an existing model or system by adding new data or algorithms.

[1899] The present invention is a system for supporting safe evacuation in an emergency situation, and is implemented with the following configuration and method.

[1900] System configuration

[1901] The basic components are as follows:

[1902] Server: Responsible for real-time acquisition and pre-processing of images and video feeds, data analysis using generative AI models, evaluation of users' emotional state using an emotion engine, generation of evacuation routes, generation of virtual training scenarios, data analysis and prevention measures proposals, and usage data updates.

[1903] Terminal: Responsible for obtaining user location information, collecting emotion engine data, providing real-time evacuation instructions, and conducting virtual drills.

[1904] User: Follows instructions from the device to take evacuation action, provides emotional data, and participates in virtual training.

[1905] System action

[1906] 1. Data Acquisition and Preprocessing

[1907] The server receives real-time images and video feeds from cameras and sensors installed within the facility, and uses noise filtering algorithms to remove noise from the data and image enhancement techniques to improve clarity.

[1908] Example: The video stream from a surveillance camera is received by the capturing module and pre-processed by the image processing engine.

[1909] 2. Situation Analysis

[1910] The pre-processed data is analyzed on a server, and a generative AI model is used to assess the location and spread of fires, smoke density, and fire scale, identifying areas with a particularly high probability of fire outbreak.

[1911] Example: A deep learning model detects fires in a specific area and visualizes their spread with 3D mapping.

[1912] 3. Obtain location information

[1913] The device (smartphone or tablet) obtains the user's current location information using the GPS function and sends that information to the server.

[1914] Example: A user launches an app and the device sends GPS data to a server.

[1915] 4. Evacuation route generation

[1916] The server generates the safest evacuation route based on the results of real-time situation analysis and the user's location information using Dijkstra's algorithm and A algorithm. The generated route is sent to the device as map data.

[1917] Example: The server calculates an evacuation route and sends map data to the device.

[1918] 5. Emotional state assessment

[1919] The device uses a camera and microphone to analyze the user's facial expressions and voice to collect emotional data, which is then sent to a server, where an emotion engine is used to evaluate the user's emotional state and dynamically adjust evacuation instructions as needed.

[1920] Example: The device analyzes the user's facial expressions, detects signs of impatience, and the server changes the audio guidance to a calmer version.

[1921] 6. Evacuation instructions

[1922] Based on the received evacuation route information and emotional state assessment, the device provides the user with visual and audio evacuation instructions, which the user follows to begin evacuation along a safe evacuation route.

[1923] Example: The device will give voice instructions such as "Turn right, walk 30 meters, and go down the stairs," and the user will follow the instructions to evacuate.

[1924] 7. Virtual Training

[1925] The server generates a scenario for the virtual training and sends it to the terminal. The terminal notifies the user of the start of the virtual training and prompts the user to participate in the training. The user participates in the virtual training and takes evacuation actions according to the simulated scenario.

[1926] Example: The server generates a virtual scenario in which a fire breaks out on the second floor of a building, and the device instructs the user to put on VR goggles and begin training.

[1927] 8. Data analysis and preventive measures

[1928] The server collects past disaster data and training data, analyzes them using a generative AI model, and evaluates the fire risk in a specific area and suggests preventive measures.

[1929] Example: The server analyzes past fire data and presents improvement proposals for fire prevention measures to managers of high-risk areas.

[1930] 9. Continuous learning and improvement

[1931] The server collects system usage data and new disaster patterns, continually updating the generative AI model, learning new algorithms and methods for optimizing evacuation routes and incorporating them into the system.

[1932] Example: The server analyzes annual evacuation drill data, proposes new evacuation routes, and reflects them in the system.

[1933] Example prompts for generative AI models

[1934] Prompt: "Generate a virtual training scenario, analyze the user's stress level, and suggest the optimal evacuation route for the user. Please provide a detailed description of the specific training scenario and its steps, and finally include evacuation instructions for the user."

[1935] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1936] Step 1:

[1937] Data acquisition and preprocessing

[1938] The server captures real-time images and video feeds from cameras and sensors installed within the facility, then uses noise filtering algorithms to remove noise from the captured data and image enhancement techniques to improve clarity.

[1939] Input: Raw data (images and video feeds) from cameras and sensors

[1940] Output: Pre-processed, clean image and video data

[1941] Specific operation: The capturing module receives the video stream from the camera, and the image processing engine removes noise and improves clarity.

[1942] Step 2:

[1943] Analysis of the situation

[1944] The server inputs the preprocessed data into a generative AI model to assess the location and spread of the fire, smoke density, and fire scale.

[1945] Input: Preprocessed image and video data

[1946] Output: Information on fire location, spread, smoke density, and fire scale

[1947] What it does: Apply deep learning models to detect fires in specific areas and visualize their spread with 3D mapping.

[1948] Step 3:

[1949] Obtaining location information

[1950] The device (smartphone or tablet) obtains the user's current location information using the GPS function and sends that information to the server.

[1951] Input: GPS data built into the device

[1952] Output: Current location of the user (latitude and longitude)

[1953] Specific operation: The device acquires GPS data and calls an API to send it to a server via the Internet.

[1954] Step 4:

[1955] Generate evacuation routes

[1956] The server uses an evacuation route generation algorithm to generate the safest evacuation route based on the results of the situation analysis and the user's location information, and sends the generated route information to the terminal.

[1957] Input: Fire location, spread, smoke de...

Claims

1. A system for supporting safe evacuation in an emergency, comprising: a means of acquiring and pre-processing images and video feeds from the scene in real time; A means of analyzing the acquired data using a generative AI model to assess the progress of the disaster; and A means for acquiring user location information and generating the safest evacuation route based on the analysis results; means for transmitting the generated evacuation route information to a user's terminal and providing audio and visual evacuation instructions; means for generating virtual training scenarios and conducting interactive evacuation drills with users; A means of analyzing past disaster data and training data to assess risk in specific areas and propose preventative measures; A means to continuously update the generative AI model based on system usage data and new disaster patterns; and A system including:

2. The system of claim 1 , wherein the system provides real-time evacuation instructions to a user's device.

3. 10. The system of claim 1, which provides an interface for the visually and hearing impaired.

Images