System

Abstract

A system is provided.SOLUTION: Acquiring water level data, precipitation data, and topographic data of a river from a weather database; converting the acquired data into a unified format; inputting the converted data into a generative AI model to predict a flood risk; and performing risk evaluation based on the predicted flood risk; A system comprising: means for visually displaying a risk level on a map; means for notifying a disaster prevention staff of a local government and residents of a risk evaluation result; means for proposing a evacuation plan based on the notified risk evaluation result; and means for collecting feedback from each local government and residents and using the feedback to improve accuracy of a model.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 flood forecasting systems have limited data collection and analysis capabilities, making it difficult to predict flood risk with high accuracy. Furthermore, there is a lack of means to provide flood risk information to local governments and residents in real time, making it difficult to develop rapid evacuation plans. Furthermore, disaster prevention systems are expensive in developing countries, and have not been widely adopted. The purpose of this invention is to solve these problems and realize safer and faster flood prevention measures. [Means for solving the problem]

[0005] The present invention provides a system that includes a means for acquiring river water level data, precipitation data, and topographical data from a meteorological database, a means for converting the acquired data into a unified format, a means for inputting the converted data into a generative AI model to predict flood risk, a means for conducting a risk assessment based on the predicted flood risk and visually displaying the risk level on a map, a means for notifying disaster prevention personnel in the local government and residents of the results of the risk assessment, a means for proposing an evacuation plan based on the notified risk assessment results, and a means for collecting feedback from each local government and residents and using it to improve the accuracy of the model. This improves the accuracy of flood risk prediction, enables the provision of information in real time, and enables rapid evacuation and effective countermeasures.

[0006] "Weather database" refers to a comprehensive data collection system provided by meteorological agencies and organizations that stores and provides weather data.

[0007] "River water level data" is numerical data that indicates the height of the river's water level, and is basic information necessary for predicting flood risk.

[0008] "Precipitation data" is data that indicates the amount of rain or snow that falls in a specific area within a certain period of time.

[0009] "Topographic data" refers to geographic information system (GIS) data that shows the topography and physical shape of the earth's surface, and is used to assess flood risk.

[0010] A "generative AI model" refers to a machine learning algorithm that discovers patterns in large amounts of data and predicts flood risk.

[0011] A "uniform format" is a data format for converting data obtained from multiple different data sources into a consistent format.

[0012] "Risk assessment" refers to the process of determining the level of flood risk based on the output of the generative AI model and assessing its impact.

[0013] "Visual display" refers to graphically representing the analysis results on a map, allowing users to intuitively understand the risk level of a particular area.

[0014] "Notifying" refers to informing local government disaster prevention officials and residents about flood risk through means such as email, SMS, and app notifications.

[0015] An "evacuation plan" is a plan that provides specific guidelines for residents to evacuate quickly and safely in the event of a high risk of flooding.

[0016] "Feedback" refers to information obtained from local governments and residents regarding actual impacts and response actions, and is information used to improve the accuracy of models and systems. [Brief explanation of the drawings]

[0017] [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

[0018] 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.

[0019] First, the terms used in the following description will be explained.

[0020] 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).

[0021] 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.

[0022] 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.

[0023] 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.

[0024] 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."

[0025] [First embodiment]

[0026] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0027] 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.

[0028] 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).

[0029] 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.

[0030] 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.

[0031] 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.

[0032] 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.

[0033] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0034] 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.

[0035] 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.

[0036] 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.

[0037] 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."

[0038] The present invention is a system that predicts river flood risk with high accuracy and provides risk information to local governments and residents in real time. This system operates in cooperation with a server, terminals, and users.

[0039] Specific Embodiments of the System

[0040] Data collection

[0041] server

[0042] River water level data, precipitation data, and topographical data are obtained from meteorological databases, which are collected in real time through collaboration with meteorological agencies, satellite databases, and on-site observation stations.

[0043] Converting acquired data into a unified format to facilitate subsequent analysis and processing, for example, integrating water level, precipitation, and topographic data from different sources into a consistent format.

[0044] Data analysis

[0045] server

[0046] The data converted into a unified format is then fed into a generative AI model, which uses advanced machine learning algorithms such as deep learning to combine historical and current flood data to predict flood risk.

[0047] Data preprocessing is performed to remove noise from each data and filter out outliers, ensuring high quality of input data for the predictive model.

[0048] Risk Assessment and Notification

[0049] server

[0050] The output of the generative AI model is used to assess the level of flood risk. This quantifies the flood risk for a specific area and displays it according to the risk level. This display is then displayed on a map as a color-coded risk map.

[0051] The system notifies local government disaster prevention officials and residents of risk assessment results in real time, providing information quickly via email, SMS, and a dedicated app.

[0052] Evacuation Planning and Response

[0053] server

[0054] Based on the results of the risk assessment, specific evacuation plans and infrastructure operation plans are generated, which are customized for each local government.

[0055] Devices (smartphones and computers of local government employees and residents)

[0056] Check the received flood risk information and suggested evacuation plans. For example, use a smartphone app to intuitively understand evacuation locations and routes on a map.

[0057] User

[0058] After an evacuation or flood occurs, feedback on actual impacts and actions is sent to the system, collecting data that can be used to improve forecasting models and response plans.

[0059] Specific examples

[0060] Example 1: Real-time flood forecasting and notification

[0061] scene

[0062] In a situation where precipitation is increasing rapidly and river water levels are rising rapidly, the server retrieves the latest data from the weather database.

[0063] server

[0064] Using a generative AI model, the risk of flooding within the next 48 hours is predicted and specific urban areas are identified as being at high risk.

[0065] Device (local government employee's smartphone)

[0066] Receive notifications from the server and check detailed risk maps and evacuation plans.

[0067] User

[0068] Based on risk information, evacuation orders are issued to residents, who then begin evacuating promptly in accordance with the instructions.

[0069] Example 2: Feedback and Improvement

[0070] scene

[0071] It was reported that after the flood occurred, residents were able to evacuate safely and ensure their safety, and the actual impact was low.

[0072] User

[0073] Feedback is sent to the system about the actual impact and evacuation actions.

[0074] server

[0075] Feedback data is collected and analyzed, and reflected in improvements to prediction models and proposed algorithms.

[0076] In this way, the present invention provides a consistent system that covers everything from flood risk prediction to notification and countermeasure proposals, enabling real-time, highly accurate flood response. In particular, the series of processes, including data collection, analysis using generative AI models, risk assessment, notification, and feedback, work together to achieve a rapid and appropriate response.

[0077] The processing flow will be explained below.

[0078] Step 1: Data collection

[0079] server

[0080] It connects to meteorological databases, satellite databases, and local observation stations to obtain the latest river water level, precipitation, and topographical data. Specifically, it uses an API key to send requests to the Meteorological Bureau's services and downloads the data in real time.

[0081] The acquired data is stored in a database and the different data formats from each data source are converted into a unified format, for example, water level data is unified to meters and precipitation data is unified to millimeters.

[0082] Step 2: Data Preprocessing

[0083] server

[0084] The acquired data is subjected to noise removal and outlier filtering. For example, abnormally high or low values ​​are detected and excluded. In addition, if there is missing data, it is supplemented based on past data or surrounding data.

[0085] The data is formatted as time series data or geographic information system (GIS) data, making it suitable for input into generative AI models.

[0086] Step 3: Flood risk forecasting

[0087] server

[0088] The formatted data is then fed into a generative AI model, which uses deep learning algorithms to predict flood risk by combining historical flood data with current weather data to calculate the probability of a flood occurring within the next 48 hours.

[0089] The forecast results are quantified and the flood risk level is evaluated. For example, the flood risk level is displayed on a three-level scale: "low," "medium," or "high."

[0090] Step 4: Risk assessment and visualization

[0091] server

[0092] The predicted flood risk level is visually displayed on a map, with high-risk areas color-coded to provide an intuitive understanding of the level of risk to a particular area.

[0093] Make a list of potentially affected facilities (hospitals, schools, evacuation centers, etc.) and identify where action is needed.

[0094] Step 5: Notification

[0095] server

[0096] Notify local government disaster prevention officials and residents of flood risk information by sending emails and SMS containing the risk level of areas requiring action and recommended action plans (evacuation locations and evacuation routes).

[0097] Devices (smartphones and computers of local government employees and residents)

[0098] Residents can check the received notifications and understand the specific measures to take. Residents can use the smartphone app to check real-time risk information and evacuation plans.

[0099] Step 6: Implement your evacuation plan

[0100] User

[0101] Evacuations are carried out promptly based on the notified flood risk information and proposed evacuation plans. Local government officials issue evacuation orders, prepare evacuation shelters, and direct traffic. Residents also follow the evacuation routes provided to them and evacuate to safe locations.

[0102] Step 7: Gather feedback

[0103] User

[0104] After a flood occurs, feedback is provided to the system regarding the actual impact and evacuation behavior, including the time required for evacuation, the effectiveness of the evacuation plan, and the extent of the damage.

[0105] server

[0106] The collected feedback data will be analyzed to improve forecasting models and evacuation plans, allowing for more accurate and rapid responses the next time floods occur.

[0107] Example 1

[0108] 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."

[0109] To predict river flood risks in real time with high accuracy based on meteorological data, promptly and appropriately notify local government disaster prevention officials and residents of risk information, and provide appropriate evacuation plans.

[0110] 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.

[0111] In this invention, the server includes means for acquiring river water level data, precipitation data, and topographical data from a weather database, means for converting the acquired data into a unified format, means for preprocessing the converted data by removing noise and filtering outliers, means for inputting the preprocessed data into a generative AI model to predict flood risk, means for conducting a risk assessment based on the predicted flood risk and visually displaying the risk level on a map, means for notifying disaster prevention personnel in the local government and residents of the results of the risk assessment, means for proposing evacuation plans based on the notified risk assessment results, and means for collecting feedback from each local government and residents and using it to improve the accuracy of the model. This enables highly accurate real-time prediction of river flood risk and the provision of prompt and appropriate risk notifications and evacuation plans.

[0112] A "weather database" is a database for storing and managing weather information, and provides data on weather such as temperature, precipitation, wind speed, wind direction, and atmospheric pressure.

[0113] "River water level data" refers to data that indicates the water level at a specific point in a river, and is used to predict river conditions and flood risk.

[0114] "Precipitation data" is data that numerically represents the amount of precipitation, such as rain or snow, that fell at a specific location within a certain period of time.

[0115] "Topographic data" refers to data that expresses geographical features and the shape of the terrain as numerical values ​​or images, and includes information such as the height of the earth's surface, undulations, and land use.

[0116] "Unified format" means converting data obtained from different sources into a consistent format and making it conform to a common standard.

[0117] "Noise removal" is a process for removing unnecessary information and outliers in data analysis and processing to improve data accuracy.

[0118] "Outlier filtering" is a process of detecting, removing, or correcting values ​​that are outside the normal range or that are unreasonable from a business perspective, contained in a data set.

[0119] "Preprocessing" refers to the preparatory steps taken to prepare data for analysis, including noise removal and filtering of outliers.

[0120] A "generative AI model" is an algorithmic model that uses machine learning and deep learning to make predictions and classifications from data.

[0121] "Risk assessment" is the evaluation and judgment of the degree and impact of predicted risks in numerical and visual form.

[0122] A "risk level" is an indicator that shows the extent to which a particular risk exists, and is usually expressed as a number or color.

[0123] "Notifying" refers to the act of transmitting specific information to a designated recipient, and can be done via email, SMS, a dedicated app, etc.

[0124] An "evacuation plan" is a plan that outlines specific procedures and routes for evacuation in the event of a danger.

[0125] "Feedback" refers to information that collects evaluations and comments regarding the operation and results of the system and is used to help improve it in the future.

[0126] The present invention is implemented as a system that predicts river flood risk with high accuracy and provides risk information to local governments and residents in real time. This system operates in cooperation with a server, terminals, and users.

[0127] Data collection

[0128] server

[0129] The server uses APIs to obtain river water level data, precipitation data, and topographical data in real time from meteorological databases, satellite databases, and local observation stations. Specifically, the server calls the meteorological database API to obtain the latest precipitation data.

[0130] Data integration

[0131] server

[0132] It converts acquired data of different formats into a unified format, generating a consistent dataset that allows subsequent processing to be performed efficiently and effectively.

[0133] Data Preprocessing

[0134] server

[0135] The server then removes noise and filters out outliers from the data converted into a unified format. Specifically, it complements missing values, normalizes the data, and detects and removes abnormally high water level data.

[0136] Flood risk prediction

[0137] server

[0138] The preprocessed data is then fed into a generative AI model, which uses advanced machine learning algorithms such as deep learning to combine historical and current flood data to predict flood risk. For example, future rainfall and current water level data are fed into the model to predict the risk of flooding within 48 hours.

[0139] Risk Assessment and Notification

[0140] server

[0141] The output of the generated AI model is analyzed to assess the level of flood risk, which then calculates a risk index for a specific area and generates a risk map that displays different colors on a map according to the risk level.

[0142] Local government disaster prevention officials and residents will be notified of the risk assessment results in real time via email, SMS, and a dedicated app, along with a risk map.

[0143] Generate evacuation plans

[0144] server

[0145] Based on the flood risk assessment results, the server generates specific evacuation plans customized for each municipality, for example, providing residents in high-risk areas with a list of evacuation sites and optimal evacuation routes.

[0146] Real-time notifications and displays

[0147] Terminal

[0148] The devices of local government officials and residents receive the risk information and evacuation plans sent from the server. For example, the smartphones of local government officials receive emergency notifications and display detailed risk maps and evacuation routes on a dedicated app.

[0149] Collecting and analyzing feedback

[0150] User

[0151] Users provide feedback to the system about the actual impacts and actions they take after a flood, for example, residents reporting on the congestion of evacuation shelters and the actual flooding situation.

[0152] server

[0153] The server analyzes the collected feedback data and uses it to improve the prediction model and response plan. For example, based on feedback that an evacuation site was overcrowded, the server can improve the evacuation plan to suggest a different evacuation site for the next time.

[0154] Prompt Sentence Examples

[0155] "Collect real-time river water level and precipitation data, use generative AI models to predict flood risk, communicate risk assessment results to city officials and residents, and develop evacuation plans."

[0156] The system covers everything from data collection to risk prediction, notification, evacuation plan generation, and feedback collection, enabling highly accurate prediction of river flood risks and supporting rapid response in real time.

[0157] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0158] Step 1: Data collection

[0159] server

[0160] Input: weather database, satellite database, data requests from local observation stations

[0161] The server uses an API to obtain river water level data, precipitation data, and topographical data in real time from meteorological databases, satellite databases, and local observation stations.

[0162] Output: Acquired data (river water level data, precipitation data, topographical data)

[0163] Step 2: Integrate the data

[0164] server

[0165] Input: Acquired river water level data, precipitation data, topographical data

[0166] The server converts the acquired data into a unified format and generates a consistent data set. The server converts and maps the data to harmonize the information obtained from different data sources.

[0167] Output: Uniform format dataset

[0168] Step 3: Preprocessing the data

[0169] server

[0170] Input: Uniform format dataset

[0171] The server then performs noise removal and outlier filtering on the data converted into a unified format, for example, by filling in missing values, normalizing the data, and detecting and excluding abnormally high water level data.

[0172] Output: Preprocessed data

[0173] Step 4: Predict flood risk

[0174] server

[0175] Input: Preprocessed data

[0176] The preprocessed data is input into a generative AI model to predict flood risk. The generative AI model uses advanced machine learning algorithms such as deep learning to combine past flood data with current data to calculate future flood risk. For example, it can predict the likelihood of a flood occurring based on rainfall data within the next 48 hours and current water level data.

[0177] Output: Predicted flood risk data

[0178] Step 5: Risk assessment and notification

[0179] server

[0180] Input: Predicted flood risk data

[0181] The output of the generated AI model is analyzed to assess the level of flood risk. The server calculates a risk index for a specific area and generates a color-coded risk map based on the risk level. The server also sends real-time risk information via email, SMS, or a dedicated app to notify local government disaster prevention officials and residents of the results of the risk assessment.

[0182] Output: Risk assessment report and risk map, notification message

[0183] Step 6: Generate an evacuation plan

[0184] server

[0185] Input: Risk Assessment Report

[0186] Based on the flood risk assessment results, the server generates specific evacuation plans customized for each local government. For example, it provides a list of evacuation sites and optimal evacuation routes for residents in high-risk areas. The evacuation plans are displayed in conjunction with risk maps.

[0187] Output: Evacuation plan

[0188] Step 7: Real-time notifications and displays

[0189] Terminal

[0190] Input: Risk assessment report, risk map, draft evacuation plan

[0191] The devices of local government officials and residents receive the risk information and evacuation plans sent from the server. For example, the smartphones of local government officials receive emergency notifications and intuitively display detailed risk maps and evacuation routes on a dedicated app.

[0192] Output: Risk information and evacuation plan displayed on the terminal

[0193] Step 8: Collect and analyze feedback

[0194] User

[0195] Input: User feedback on the actual evacuation situation and impact

[0196] Users provide feedback to the system about the actual impacts and actions they take after a flood, for example reporting on the congestion of evacuation shelters and the actual flooding situation.

[0197] Output: Feedback data sent

[0198] server

[0199] Input: Collected feedback data

[0200] The server collects and analyzes the feedback data and uses it to improve the prediction model and response plan. For example, based on feedback that an evacuation site was overcrowded, it will suggest a different evacuation site for next time.

[0201] Output: Improved forecast model and evacuation plan

[0202] In this way, this system predicts river flood risks with high accuracy and in real time through step-by-step processing, providing appropriate information and supporting evacuation planning.

[0203] (Application example 1)

[0204] 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."

[0205] In recent years, climate change has led to an increase in extreme weather events, increasing the risk of river flooding. However, existing systems have difficulty accurately predicting flood risk in real time and responding quickly. Furthermore, evacuation plans and evacuation route suggestions are not provided properly, which can lead to problems in ensuring the safety of residents. Additionally, there is a lack of a feedback function to continuously improve the prediction model based on collected data.

[0206] 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.

[0207] In this invention, the server includes: means for acquiring river water level data, precipitation data, and topographical data from a weather database; means for converting the acquired data into a unified format; and means for inputting the converted data into a generative AI model to predict flood risk. This enables real-time, highly accurate prediction of flood risk and prompt response. The server also includes means for performing a risk assessment based on the predicted flood risk and visually displaying the risk level on a map; and means for notifying local government disaster prevention personnel and residents of the risk assessment results. This enables risk information to be provided promptly to residents and local governments. The server also includes means for proposing an evacuation plan and an optimal evacuation route based on the notified risk assessment results, and means for displaying evacuation sites in a specific area based on the user's location information. The server also includes means for collecting feedback from local governments and residents and using it to improve the accuracy of the model, thereby ensuring the safety of residents and continuously improving the prediction model.

[0208] "Weather database" refers to a database that collects, stores, and provides weather information, including river water level data, precipitation data, and weather forecast data.

[0209] "River water level data" refers to data that indicates the height of the water level in a river, and is information necessary for assessing flood risk.

[0210] "Precipitation data" refers to data that indicates the amount of precipitation over a certain period of time in a specific area, and is an important factor in predicting the risk of flooding.

[0211] "Topographical data" refers to data containing information about the topography of an area, which is useful for planning the extent of flood impacts and evacuation routes.

[0212] A "unified format" is a consistent data format for converting data obtained from different data sources into a format that is easy to analyze.

[0213] A "generative AI model" is an artificial intelligence model used to predict flood risk based on historical and current data, and specifically includes deep learning models.

[0214] "Risk assessment" is the process of quantifying and assessing the flood risk of a specific area based on the output of a generative AI model.

[0215] "Visually displaying" means presenting the assessed risk to the user in a visual format such as a map or graph.

[0216] "Notifying" means informing users of the assessed flood risk and related information via email, SMS, smartphone apps, etc.

[0217] An "evacuation plan" is a set of instructions and route guidance designed to enable residents to evacuate safely when the risk of flooding increases.

[0218] The "optimal evacuation route" is a route that suggests the safest and quickest evacuation route based on the user's current location.

[0219] "Feedback" refers to residents and local governments sending information back to the system about actual evacuation behavior and the impact of flooding, which is used to improve the accuracy of the model.

[0220] The system proposed in this invention aims to predict river flood risk in real time and provide prompt and appropriate information to residents and local governments. To achieve this, it is necessary for the server, terminals, and users to work together.

[0221] server

[0222] Data collection

[0223] The server obtains river water level data, precipitation data, and topographical data from a meteorological database, which uses a database that collects, stores, and provides meteorological information.

[0224] Converting acquired data into a unified format: Data obtained from different data sources is converted into a unified data format to make it easier to analyze.

[0225] Data analysis

[0226] The data converted into a unified format is then input into a generative AI model, an artificial intelligence model used to predict flood risk based on past and current data. A deep learning model is typically used.

[0227] Generative AI models are implemented using advanced machine learning frameworks such as TensorFlow and Keras, which denoise data and filter outliers to generate high-quality input data.

[0228] Risk Assessment and Notification

[0229] The level of flood risk is assessed based on the output of the generative AI model. The risk assessment quantifies the flood risk of a specific area and displays it according to the risk level.

[0230] The results of the risk assessment are visually displayed on a map and are notified in real time to local government disaster prevention officials and residents via email, SMS, and a dedicated smartphone app.

[0231] Evacuation Planning and Response

[0232] Based on the risk assessment results, the system proposes specific evacuation plans and optimal evacuation routes. Based on the user's location information, it displays evacuation sites in specific areas.

[0233] Evacuation plans can be customized to suit the needs of each local government.

[0234] Terminal

[0235] The terminal (smartphone or PC) functions as a device for checking the received flood risk information and the proposed evacuation plan. For example, using a smartphone app, users can intuitively check evacuation locations and evacuation routes on a map.

[0236] User

[0237] Users act quickly based on the risk information provided. After an evacuation or flood occurs, they provide feedback to the system about the actual impact and actions taken. This feedback gathers data that can be used to improve predictive models and response plans.

[0238] Specific examples

[0239] Real-time flood forecasting and notifications

[0240] When residents press the "Check current flood risk" button, the latest flood risk map is retrieved from the server and displayed on the screen. When residents select "Show nearest evacuation site," the smartphone app identifies the user's current location and displays the optimal evacuation site and route on the map.

[0241] Prompt Sentence Examples

[0242] 1. "Check your current flood risk"

[0243] 2. "Display the nearest evacuation site"

[0244] Through these functions, this system aims to ensure the safety of residents and provide rapid evacuation support, while also improving the accuracy of the model.

[0245] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0246] Step 1:

[0247] The server retrieves river water level data, precipitation data, and topography data from the weather database. This includes specific operations such as collecting data in real time using APIs. The input data includes information on water level, precipitation, and topography, and the output is raw data retrieved in bulk.

[0248] Step 2:

[0249] The server converts the acquired data into a unified format. Specifically, it converts data in different formats into a standardized format and processes it to make it easier to analyze. The input data is raw data, and the output data is data converted into a unified format. This process includes data normalization and unit unification.

[0250] Step 3:

[0251] The server inputs the converted data into a unified format into a generative AI model to predict flood risk. The generative AI model is implemented using TensorFlow and Keras, and predicts risk based on past and current data. The input data is standardized data, and the output data is the predicted flood risk level. This step also removes noise from the data and filters out outliers.

[0252] Step 4:

[0253] The server performs a risk assessment based on the output of the generative AI model, quantifying and assessing the flood risk of a specific area. The resulting risk level is visually displayed on a map. The input data is the predicted flood risk level, and the output data is a visually displayed risk map. Specifically, the risk level is displayed color-coded on the map application.

[0254] Step 5:

[0255] The server notifies local government disaster prevention personnel and residents of the risk assessment results. Specifically, notifications are sent via email, SMS, and a dedicated smartphone app. The input data is the risk assessment results, and the output data is the notification message. This process also includes prioritizing notifications and selecting recipients.

[0256] Step 6:

[0257] The server proposes an evacuation plan and optimal evacuation route based on the notified risk assessment results. It displays evacuation locations in a specific area based on the user's location information. The input data are the risk assessment results and the user's location information, and the output data are the evacuation plan and evacuation route. Specifically, it displays the optimal evacuation route on a map application.

[0258] Step 7:

[0259] The device checks the received flood risk information and proposed evacuation plan and presents it to the user. The user uses a smartphone app to intuitively check evacuation locations and routes on a map. The input data is the notified risk information and evacuation plan, and the output data is the displayed evacuation map.

[0260] Step 8:

[0261] Users send feedback to the system about the actual impacts and actions taken after an evacuation or flood occurs. The input data is feedback information about evacuation actions and impacts, and the output data is the data stored on the server that received it. In this step, the prediction model and response plan are improved based on the collected feedback.

[0262] In this way, each step works in coordination to predict river flood risks with high accuracy and enable appropriate responses in real time.

[0263] 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.

[0264] This invention combines a system that obtains river water level data, precipitation data, and topographical data from meteorological and satellite databases, and uses this data to predict flood risk with high accuracy, with an emotion engine that recognizes the user's emotions. This system operates in cooperation with a server, terminals, and users.

[0265] Specific Embodiments of the System

[0266] Data collection and analysis

[0267] server

[0268] Obtain the latest river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases. This includes obtaining data from the Meteorological Agency using an API key and analyzing image data from the satellite database.

[0269] The acquired data is converted into a unified format, missing data is imputed, and noise is filtered out, and the converted data is used as input for generative AI models.

[0270] Flood risk forecasting

[0271] server

[0272] After the pre-processing process is complete, the data is fed into a generative AI model, which uses deep learning algorithms to predict flood risk. The model takes into account historical flood data, current water levels, precipitation, and topographical data to calculate the risk of flooding within the next 48 hours.

[0273] The prediction results are quantified according to the risk level, and a risk assessment is carried out based on this, with the risk level of a specific area being visually displayed on a map using different colors.

[0274] Risk notification and evacuation planning

[0275] server

[0276] Flood risk information is sent to local government disaster prevention officials and residents via email, SMS, and a dedicated app, and includes specific risk levels and evacuation instructions.

[0277] Devices (smartphones and computers of local government employees and residents)

[0278] Residents can receive the information and check the risk level. Detailed evacuation plans and evacuation routes are also provided at the same time, allowing them to take action quickly.

[0279] Application of Emotion Engine

[0280] server

[0281] The emotion engine analyzes the user's emotional state when receiving and responding to notifications. For example, it analyzes emails, in-app reaction data, and text data obtained from feedback forms to recognize the user's emotions in real time.

[0282] The notification language and content are adjusted based on the user's emotional state. Specifically, if a user is in a panic despite the urgency of the situation, the notification language will be changed to one that helps the user remain calm.

[0283] Feedback and Improvements

[0284] User

[0285] After the evacuation and flood events, provide feedback on emotional states and actual evacuation behavior, for example, recording fear and difficulties felt during evacuation and the effectiveness of evacuation routes.

[0286] server

[0287] The collected feedback data will be analyzed to improve forecasting models and evacuation plans, and sentiment data will also be analyzed to improve the quality of notifications and plans for the next disaster response.

[0288] Specific examples

[0289] Example 1: Flood risk prediction and emotional response

[0290] scene

[0291] A heavy rain warning is issued, causing river water levels to rise rapidly. The server retrieves the latest water level and precipitation data from the weather database and uses a deep learning model to predict flood risk.

[0292] server

[0293] Determine whether a particular area is at high risk and assess the risk level.

[0294] The emotion engine analyzed that some users had panicked during previous evacuations, and this time the notification sent a message encouraging those users to remain calm and act quickly.

[0295] Device (resident's smartphone)

[0296] Residents receive notifications, reassuring instructions along with evacuation routes, and take action quickly.

[0297] Example 2: Feedback collection and system improvement

[0298] scene

[0299] Flooding occurs and residents are evacuated. After evacuation, residents provide feedback on their feelings and the evacuation situation.

[0300] User

[0301] In the feedback form, participants provided specific feedback such as, "I panicked during the evacuation, but the notification message was very helpful."

[0302] server

[0303] The feedback and sentiment data will be analyzed and reflected in response measures for the next flood, and the sentiment engine algorithm will be improved to provide more effective risk notifications and evacuation plans.

[0304] In this way, the present invention is a system that combines highly accurate flood risk prediction with user emotional responses, encouraging prompt and appropriate evacuation behavior and ensuring safety. In particular, the introduction of an emotion engine enables flexible responses based on user reactions, further enhancing the effectiveness of disaster prevention measures.

[0305] The processing flow will be explained below.

[0306] Step 1: Data collection

[0307] server

[0308] Connect to meteorological and satellite databases to get real-time river water level data, precipitation data, and topographic data. Use an API key to request data from the meteorological bureau and download the latest data.

[0309] The acquired data is stored in an internal database and data from different data sources is converted into a unified format, for example, water level data is standardized to meters and precipitation data is converted to millimeters.

[0310] Step 2: Data Preprocessing

[0311] server

[0312] The acquired data is subjected to noise removal and outlier filtering. For example, abnormally high or low measurement values ​​are detected and excluded. In addition, if there is missing data, it is supplemented based on past data or surrounding data.

[0313] The data is formatted as time series data so that it can be input into a generative AI model.

[0314] Step 3: Flood risk forecasting

[0315] server

[0316] The formatted data is then fed into a generative AI model, which uses deep learning algorithms to predict flood risk. The generative AI model calculates the probability of flooding within the next 48 hours based on historical flood data, current water levels, precipitation, and topographical data.

[0317] The forecast results are quantified and the flood risk level for each area is evaluated. For example, flood risk levels are displayed on a three-level scale: "low," "medium," and "high."

[0318] Step 4: Risk assessment and visualization

[0319] server

[0320] The predicted flood risk level is visually displayed on a map, with high-risk areas color-coded to provide an intuitive understanding of the level of risk to a particular area.

[0321] Make a list of potentially affected facilities (hospitals, schools, evacuation centers, etc.) and identify where action is needed.

[0322] Step 5: Risk notification and sentiment analysis

[0323] server

[0324] Flood risk information is sent to local government disaster prevention officials and residents via email, SMS, and a dedicated app, and includes specific risk levels and evacuation instructions.

[0325] The emotion engine collects user reaction data when receiving notifications and analyzes their emotional state. For example, it recognizes emotions based on the user's immediate reaction when opening a notification or text input from a feedback form.

[0326] Step 6: Evacuation planning and emotional response

[0327] server

[0328] The content and presentation of notification messages are adjusted based on the user's emotional state analyzed by the emotion engine. For example, if a user is in a panic despite the urgency of the situation, a message encouraging them to remain calm will be sent.

[0329] Devices (smartphones and computers of local government employees and residents)

[0330] Along with the risk information provided, residents can confirm appropriate evacuation plans and routes. For example, residents can visually check specific evacuation locations and routes via a smartphone app.

[0331] Step 7: Evacuation implementation and feedback

[0332] User

[0333] Evacuate promptly based on the notified flood risk information and the proposed evacuation plan. After completing the evacuation, provide feedback on the emotions and evacuation behavior during the evacuation. For example, enter information such as "the fear and difficulty felt during the evacuation" and "the effectiveness of the evacuation route" in the feedback form.

[0334] server

[0335] The collected feedback data will be analyzed to improve prediction models and evacuation plans. In particular, analyzing emotion data at the same time will improve the quality of notifications and plans for the next disaster response.

[0336] As a result, this invention provides a system that integrates highly accurate flood risk prediction with flexible responses based on user emotions, encouraging prompt and appropriate evacuation behavior. The introduction of an emotion engine makes it possible to realize notifications and evacuation plans based on user reactions, further enhancing the effectiveness of disaster prevention measures.

[0337] Example 2

[0338] 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."

[0339] Current flood risk prediction systems often lack real-time capabilities and accuracy, resulting in delays in evacuation notifications and evacuation plan proposals to residents. These systems also lack the ability to respond to users' emotional states, which can lead to panic among residents in emergencies. As a result, appropriate evacuation behavior is not possible, and safety is not ensured. Furthermore, there is also the issue that feedback is not fully utilized, resulting in the time required to improve the models.

[0340] 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.

[0341] In this invention, the server includes a means for acquiring river water level data, precipitation data, and topographical data from a meteorological database and a satellite database; a means for converting the data into a unified format, filling in missing data, and performing noise filtering; and a means for inputting the converted data into a generative AI model to predict flood risk. This enables real-time and highly accurate prediction of flood risk. Furthermore, risk assessments are performed and displayed visually on a map, and an emotion recognition engine is used to analyze the user's emotional state and adjust notification content to encourage appropriate evacuation behavior. Feedback data can be collected and used to improve the accuracy of the model, further enhancing the effectiveness of future disaster responses.

[0342] A "weather database" is a data storage system for collecting, storing, and providing weather information. It is primarily managed and operated by meteorological agencies and related organizations.

[0343] A "satellite database" is a system that stores data collected from artificial satellites for the purpose of Earth observation. It provides a variety of information, including topographical and meteorological data.

[0344] "River water level data" is numerical information measuring the water surface height of a specific river. It is an important element in flood prediction.

[0345] "Precipitation data" is data that numerically indicates the amount of precipitation (rain, snow, fog, etc.) that fell in a specific area within a certain period of time. It is used for weather forecasting and flood prediction.

[0346] "Terrain data" is data that contains information about the shape, height, and structure of the Earth's surface. It is primarily used for mapping and environmental modeling.

[0347] "Missing data" refers to the absence of required data in a dataset. Also known as missing values.

[0348] "Noise filtering" is a process to remove unnecessary information (noise) from data. It is performed to improve the quality of data.

[0349] A "generative AI model" is a computational model designed to generate new data or information using artificial intelligence, often involving deep learning algorithms.

[0350] "Flood risk" refers to the possibility of flooding occurring in a particular area and the extent of its impact. It is used in risk assessment.

[0351] An "emotion recognition engine" is software that analyzes a user's emotional state from input such as text data and voice data, enabling the system to respond based on the user's reaction.

[0352] An "evacuation plan" is a plan that shows routes and methods for safe evacuation in the event of a disaster. It is provided to ensure the safety of residents.

[0353] "Feedback" refers to opinions, impressions, and evaluation information provided by system users. It is used to improve the system.

[0354] A "municipal government" is a local government or related institution that administers a particular area and ensures the welfare and safety of its residents.

[0355] "Residents" refers to people who live in a specific area. They are often users of the system.

[0356] A "deep learning model" is a machine learning model consisting of a multi-layered neural network. It is used for advanced data analysis and prediction.

[0357] The present invention is a system that predicts flood risk using data acquired from a meteorological database and a satellite database, and provides notifications that correspond to the user's emotional state. A specific embodiment of this system is described below.

[0358] Data collection and analysis

[0359] server

[0360] The server obtains the latest river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases. APIs are used to obtain data from the meteorological bureau and satellite databases. For example, API requests are sent using the Python requests library. The obtained data is stored in a database management system such as PostgreSQL or MySQL.

[0361] Data Preprocessing

[0362] server

[0363] The server converts the acquired data into a unified format, imputes missing data, and performs noise filtering. Python's pandas library is used to unify different data formats and impute missing values. Statistical methods such as median are used to improve data quality during this process. After preprocessing, the data is formatted as input for the generative AI model and converted into NumPy arrays or TensorFlow tensors.

[0364] Flood risk forecasting

[0365] server

[0366] The server inputs the preprocessed data into the generative AI model. A deep learning model is built using TensorFlow and PyTorch, and the data is used for prediction. The model takes into account past flood data, current water levels, precipitation, and topographical data to predict flood risk within the next 48 hours. The prediction results are quantified based on risk levels, and the risk level for specific areas is visually displayed on a map. Map libraries such as Google Maps API and Leaflet are used to display the map.

[0367] Risk notification and evacuation planning

[0368] server

[0369] The server then notifies local government disaster prevention officials and residents of flood risk information via email, SMS, or a dedicated app. For example, AWS SNS (Simple Notification Service) can be used to send SMS notifications.

[0370] Terminal

[0371] The device receives the information and checks the risk level. A push notification is sent to the smartphone app, and by tapping the notification, detailed information is displayed. Evacuation plans and evacuation routes are also presented, and the optimal evacuation route can be displayed using the Google Maps API.

[0372] Emotion recognition and notification adjustment

[0373] server

[0374] The server uses an emotion engine to analyze the emotional state of users receiving notifications in real time. For example, it uses the Natural Language Toolkit (NLTK) and BERT models to analyze notification emails and app response data. Based on the analysis results, it adjusts the wording and content of notifications. For example, it sends an encouraging message to a panicked user to help them stay calm.

[0375] Feedback and System Improvement

[0376] User

[0377] Users provide feedback after an evacuation or flood occurs. For example, they can write in the app's feedback form that they felt scared during the evacuation but found the notifications helpful.

[0378] server

[0379] The server analyzes the collected feedback data to help improve the predictive models and evacuation plans. It uses data mining techniques and sentiment analysis algorithms to analyze the feedback, which will enable it to provide more effective risk notifications and evacuation plans in the next disaster response.

[0380] Specific examples

[0381] The following is a specific example of how this system works.

[0382] Scene 1: Processing in the event of a heavy rain warning

[0383] The server uses an API to retrieve the latest precipitation and river water level data from the weather database.

[0384] Convert the acquired data into a unified format and fill in any missing data.

[0385] The preprocessed data is input into a generative AI model to predict flood risk.

[0386] High-risk areas are identified and marked in red on the map.

[0387] Scene 2: Notifications and Emotion Recognition

[0388] The server sends an SMS to users who live in high-risk areas, for example, "Your area is at high risk of flooding. Please begin evacuation."

[0389] The user views the SMS on their smartphone and checks the evacuation route.

[0390] The server uses an emotion engine to analyze the user's reaction and detect whether they are in a panic state.

[0391] The server sends a calming message to panicked users, for example, "Remain calm and evacuate. You are safe."

[0392] Based on these examples, the system can ensure the safety of residents by providing highly accurate predictions of flood risk and prompt evacuation notifications that respond to users' emotions.

[0393] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0394] Step 1: Data collection

[0395] server

[0396] The server retrieves river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases, and sends API requests using the Python requests library.

[0397] Input: API request (weather database and satellite database)

[0398] Output: Acquired river water level data, precipitation data, topographical data

[0399] The acquired data is stored in a database management system such as PostgreSQL or MySQL.

[0400] Step 2: Data Preprocessing

[0401] server

[0402] The data retrieved by the server is converted into a unified format. The data is cleaned and formatted consistently using the Python pandas library.

[0403] The server imputes missing data and performs noise filtering, for example by imputing missing values ​​with the median to ensure data consistency.

[0404] Input: Acquired river water level data, precipitation data, topographical data

[0405] Output: Preprocessed data (unified format, missing data imputation, noise filtering)

[0406] The preprocessed data is converted into NumPy arrays or TensorFlow tensors.

[0407] Step 3: Flood risk forecasting

[0408] server

[0409] The server inputs the preprocessed data into a generative AI model, which uses a deep learning model built using TensorFlow or PyTorch.

[0410] The model takes into account historical flood data, current water levels, rainfall and topographical data to predict flood risk within the next 48 hours.

[0411] Input: Preprocessed data (NumPy arrays or TensorFlow tensors)

[0412] Output: Flood risk rating for each region (ranging from 0 to 1)

[0413] High-risk areas are color-coded on a map, and the map is displayed using Google Maps API and Leaflet.

[0414] Step 4: Risk notification

[0415] server

[0416] The server notifies local government disaster prevention officials and residents of flood risk information via email, SMS, and a dedicated app.

[0417] Send SMS notifications using AWS SNS (Simple Notification Service).

[0418] Input: Flood risk assessment results, notification messages based on risk level

[0419] Output: Notifications (email, SMS) sent to local government disaster prevention officials and residents

[0420] Terminal

[0421] The device receives the information and checks the risk level. A push notification is sent to the smartphone app, displaying detailed information.

[0422] The device will present the user with an evacuation plan and evacuation route, using the Google Maps API to display safe evacuation routes.

[0423] Input: Risk notification message, evacuation plan information

[0424] Output: Notifications and evacuation routes displayed to the user

[0425] Step 5: Emotion recognition and notification adjustment

[0426] server

[0427] The server uses an emotion engine to analyze the emotional state of the user receiving the notification information in real time. It analyzes the text data using NLTK (Natural Language Toolkit) and the BERT model.

[0428] Tailor the wording and content of notifications based on the user's emotional state: if a user is panicking, send them a notification that helps them stay calm.

[0429] Input: User emotion data (text, reaction data)

[0430] Output: Adjusted notification message

[0431] Example: A message encouraging people to remain calm and evacuate.

[0432] Step 6: Gather feedback and improve the system

[0433] User

[0434] After an evacuation or flood occurs, users provide feedback on their emotional state and actual evacuation behavior. For example, they can write in the in-app feedback form, "I was able to make appropriate decisions quickly when evacuating."

[0435] Input: Feedback form input (text data)

[0436] Output: Feedback data provided

[0437] server

[0438] The server analyzes the collected feedback data to help improve prediction models and evacuation plans. It uses data mining techniques and sentiment analysis algorithms to analyze the feedback.

[0439] Input: Feedback data (emotional information, behavioral records)

[0440] Output: Improved models and evacuation plans

[0441] Improved emotion engine algorithms will provide more effective risk notifications and evacuation plans for the next flood.

[0442] This allows the system to achieve highly accurate flood risk prediction and user response at each step, promoting safe evacuation behavior.

[0443] (Application example 2)

[0444] 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."

[0445] Conventional flood risk prediction systems primarily use meteorological and satellite data to predict flood risks and notify local governments and residents. However, they do not take into account the emotional state of passengers, particularly in autonomous vehicles, which can lead to confusion and anxiety caused by emotions such as panic. Furthermore, they lack the ability to automatically avoid areas with a high risk of flooding, which can make it difficult to fully ensure passenger safety. A system that can effectively solve these issues is needed.

[0446] 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.

[0447] In this invention, the server includes means for acquiring river water level data, precipitation data, and topographical data from a meteorological database and a satellite database, means for converting the acquired data into a unified format, and means for inputting the converted data into a generative AI model to predict flood risk, thereby enabling highly accurate flood risk prediction and risk assessment.

[0448] The server also includes a means for performing a risk assessment based on the predicted flood risk and visually displaying the risk level on a map, a means for notifying local government disaster prevention personnel and residents of the risk assessment results, a means for proposing an evacuation plan based on the notified risk assessment results, a means for collecting feedback from each local government and residents and using it to improve the accuracy of the model, a means for recognizing the emotional state of passengers, a means for customizing the content of the notification based on the emotional state, and a means for the autonomous vehicle to automatically select a route that avoids the flood risk. This enables flexible notifications according to the emotional state of passengers and enables the autonomous vehicle to select a safe route.

[0449] A "weather database" is a database that includes various weather data (e.g., precipitation data, temperature data, wind speed data, etc.).

[0450] A "satellite database" is a database that includes topographical data and water level data obtained from artificial satellites.

[0451] "River water level data" is data that indicates information on the water level in a specific river basin.

[0452] "Precipitation data" is data that indicates information about the amount of precipitation in a specific area.

[0453] "Topography data" refers to data that indicates information such as the topography and altitude of a specific area.

[0454] A "unified format" is a format for converting data obtained from different data sources into a consistent format.

[0455] A "generative AI model" is a machine learning model that makes predictions and classifications based on data.

[0456] "Flood risk" is an indicator of the likelihood of flooding occurring in a particular area within a certain period of time.

[0457] "Risk assessment" is the process of determining danger based on acquired data and indicating the risk level using numbers, colors, etc.

[0458] A "local government disaster prevention officer" is an official responsible for disaster prevention measures in a local government.

[0459] "Residents" refers to people who live in a particular area.

[0460] An "evacuation plan" is a plan for safely evacuating in the event of a disaster such as a flood.

[0461] "Feedback" refers to information that collects opinions and impressions from users and is used to improve the system.

[0462] "Passenger emotional state" is an index that indicates the emotional state that passengers are feeling (e.g., relief, anxiety, panic, etc.).

[0463] "Customizing notification content" means changing the content of a signal or message depending on the emotional state of the recipient.

[0464] An "autonomously selected route" is a safe route that an autonomous vehicle selects to avoid hazards such as flooding.

[0465] This invention is a system for predicting flood risks and safely operating autonomous vehicles. This system obtains river water level data, precipitation data, and topographical data from meteorological and satellite databases, and predicts flood risks using a generative AI model based on this data. Furthermore, the system is characterized by its ability to recognize passenger emotions and customize notification content.

[0466] Data collection and analysis

[0467] server

[0468] Obtain the latest river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases. This includes obtaining data from the Meteorological Agency using an API key and analyzing image data from the satellite database.

[0469] The acquired data is converted into a unified format, missing data is imputed, and noise is filtered out, and the converted data is used as input for generative AI models.

[0470] Flood risk forecasting and risk assessment

[0471] server

[0472] After the pre-processing process is complete, the data is fed into a generative AI model, which uses deep learning algorithms to predict flood risk. The model takes into account historical flood data, current water levels, precipitation, and topographical data to calculate the risk of flooding within the next 48 hours.

[0473] The prediction results are quantified according to the risk level, and a risk assessment is carried out based on this, with the risk level of a specific area being visually displayed on a map using different colors.

[0474] Risk notification and evacuation planning

[0475] server

[0476] Flood risk information is sent to local government disaster prevention officials and residents via email, SMS, and a dedicated app, and includes specific risk levels and evacuation instructions.

[0477] Devices (smartphones and personal computers of local government employees and residents)

[0478] Residents can receive the information and check the risk level. Detailed evacuation plans and evacuation routes are also provided at the same time, allowing them to take action quickly.

[0479] Route selection for autonomous vehicles

[0480] server

[0481] This system suggests safe routes for autonomous vehicles to avoid areas at high risk of flooding, allowing them to operate on routes with lower flood risk.

[0482] Terminal (on-board system for autonomous vehicles)

[0483] The in-vehicle system automatically selects the route for the autonomous vehicle based on the proposed route, and the route information is communicated to passengers via the in-vehicle display and audio system.

[0484] Application of Emotion Engine

[0485] server

[0486] Using an emotion engine, the system analyzes passengers' facial expressions using an in-car camera and recognizes their emotional state in real time. If a passenger is panicking, it will customize the notification content to match their emotion and display a reassuring message.

[0487] Terminal (on-board system for autonomous vehicles)

[0488] Customized notifications based on emotion recognition are displayed on in-car displays to encourage passengers to take appropriate action.

[0489] Feedback and model improvement

[0490] User

[0491] After the evacuation or flood occurs, feedback is provided on emotional state and actual evacuation behavior through a feedback form.

[0492] server

[0493] The feedback data will be analyzed and the results will be reflected in improvements to forecasting models and evacuation plans. Sentiment data will also be analyzed at the same time to improve the quality of notifications and plans for the next disaster response.

[0494] Examples of specific examples and prompts

[0495] Example scene:

[0496] When an autonomous vehicle is operating in heavy rain, the on-board system detects a flood risk area, analyzes the passengers' facial expressions, and selects an appropriate evacuation route. The in-car display shows, "A flood warning has been issued. Please remain calm and follow a safe route."

[0497] Example prompt sentence:

[0498] "Heavy rain warnings are issued and autonomous vehicles detect flood risk areas. If passengers start to panic, choose safe routes and display reassuring messages."

[0499] In this way, this invention is a system that combines highly accurate flood risk prediction with passenger emotion recognition, encouraging swift and appropriate evacuation behavior and ensuring safety. In particular, the introduction of an emotion engine enables flexible responses based on passenger reactions, further enhancing the effectiveness of disaster prevention measures.

[0500] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0501] Step 1:

[0502] The server retrieves river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases. The API key and parameters for the data to be retrieved are used as input. This outputs the latest meteorological and topographical data.

[0503] Step 2:

[0504] The server converts the acquired data into a unified format. The acquired river water level data, precipitation data, and topographical data are used as input. The data conversion process complements missing data and filters noise, preparing the data in a format that can be input into the generative AI model. The output is preprocessed data in a unified format.

[0505] Step 3:

[0506] The server inputs the preprocessed data in a unified format into a generative AI model to predict flood risk. The preprocessed data set is used as input. The generative AI model uses a deep learning algorithm to analyze past flood data, which outputs a numerical value for the risk of flooding within the next 48 hours.

[0507] Step 4:

[0508] The server performs risk assessment based on the predicted flood risk. Flood risk prediction data from the generative AI model is used as input. The risk assessment process quantifies the risk level of a specific area and visually displays it in color on a map. The output is risk level information on the map.

[0509] Step 5:

[0510] The server notifies local government disaster prevention personnel and residents of the risk assessment results. The inputs are the risk assessment results and notification information (email address, phone number, etc.). The notification process sends information including the risk level for a specific area and evacuation instructions via email, SMS, or a dedicated app. The output is a notification message sent to local government officials and residents.

[0511] Step 6:

[0512] The terminal (smartphone or personal computer of local government officials or residents) receives the notified information and checks the risk level and evacuation plan. The received notification message is used as input. The risk level and evacuation route are displayed on the terminal, urging the user to take prompt action. The output is display information that serves as a guide for the user's actions.

[0513] Step 7:

[0514] The server analyzes the passenger's facial expressions using the in-car camera and recognizes their emotional state in real time. It uses image data acquired from the in-car camera as input. It uses an emotion recognition engine to analyze the facial expression data and determine the passenger's emotional state. The output is passenger's emotional state information.

[0515] Step 8:

[0516] The server customizes the notification content based on the passenger's emotional state. The inputs are the emotional state information and the risk assessment results. In the notification customization process, if the passenger is in a panic state, the message is changed to one that gives a sense of security. The output is the customized notification message.

[0517] Step 9:

[0518] The server proposes safe routes to the autonomous vehicle to avoid areas with high flood risk. It uses the risk assessment results as input. It calculates a safe path using a route selection algorithm and sends it to the autonomous vehicle. The output is the route information for the autonomous vehicle.

[0519] Step 10:

[0520] The terminal (the autonomous vehicle's on-board system) sets the autonomous driving route based on the proposed route. Route information from the server is used as input. The route information is notified to passengers via the in-car display and audio system, ensuring safe driving. The output is the autonomous vehicle's driving route.

[0521] Step 11:

[0522] After an evacuation or flood occurs, users provide feedback about their emotional state and actual evacuation behavior through a feedback form. The evacuation experience and emotional state are used as input. The feedback data is recorded and sent to the server. The output is the feedback data sent to the server.

[0523] Step 12:

[0524] The server analyzes the collected feedback data to help improve the forecasting model and evacuation plan. It uses the feedback data as input. It uses analytical algorithms to analyze the data and incorporate it into the model and plan. The output is an improved forecasting model and evacuation plan.

[0525] 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.

[0526] 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.

[0527] 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.

[0528] [Second embodiment]

[0529] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0530] 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.

[0531] 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).

[0532] 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.

[0533] 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.

[0534] 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).

[0535] 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.

[0536] 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.

[0537] 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.

[0538] 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.

[0539] 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.

[0540] 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."

[0541] The present invention is a system that predicts river flood risk with high accuracy and provides risk information to local governments and residents in real time. This system operates in cooperation with a server, terminals, and users.

[0542] Specific Embodiments of the System

[0543] Data collection

[0544] server

[0545] River water level data, precipitation data, and topographical data are obtained from meteorological databases, which are collected in real time through collaboration with meteorological agencies, satellite databases, and on-site observation stations.

[0546] Converting acquired data into a unified format to facilitate subsequent analysis and processing, for example, integrating water level, precipitation, and topographic data from different sources into a consistent format.

[0547] Data analysis

[0548] server

[0549] The data converted into a unified format is then fed into a generative AI model, which uses advanced machine learning algorithms such as deep learning to combine historical and current flood data to predict flood risk.

[0550] Data preprocessing is performed to remove noise from each data and filter out outliers, ensuring high quality of input data for the predictive model.

[0551] Risk Assessment and Notification

[0552] server

[0553] The output of the generative AI model is used to assess the level of flood risk. This quantifies the flood risk for a specific area and displays it according to the risk level. This display is then displayed on a map as a color-coded risk map.

[0554] The system notifies local government disaster prevention officials and residents of risk assessment results in real time, providing information quickly via email, SMS, and a dedicated app.

[0555] Evacuation Planning and Response

[0556] server

[0557] Based on the results of the risk assessment, specific evacuation plans and infrastructure operation plans are generated, which are customized for each local government.

[0558] Devices (smartphones and computers of local government employees and residents)

[0559] Check the received flood risk information and suggested evacuation plans. For example, use a smartphone app to intuitively understand evacuation locations and routes on a map.

[0560] User

[0561] After an evacuation or flood occurs, feedback on actual impacts and actions is sent to the system, collecting data that can be used to improve forecasting models and response plans.

[0562] Specific examples

[0563] Example 1: Real-time flood forecasting and notification

[0564] scene

[0565] In a situation where precipitation is increasing rapidly and river water levels are rising rapidly, the server retrieves the latest data from the weather database.

[0566] server

[0567] Using a generative AI model, the risk of flooding within the next 48 hours is predicted and specific urban areas are identified as being at high risk.

[0568] Device (local government employee's smartphone)

[0569] Receive notifications from the server and check detailed risk maps and evacuation plans.

[0570] User

[0571] Based on risk information, evacuation orders are issued to residents, who then begin evacuating promptly in accordance with the instructions.

[0572] Example 2: Feedback and Improvement

[0573] scene

[0574] It was reported that after the flood occurred, residents were able to evacuate safely and ensure their safety, and the actual impact was low.

[0575] User

[0576] Feedback is sent to the system about the actual impact and evacuation actions.

[0577] server

[0578] Feedback data is collected and analyzed, and reflected in improvements to prediction models and proposed algorithms.

[0579] In this way, the present invention provides a consistent system that covers everything from flood risk prediction to notification and countermeasure proposals, enabling real-time, highly accurate flood response. In particular, the series of processes, including data collection, analysis using generative AI models, risk assessment, notification, and feedback, work together to achieve a rapid and appropriate response.

[0580] The processing flow will be explained below.

[0581] Step 1: Data collection

[0582] server

[0583] It connects to meteorological databases, satellite databases, and local observation stations to obtain the latest river water level, precipitation, and topographical data. Specifically, it uses an API key to send requests to the Meteorological Bureau's services and downloads the data in real time.

[0584] The acquired data is stored in a database and the different data formats from each data source are converted into a unified format, for example, water level data is unified to meters and precipitation data is unified to millimeters.

[0585] Step 2: Data Preprocessing

[0586] server

[0587] The acquired data is subjected to noise removal and outlier filtering. For example, abnormally high or low values ​​are detected and excluded. In addition, if there is missing data, it is supplemented based on past data or surrounding data.

[0588] The data is formatted as time series data or geographic information system (GIS) data, making it suitable for input into generative AI models.

[0589] Step 3: Flood risk forecasting

[0590] server

[0591] The formatted data is then fed into a generative AI model, which uses deep learning algorithms to predict flood risk by combining historical flood data with current weather data to calculate the probability of a flood occurring within the next 48 hours.

[0592] The forecast results are quantified and the flood risk level is evaluated. For example, the flood risk level is displayed on a three-level scale: "low," "medium," or "high."

[0593] Step 4: Risk assessment and visualization

[0594] server

[0595] The predicted flood risk level is visually displayed on a map, with high-risk areas color-coded to provide an intuitive understanding of the level of risk to a particular area.

[0596] Make a list of potentially affected facilities (hospitals, schools, evacuation centers, etc.) and identify where action is needed.

[0597] Step 5: Notification

[0598] server

[0599] Notify local government disaster prevention officials and residents of flood risk information by sending emails and SMS containing the risk level of areas requiring action and recommended action plans (evacuation locations and evacuation routes).

[0600] Devices (smartphones and computers of local government employees and residents)

[0601] Residents can check the received notifications and understand the specific measures to take. Residents can use the smartphone app to check real-time risk information and evacuation plans.

[0602] Step 6: Implement your evacuation plan

[0603] User

[0604] Evacuations are carried out promptly based on the notified flood risk information and proposed evacuation plans. Local government officials issue evacuation orders, prepare evacuation shelters, and direct traffic. Residents also follow the evacuation routes provided to them and evacuate to safe locations.

[0605] Step 7: Gather feedback

[0606] User

[0607] After a flood occurs, feedback is provided to the system regarding the actual impact and evacuation behavior, including the time required for evacuation, the effectiveness of the evacuation plan, and the extent of the damage.

[0608] server

[0609] The collected feedback data will be analyzed to improve forecasting models and evacuation plans, allowing for more accurate and rapid responses the next time floods occur.

[0610] Example 1

[0611] 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."

[0612] To predict river flood risks in real time with high accuracy based on meteorological data, promptly and appropriately notify local government disaster prevention officials and residents of risk information, and provide appropriate evacuation plans.

[0613] 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.

[0614] In this invention, the server includes means for acquiring river water level data, precipitation data, and topographical data from a weather database, means for converting the acquired data into a unified format, means for preprocessing the converted data by removing noise and filtering outliers, means for inputting the preprocessed data into a generative AI model to predict flood risk, means for conducting a risk assessment based on the predicted flood risk and visually displaying the risk level on a map, means for notifying disaster prevention personnel in the local government and residents of the results of the risk assessment, means for proposing evacuation plans based on the notified risk assessment results, and means for collecting feedback from each local government and residents and using it to improve the accuracy of the model. This enables highly accurate real-time prediction of river flood risk and the provision of prompt and appropriate risk notifications and evacuation plans.

[0615] A "weather database" is a database for storing and managing weather information, and provides data on weather such as temperature, precipitation, wind speed, wind direction, and atmospheric pressure.

[0616] "River water level data" refers to data that indicates the water level at a specific point in a river, and is used to predict river conditions and flood risk.

[0617] "Precipitation data" is data that numerically represents the amount of precipitation, such as rain or snow, that fell at a specific location within a certain period of time.

[0618] "Topographic data" refers to data that expresses geographical features and the shape of the terrain as numerical values ​​or images, and includes information such as the height of the earth's surface, undulations, and land use.

[0619] "Unified format" means converting data obtained from different sources into a consistent format and making it conform to a common standard.

[0620] "Noise removal" is a process for removing unnecessary information and outliers in data analysis and processing to improve data accuracy.

[0621] "Outlier filtering" is a process of detecting, removing, or correcting values ​​that are outside the normal range or that are unreasonable from a business perspective, contained in a data set.

[0622] "Preprocessing" refers to the preparatory steps taken to prepare data for analysis, including noise removal and filtering of outliers.

[0623] A "generative AI model" is an algorithmic model that uses machine learning and deep learning to make predictions and classifications from data.

[0624] "Risk assessment" is the evaluation and judgment of the degree and impact of predicted risks in numerical and visual form.

[0625] A "risk level" is an indicator that shows the extent to which a particular risk exists, and is usually expressed as a number or color.

[0626] "Notifying" refers to the act of transmitting specific information to a designated recipient, and can be done via email, SMS, a dedicated app, etc.

[0627] An "evacuation plan" is a plan that outlines specific procedures and routes for evacuation in the event of a danger.

[0628] "Feedback" refers to information that collects evaluations and comments regarding the operation and results of the system and is used to help improve it in the future.

[0629] The present invention is implemented as a system that predicts river flood risk with high accuracy and provides risk information to local governments and residents in real time. This system operates in cooperation with a server, terminals, and users.

[0630] Data collection

[0631] server

[0632] The server uses APIs to obtain river water level data, precipitation data, and topographical data in real time from meteorological databases, satellite databases, and local observation stations. Specifically, the server calls the meteorological database API to obtain the latest precipitation data.

[0633] Data integration

[0634] server

[0635] It converts acquired data of different formats into a unified format, generating a consistent dataset that allows subsequent processing to be performed efficiently and effectively.

[0636] Data Preprocessing

[0637] server

[0638] The server then removes noise and filters out outliers from the data converted into a unified format. Specifically, it complements missing values, normalizes the data, and detects and removes abnormally high water level data.

[0639] Flood risk prediction

[0640] server

[0641] The preprocessed data is then fed into a generative AI model, which uses advanced machine learning algorithms such as deep learning to combine historical and current flood data to predict flood risk. For example, future rainfall and current water level data are fed into the model to predict the risk of flooding within 48 hours.

[0642] Risk Assessment and Notification

[0643] server

[0644] The output of the generated AI model is analyzed to assess the level of flood risk, which then calculates a risk index for a specific area and generates a risk map that displays different colors on a map according to the risk level.

[0645] Local government disaster prevention officials and residents will be notified of the risk assessment results in real time via email, SMS, and a dedicated app, along with a risk map.

[0646] Generate evacuation plans

[0647] server

[0648] Based on the flood risk assessment results, the server generates specific evacuation plans customized for each municipality, for example, providing residents in high-risk areas with a list of evacuation sites and optimal evacuation routes.

[0649] Real-time notifications and displays

[0650] Terminal

[0651] The devices of local government officials and residents receive the risk information and evacuation plans sent from the server. For example, the smartphones of local government officials receive emergency notifications and display detailed risk maps and evacuation routes on a dedicated app.

[0652] Collecting and analyzing feedback

[0653] User

[0654] Users provide feedback to the system about the actual impacts and actions they take after a flood, for example, residents reporting on the congestion of evacuation shelters and the actual flooding situation.

[0655] server

[0656] The server analyzes the collected feedback data and uses it to improve the prediction model and response plan. For example, based on feedback that an evacuation site was overcrowded, the server can improve the evacuation plan to suggest a different evacuation site for the next time.

[0657] Prompt Sentence Examples

[0658] "Collect real-time river water level and precipitation data, use generative AI models to predict flood risk, communicate risk assessment results to city officials and residents, and develop evacuation plans."

[0659] The system covers everything from data collection to risk prediction, notification, evacuation plan generation, and feedback collection, enabling highly accurate prediction of river flood risks and supporting rapid response in real time.

[0660] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0661] Step 1: Data collection

[0662] server

[0663] Input: weather database, satellite database, data requests from local observation stations

[0664] The server uses an API to obtain river water level data, precipitation data, and topographical data in real time from meteorological databases, satellite databases, and local observation stations.

[0665] Output: Acquired data (river water level data, precipitation data, topographical data)

[0666] Step 2: Integrate the data

[0667] server

[0668] Input: Acquired river water level data, precipitation data, topographical data

[0669] The server converts the acquired data into a unified format and generates a consistent data set. The server converts and maps the data to harmonize the information obtained from different data sources.

[0670] Output: Uniform format dataset

[0671] Step 3: Preprocessing the data

[0672] server

[0673] Input: Uniform format dataset

[0674] The server then performs noise removal and outlier filtering on the data converted into a unified format, for example, by filling in missing values, normalizing the data, and detecting and excluding abnormally high water level data.

[0675] Output: Preprocessed data

[0676] Step 4: Predict flood risk

[0677] server

[0678] Input: Preprocessed data

[0679] The preprocessed data is input into a generative AI model to predict flood risk. The generative AI model uses advanced machine learning algorithms such as deep learning to combine past flood data with current data to calculate future flood risk. For example, it can predict the likelihood of a flood occurring based on rainfall data within the next 48 hours and current water level data.

[0680] Output: Predicted flood risk data

[0681] Step 5: Risk assessment and notification

[0682] server

[0683] Input: Predicted flood risk data

[0684] The output of the generated AI model is analyzed to assess the level of flood risk. The server calculates a risk index for a specific area and generates a color-coded risk map based on the risk level. The server also sends real-time risk information via email, SMS, or a dedicated app to notify local government disaster prevention officials and residents of the results of the risk assessment.

[0685] Output: Risk assessment report and risk map, notification message

[0686] Step 6: Generate an evacuation plan

[0687] server

[0688] Input: Risk Assessment Report

[0689] Based on the flood risk assessment results, the server generates specific evacuation plans customized for each local government. For example, it provides a list of evacuation sites and optimal evacuation routes for residents in high-risk areas. The evacuation plans are displayed in conjunction with risk maps.

[0690] Output: Evacuation plan

[0691] Step 7: Real-time notifications and displays

[0692] Terminal

[0693] Input: Risk assessment report, risk map, draft evacuation plan

[0694] The devices of local government officials and residents receive the risk information and evacuation plans sent from the server. For example, the smartphones of local government officials receive emergency notifications and intuitively display detailed risk maps and evacuation routes on a dedicated app.

[0695] Output: Risk information and evacuation plan displayed on the terminal

[0696] Step 8: Collect and analyze feedback

[0697] User

[0698] Input: User feedback on the actual evacuation situation and impact

[0699] Users provide feedback to the system about the actual impacts and actions they take after a flood, for example reporting on the congestion of evacuation shelters and the actual flooding situation.

[0700] Output: Feedback data sent

[0701] server

[0702] Input: Collected feedback data

[0703] The server collects and analyzes the feedback data and uses it to improve the prediction model and response plan. For example, based on feedback that an evacuation site was overcrowded, it will suggest a different evacuation site for next time.

[0704] Output: Improved forecast model and evacuation plan

[0705] In this way, this system predicts river flood risks with high accuracy and in real time through step-by-step processing, providing appropriate information and supporting evacuation planning.

[0706] (Application example 1)

[0707] 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."

[0708] In recent years, climate change has led to an increase in extreme weather events, increasing the risk of river flooding. However, existing systems have difficulty accurately predicting flood risk in real time and responding quickly. Furthermore, evacuation plans and evacuation route suggestions are not provided properly, which can lead to problems in ensuring the safety of residents. Additionally, there is a lack of a feedback function to continuously improve the prediction model based on collected data.

[0709] 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.

[0710] In this invention, the server includes: means for acquiring river water level data, precipitation data, and topographical data from a weather database; means for converting the acquired data into a unified format; and means for inputting the converted data into a generative AI model to predict flood risk. This enables real-time, highly accurate prediction of flood risk and prompt response. The server also includes means for performing a risk assessment based on the predicted flood risk and visually displaying the risk level on a map; and means for notifying local government disaster prevention personnel and residents of the risk assessment results. This enables risk information to be provided promptly to residents and local governments. The server also includes means for proposing an evacuation plan and an optimal evacuation route based on the notified risk assessment results, and means for displaying evacuation sites in a specific area based on the user's location information. The server also includes means for collecting feedback from local governments and residents and using it to improve the accuracy of the model, thereby ensuring the safety of residents and continuously improving the prediction model.

[0711] "Weather database" refers to a database that collects, stores, and provides weather information, including river water level data, precipitation data, and weather forecast data.

[0712] "River water level data" refers to data that indicates the height of the water level in a river, and is information necessary for assessing flood risk.

[0713] "Precipitation data" refers to data that indicates the amount of precipitation over a certain period of time in a specific area, and is an important factor in predicting the risk of flooding.

[0714] "Topographical data" refers to data containing information about the topography of an area, which is useful for planning the extent of flood impacts and evacuation routes.

[0715] A "unified format" is a consistent data format for converting data obtained from different data sources into a format that is easy to analyze.

[0716] A "generative AI model" is an artificial intelligence model used to predict flood risk based on historical and current data, and specifically includes deep learning models.

[0717] "Risk assessment" is the process of quantifying and assessing the flood risk of a specific area based on the output of a generative AI model.

[0718] "Visually displaying" means presenting the assessed risk to the user in a visual format such as a map or graph.

[0719] "Notifying" means informing users of the assessed flood risk and related information via email, SMS, smartphone apps, etc.

[0720] An "evacuation plan" is a set of instructions and route guidance designed to enable residents to evacuate safely when the risk of flooding increases.

[0721] The "optimal evacuation route" is a route that suggests the safest and quickest evacuation route based on the user's current location.

[0722] "Feedback" refers to residents and local governments sending information back to the system about actual evacuation behavior and the impact of flooding, which is used to improve the accuracy of the model.

[0723] The system proposed in this invention aims to predict river flood risk in real time and provide prompt and appropriate information to residents and local governments. To achieve this, it is necessary for the server, terminals, and users to work together.

[0724] server

[0725] Data collection

[0726] The server obtains river water level data, precipitation data, and topographical data from a meteorological database, which uses a database that collects, stores, and provides meteorological information.

[0727] Converting acquired data into a unified format: Data obtained from different data sources is converted into a unified data format to make it easier to analyze.

[0728] Data analysis

[0729] The data converted into a unified format is then input into a generative AI model, an artificial intelligence model used to predict flood risk based on past and current data. A deep learning model is typically used.

[0730] Generative AI models are implemented using advanced machine learning frameworks such as TensorFlow and Keras, which denoise data and filter outliers to generate high-quality input data.

[0731] Risk Assessment and Notification

[0732] The level of flood risk is assessed based on the output of the generative AI model. The risk assessment quantifies the flood risk of a specific area and displays it according to the risk level.

[0733] The results of the risk assessment are visually displayed on a map and are notified in real time to local government disaster prevention officials and residents via email, SMS, and a dedicated smartphone app.

[0734] Evacuation Planning and Response

[0735] Based on the risk assessment results, the system proposes specific evacuation plans and optimal evacuation routes. Based on the user's location information, it displays evacuation sites in specific areas.

[0736] Evacuation plans can be customized to suit the needs of each local government.

[0737] Terminal

[0738] The terminal (smartphone or PC) functions as a device for checking the received flood risk information and the proposed evacuation plan. For example, using a smartphone app, users can intuitively check evacuation locations and evacuation routes on a map.

[0739] User

[0740] Users act quickly based on the risk information provided. After an evacuation or flood occurs, they provide feedback to the system about the actual impact and actions taken. This feedback gathers data that can be used to improve predictive models and response plans.

[0741] Specific examples

[0742] Real-time flood forecasting and notifications

[0743] When residents press the "Check current flood risk" button, the latest flood risk map is retrieved from the server and displayed on the screen. When residents select "Show nearest evacuation site," the smartphone app identifies the user's current location and displays the optimal evacuation site and route on the map.

[0744] Prompt Sentence Examples

[0745] 1. "Check your current flood risk"

[0746] 2. "Display the nearest evacuation site"

[0747] Through these functions, this system aims to ensure the safety of residents and provide rapid evacuation support, while also improving the accuracy of the model.

[0748] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0749] Step 1:

[0750] The server retrieves river water level data, precipitation data, and topography data from the weather database. This includes specific operations such as collecting data in real time using APIs. The input data includes information on water level, precipitation, and topography, and the output is raw data retrieved in bulk.

[0751] Step 2:

[0752] The server converts the acquired data into a unified format. Specifically, it converts data in different formats into a standardized format and processes it to make it easier to analyze. The input data is raw data, and the output data is data converted into a unified format. This process includes data normalization and unit unification.

[0753] Step 3:

[0754] The server inputs the converted data into a unified format into a generative AI model to predict flood risk. The generative AI model is implemented using TensorFlow and Keras, and predicts risk based on past and current data. The input data is standardized data, and the output data is the predicted flood risk level. This step also removes noise from the data and filters out outliers.

[0755] Step 4:

[0756] The server performs a risk assessment based on the output of the generative AI model, quantifying and assessing the flood risk of a specific area. The resulting risk level is visually displayed on a map. The input data is the predicted flood risk level, and the output data is a visually displayed risk map. Specifically, the risk level is displayed color-coded on the map application.

[0757] Step 5:

[0758] The server notifies local government disaster prevention personnel and residents of the risk assessment results. Specifically, notifications are sent via email, SMS, and a dedicated smartphone app. The input data is the risk assessment results, and the output data is the notification message. This process also includes prioritizing notifications and selecting recipients.

[0759] Step 6:

[0760] The server proposes an evacuation plan and optimal evacuation route based on the notified risk assessment results. It displays evacuation locations in a specific area based on the user's location information. The input data are the risk assessment results and the user's location information, and the output data are the evacuation plan and evacuation route. Specifically, it displays the optimal evacuation route on a map application.

[0761] Step 7:

[0762] The device checks the received flood risk information and proposed evacuation plan and presents it to the user. The user uses a smartphone app to intuitively check evacuation locations and routes on a map. The input data is the notified risk information and evacuation plan, and the output data is the displayed evacuation map.

[0763] Step 8:

[0764] Users send feedback to the system about the actual impacts and actions taken after an evacuation or flood occurs. The input data is feedback information about evacuation actions and impacts, and the output data is the data stored on the server that received it. In this step, the prediction model and response plan are improved based on the collected feedback.

[0765] In this way, each step works in coordination to predict river flood risks with high accuracy and enable appropriate responses in real time.

[0766] 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.

[0767] This invention combines a system that obtains river water level data, precipitation data, and topographical data from meteorological and satellite databases, and uses this data to predict flood risk with high accuracy, with an emotion engine that recognizes the user's emotions. This system operates in cooperation with a server, terminals, and users.

[0768] Specific Embodiments of the System

[0769] Data collection and analysis

[0770] server

[0771] Obtain the latest river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases. This includes obtaining data from the Meteorological Agency using an API key and analyzing image data from the satellite database.

[0772] The acquired data is converted into a unified format, missing data is imputed, and noise is filtered out, and the converted data is used as input for generative AI models.

[0773] Flood risk forecasting

[0774] server

[0775] After the pre-processing process is complete, the data is fed into a generative AI model, which uses deep learning algorithms to predict flood risk. The model takes into account historical flood data, current water levels, precipitation, and topographical data to calculate the risk of flooding within the next 48 hours.

[0776] The prediction results are quantified according to the risk level, and a risk assessment is carried out based on this, with the risk level of a specific area being visually displayed on a map using different colors.

[0777] Risk notification and evacuation planning

[0778] server

[0779] Flood risk information is sent to local government disaster prevention officials and residents via email, SMS, and a dedicated app, and includes specific risk levels and evacuation instructions.

[0780] Devices (smartphones and computers of local government employees and residents)

[0781] Residents can receive the information and check the risk level. Detailed evacuation plans and evacuation routes are also provided at the same time, allowing them to take action quickly.

[0782] Application of Emotion Engine

[0783] server

[0784] The emotion engine analyzes the user's emotional state when receiving and responding to notifications. For example, it analyzes emails, in-app reaction data, and text data obtained from feedback forms to recognize the user's emotions in real time.

[0785] The notification language and content are adjusted based on the user's emotional state. Specifically, if a user is in a panic despite the urgency of the situation, the notification language will be changed to one that helps the user remain calm.

[0786] Feedback and Improvements

[0787] User

[0788] After the evacuation and flood events, provide feedback on emotional states and actual evacuation behavior, for example, recording fear and difficulties felt during evacuation and the effectiveness of evacuation routes.

[0789] server

[0790] The collected feedback data will be analyzed to improve forecasting models and evacuation plans, and sentiment data will also be analyzed to improve the quality of notifications and plans for the next disaster response.

[0791] Specific examples

[0792] Example 1: Flood risk prediction and emotional response

[0793] scene

[0794] A heavy rain warning is issued, causing river water levels to rise rapidly. The server retrieves the latest water level and precipitation data from the weather database and uses a deep learning model to predict flood risk.

[0795] server

[0796] Determine whether a particular area is at high risk and assess the risk level.

[0797] The emotion engine analyzed that some users had panicked during previous evacuations, and this time the notification sent a message encouraging those users to remain calm and act quickly.

[0798] Device (resident's smartphone)

[0799] Residents receive notifications, reassuring instructions along with evacuation routes, and take action quickly.

[0800] Example 2: Feedback collection and system improvement

[0801] scene

[0802] Flooding occurs and residents are evacuated. After evacuation, residents provide feedback on their feelings and the evacuation situation.

[0803] User

[0804] In the feedback form, participants provided specific feedback such as, "I panicked during the evacuation, but the notification message was very helpful."

[0805] server

[0806] The feedback and sentiment data will be analyzed and reflected in response measures for the next flood, and the sentiment engine algorithm will be improved to provide more effective risk notifications and evacuation plans.

[0807] In this way, the present invention is a system that combines highly accurate flood risk prediction with user emotional responses, encouraging prompt and appropriate evacuation behavior and ensuring safety. In particular, the introduction of an emotion engine enables flexible responses based on user reactions, further enhancing the effectiveness of disaster prevention measures.

[0808] The processing flow will be explained below.

[0809] Step 1: Data collection

[0810] server

[0811] Connect to meteorological and satellite databases to get real-time river water level data, precipitation data, and topographic data. Use an API key to request data from the meteorological bureau and download the latest data.

[0812] The acquired data is stored in an internal database and data from different data sources is converted into a unified format, for example, water level data is standardized to meters and precipitation data is converted to millimeters.

[0813] Step 2: Data Preprocessing

[0814] server

[0815] The acquired data is subjected to noise removal and outlier filtering. For example, abnormally high or low measurement values ​​are detected and excluded. In addition, if there is missing data, it is supplemented based on past data or surrounding data.

[0816] The data is formatted as time series data so that it can be input into a generative AI model.

[0817] Step 3: Flood risk forecasting

[0818] server

[0819] The formatted data is then fed into a generative AI model, which uses deep learning algorithms to predict flood risk. The generative AI model calculates the probability of flooding within the next 48 hours based on historical flood data, current water levels, precipitation, and topographical data.

[0820] The forecast results are quantified and the flood risk level for each area is evaluated. For example, flood risk levels are displayed on a three-level scale: "low," "medium," and "high."

[0821] Step 4: Risk assessment and visualization

[0822] server

[0823] The predicted flood risk level is visually displayed on a map, with high-risk areas color-coded to provide an intuitive understanding of the level of risk to a particular area.

[0824] Make a list of potentially affected facilities (hospitals, schools, evacuation centers, etc.) and identify where action is needed.

[0825] Step 5: Risk notification and sentiment analysis

[0826] server

[0827] Flood risk information is sent to local government disaster prevention officials and residents via email, SMS, and a dedicated app, and includes specific risk levels and evacuation instructions.

[0828] The emotion engine collects user reaction data when receiving notifications and analyzes their emotional state. For example, it recognizes emotions based on the user's immediate reaction when opening a notification or text input from a feedback form.

[0829] Step 6: Evacuation planning and emotional response

[0830] server

[0831] The content and presentation of notification messages are adjusted based on the user's emotional state analyzed by the emotion engine. For example, if a user is in a panic despite the urgency of the situation, a message encouraging them to remain calm will be sent.

[0832] Devices (smartphones and computers of local government employees and residents)

[0833] Along with the risk information provided, residents can confirm appropriate evacuation plans and routes. For example, residents can visually check specific evacuation locations and routes via a smartphone app.

[0834] Step 7: Evacuation implementation and feedback

[0835] User

[0836] Evacuate promptly based on the notified flood risk information and the proposed evacuation plan. After completing the evacuation, provide feedback on the emotions and evacuation behavior during the evacuation. For example, enter information such as "the fear and difficulty felt during the evacuation" and "the effectiveness of the evacuation route" in the feedback form.

[0837] server

[0838] The collected feedback data will be analyzed to improve prediction models and evacuation plans. In particular, analyzing emotion data at the same time will improve the quality of notifications and plans for the next disaster response.

[0839] As a result, this invention provides a system that integrates highly accurate flood risk prediction with flexible responses based on user emotions, encouraging prompt and appropriate evacuation behavior. The introduction of an emotion engine makes it possible to realize notifications and evacuation plans based on user reactions, further enhancing the effectiveness of disaster prevention measures.

[0840] Example 2

[0841] 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."

[0842] Current flood risk prediction systems often lack real-time capabilities and accuracy, resulting in delays in evacuation notifications and evacuation plan proposals to residents. These systems also lack the ability to respond to users' emotional states, which can lead to panic among residents in emergencies. As a result, appropriate evacuation behavior is not possible, and safety is not ensured. Furthermore, there is also the issue that feedback is not fully utilized, resulting in the time required to improve the models.

[0843] 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.

[0844] In this invention, the server includes a means for acquiring river water level data, precipitation data, and topographical data from a meteorological database and a satellite database; a means for converting the data into a unified format, filling in missing data, and performing noise filtering; and a means for inputting the converted data into a generative AI model to predict flood risk. This enables real-time and highly accurate prediction of flood risk. Furthermore, risk assessments are performed and displayed visually on a map, and an emotion recognition engine is used to analyze the user's emotional state and adjust notification content to encourage appropriate evacuation behavior. Feedback data can be collected and used to improve the accuracy of the model, further enhancing the effectiveness of future disaster responses.

[0845] A "weather database" is a data storage system for collecting, storing, and providing weather information. It is primarily managed and operated by meteorological agencies and related organizations.

[0846] A "satellite database" is a system that stores data collected from artificial satellites for the purpose of Earth observation. It provides a variety of information, including topographical and meteorological data.

[0847] "River water level data" is numerical information measuring the water surface height of a specific river. It is an important element in flood prediction.

[0848] "Precipitation data" is data that numerically indicates the amount of precipitation (rain, snow, fog, etc.) that fell in a specific area within a certain period of time. It is used for weather forecasting and flood prediction.

[0849] "Terrain data" is data that contains information about the shape, height, and structure of the Earth's surface. It is primarily used for mapping and environmental modeling.

[0850] "Missing data" refers to the absence of required data in a dataset. Also known as missing values.

[0851] "Noise filtering" is a process to remove unnecessary information (noise) from data. It is performed to improve the quality of data.

[0852] A "generative AI model" is a computational model designed to generate new data or information using artificial intelligence, often involving deep learning algorithms.

[0853] "Flood risk" refers to the possibility of flooding occurring in a particular area and the extent of its impact. It is used in risk assessment.

[0854] An "emotion recognition engine" is software that analyzes a user's emotional state from input such as text data and voice data, enabling the system to respond based on the user's reaction.

[0855] An "evacuation plan" is a plan that shows routes and methods for safe evacuation in the event of a disaster. It is provided to ensure the safety of residents.

[0856] "Feedback" refers to opinions, impressions, and evaluation information provided by system users. It is used to improve the system.

[0857] A "municipal government" is a local government or related institution that administers a particular area and ensures the welfare and safety of its residents.

[0858] "Residents" refers to people who live in a specific area. They are often users of the system.

[0859] A "deep learning model" is a machine learning model consisting of a multi-layered neural network. It is used for advanced data analysis and prediction.

[0860] The present invention is a system that predicts flood risk using data acquired from a meteorological database and a satellite database, and provides notifications that correspond to the user's emotional state. A specific embodiment of this system is described below.

[0861] Data collection and analysis

[0862] server

[0863] The server obtains the latest river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases. APIs are used to obtain data from the meteorological bureau and satellite databases. For example, API requests are sent using the Python requests library. The obtained data is stored in a database management system such as PostgreSQL or MySQL.

[0864] Data Preprocessing

[0865] server

[0866] The server converts the acquired data into a unified format, imputes missing data, and performs noise filtering. Python's pandas library is used to unify different data formats and impute missing values. Statistical methods such as median are used to improve data quality during this process. After preprocessing, the data is formatted as input for the generative AI model and converted into NumPy arrays or TensorFlow tensors.

[0867] Flood risk forecasting

[0868] server

[0869] The server inputs the preprocessed data into the generative AI model. A deep learning model is built using TensorFlow and PyTorch, and the data is used for prediction. The model takes into account past flood data, current water levels, precipitation, and topographical data to predict flood risk within the next 48 hours. The prediction results are quantified based on risk levels, and the risk level for specific areas is visually displayed on a map. Map libraries such as Google Maps API and Leaflet are used to display the map.

[0870] Risk notification and evacuation planning

[0871] server

[0872] The server then notifies local government disaster prevention officials and residents of flood risk information via email, SMS, or a dedicated app. For example, AWS SNS (Simple Notification Service) can be used to send SMS notifications.

[0873] Terminal

[0874] The device receives the information and checks the risk level. A push notification is sent to the smartphone app, and by tapping the notification, detailed information is displayed. Evacuation plans and evacuation routes are also presented, and the optimal evacuation route can be displayed using the Google Maps API.

[0875] Emotion recognition and notification adjustment

[0876] server

[0877] The server uses an emotion engine to analyze the emotional state of users receiving notifications in real time. For example, it uses the Natural Language Toolkit (NLTK) and BERT models to analyze notification emails and app response data. Based on the analysis results, it adjusts the wording and content of notifications. For example, it sends an encouraging message to a panicked user to help them stay calm.

[0878] Feedback and System Improvement

[0879] User

[0880] Users provide feedback after an evacuation or flood occurs. For example, they can write in the app's feedback form that they felt scared during the evacuation but found the notifications helpful.

[0881] server

[0882] The server analyzes the collected feedback data to help improve the predictive models and evacuation plans. It uses data mining techniques and sentiment analysis algorithms to analyze the feedback, which will enable it to provide more effective risk notifications and evacuation plans in the next disaster response.

[0883] Specific examples

[0884] The following is a specific example of how this system works.

[0885] Scene 1: Processing in the event of a heavy rain warning

[0886] The server uses an API to retrieve the latest precipitation and river water level data from the weather database.

[0887] Convert the acquired data into a unified format and fill in any missing data.

[0888] The preprocessed data is input into a generative AI model to predict flood risk.

[0889] High-risk areas are identified and marked in red on the map.

[0890] Scene 2: Notifications and Emotion Recognition

[0891] The server sends an SMS to users who live in high-risk areas, for example, "Your area is at high risk of flooding. Please begin evacuation."

[0892] The user views the SMS on their smartphone and checks the evacuation route.

[0893] The server uses an emotion engine to analyze the user's reaction and detect whether they are in a panic state.

[0894] The server sends a calming message to panicked users, for example, "Remain calm and evacuate. You are safe."

[0895] Based on these examples, the system can ensure the safety of residents by providing highly accurate predictions of flood risk and prompt evacuation notifications that respond to users' emotions.

[0896] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0897] Step 1: Data collection

[0898] server

[0899] The server retrieves river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases, and sends API requests using the Python requests library.

[0900] Input: API request (weather database and satellite database)

[0901] Output: Acquired river water level data, precipitation data, topographical data

[0902] The acquired data is stored in a database management system such as PostgreSQL or MySQL.

[0903] Step 2: Data Preprocessing

[0904] server

[0905] The data retrieved by the server is converted into a unified format. The data is cleaned and formatted consistently using the Python pandas library.

[0906] The server imputes missing data and performs noise filtering, for example by imputing missing values ​​with the median to ensure data consistency.

[0907] Input: Acquired river water level data, precipitation data, topographical data

[0908] Output: Preprocessed data (unified format, missing data imputation, noise filtering)

[0909] The preprocessed data is converted into NumPy arrays or TensorFlow tensors.

[0910] Step 3: Flood risk forecasting

[0911] server

[0912] The server inputs the preprocessed data into a generative AI model, which uses a deep learning model built using TensorFlow or PyTorch.

[0913] The model takes into account historical flood data, current water levels, rainfall and topographical data to predict flood risk within the next 48 hours.

[0914] Input: Preprocessed data (NumPy arrays or TensorFlow tensors)

[0915] Output: Flood risk rating for each region (ranging from 0 to 1)

[0916] High-risk areas are color-coded on a map, and the map is displayed using Google Maps API and Leaflet.

[0917] Step 4: Risk notification

[0918] server

[0919] The server notifies local government disaster prevention officials and residents of flood risk information via email, SMS, and a dedicated app.

[0920] Send SMS notifications using AWS SNS (Simple Notification Service).

[0921] Input: Flood risk assessment results, notification messages based on risk level

[0922] Output: Notifications (email, SMS) sent to local government disaster prevention officials and residents

[0923] Terminal

[0924] The device receives the information and checks the risk level. A push notification is sent to the smartphone app, displaying detailed information.

[0925] The device will present the user with an evacuation plan and evacuation route, using the Google Maps API to display safe evacuation routes.

[0926] Input: Risk notification message, evacuation plan information

[0927] Output: Notifications and evacuation routes displayed to the user

[0928] Step 5: Emotion recognition and notification adjustment

[0929] server

[0930] The server uses an emotion engine to analyze the emotional state of the user receiving the notification information in real time. It analyzes the text data using NLTK (Natural Language Toolkit) and the BERT model.

[0931] Tailor the wording and content of notifications based on the user's emotional state: if a user is panicking, send them a notification that helps them stay calm.

[0932] Input: User emotion data (text, reaction data)

[0933] Output: Adjusted notification message

[0934] Example: A message encouraging people to remain calm and evacuate.

[0935] Step 6: Gather feedback and improve the system

[0936] User

[0937] After an evacuation or flood occurs, users provide feedback on their emotional state and actual evacuation behavior. For example, they can write in the in-app feedback form, "I was able to make appropriate decisions quickly when evacuating."

[0938] Input: Feedback form input (text data)

[0939] Output: Feedback data provided

[0940] server

[0941] The server analyzes the collected feedback data to help improve prediction models and evacuation plans. It uses data mining techniques and sentiment analysis algorithms to analyze the feedback.

[0942] Input: Feedback data (emotional information, behavioral records)

[0943] Output: Improved models and evacuation plans

[0944] Improved emotion engine algorithms will provide more effective risk notifications and evacuation plans for the next flood.

[0945] This allows the system to achieve highly accurate flood risk prediction and user response at each step, promoting safe evacuation behavior.

[0946] (Application example 2)

[0947] 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."

[0948] Conventional flood risk prediction systems primarily use meteorological and satellite data to predict flood risks and notify local governments and residents. However, they do not take into account the emotional state of passengers, particularly in autonomous vehicles, which can lead to confusion and anxiety caused by emotions such as panic. Furthermore, they lack the ability to automatically avoid areas with a high risk of flooding, which can make it difficult to fully ensure passenger safety. A system that can effectively solve these issues is needed.

[0949] 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.

[0950] In this invention, the server includes means for acquiring river water level data, precipitation data, and topographical data from a meteorological database and a satellite database, means for converting the acquired data into a unified format, and means for inputting the converted data into a generative AI model to predict flood risk, thereby enabling highly accurate flood risk prediction and risk assessment.

[0951] The server also includes a means for performing a risk assessment based on the predicted flood risk and visually displaying the risk level on a map, a means for notifying local government disaster prevention personnel and residents of the risk assessment results, a means for proposing an evacuation plan based on the notified risk assessment results, a means for collecting feedback from each local government and residents and using it to improve the accuracy of the model, a means for recognizing the emotional state of passengers, a means for customizing the content of the notification based on the emotional state, and a means for the autonomous vehicle to automatically select a route that avoids the flood risk. This enables flexible notifications according to the emotional state of passengers and enables the autonomous vehicle to select a safe route.

[0952] A "weather database" is a database that includes various weather data (e.g., precipitation data, temperature data, wind speed data, etc.).

[0953] A "satellite database" is a database that includes topographical data and water level data obtained from artificial satellites.

[0954] "River water level data" is data that indicates information on the water level in a specific river basin.

[0955] "Precipitation data" is data that indicates information about the amount of precipitation in a specific area.

[0956] "Topography data" refers to data that indicates information such as the topography and altitude of a specific area.

[0957] A "unified format" is a format for converting data obtained from different data sources into a consistent format.

[0958] A "generative AI model" is a machine learning model that makes predictions and classifications based on data.

[0959] "Flood risk" is an indicator of the likelihood of flooding occurring in a particular area within a certain period of time.

[0960] "Risk assessment" is the process of determining danger based on acquired data and indicating the risk level using numbers, colors, etc.

[0961] A "local government disaster prevention officer" is an official responsible for disaster prevention measures in a local government.

[0962] "Residents" refers to people who live in a particular area.

[0963] An "evacuation plan" is a plan for safely evacuating in the event of a disaster such as a flood.

[0964] "Feedback" refers to information that collects opinions and impressions from users and is used to improve the system.

[0965] "Passenger emotional state" is an index that indicates the emotional state that passengers are feeling (e.g., relief, anxiety, panic, etc.).

[0966] "Customizing notification content" means changing the content of a signal or message depending on the emotional state of the recipient.

[0967] An "autonomously selected route" is a safe route that an autonomous vehicle selects to avoid hazards such as flooding.

[0968] This invention is a system for predicting flood risks and safely operating autonomous vehicles. This system obtains river water level data, precipitation data, and topographical data from meteorological and satellite databases, and predicts flood risks using a generative AI model based on this data. Furthermore, the system is characterized by its ability to recognize passenger emotions and customize notification content.

[0969] Data collection and analysis

[0970] server

[0971] Obtain the latest river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases. This includes obtaining data from the Meteorological Agency using an API key and analyzing image data from the satellite database.

[0972] The acquired data is converted into a unified format, missing data is imputed, and noise is filtered out, and the converted data is used as input for generative AI models.

[0973] Flood risk forecasting and risk assessment

[0974] server

[0975] After the pre-processing process is complete, the data is fed into a generative AI model, which uses deep learning algorithms to predict flood risk. The model takes into account historical flood data, current water levels, precipitation, and topographical data to calculate the risk of flooding within the next 48 hours.

[0976] The prediction results are quantified according to the risk level, and a risk assessment is carried out based on this, with the risk level of a specific area being visually displayed on a map using different colors.

[0977] Risk notification and evacuation planning

[0978] server

[0979] Flood risk information is sent to local government disaster prevention officials and residents via email, SMS, and a dedicated app, and includes specific risk levels and evacuation instructions.

[0980] Devices (smartphones and personal computers of local government employees and residents)

[0981] Residents can receive the information and check the risk level. Detailed evacuation plans and evacuation routes are also provided at the same time, allowing them to take action quickly.

[0982] Route selection for autonomous vehicles

[0983] server

[0984] This system suggests safe routes for autonomous vehicles to avoid areas at high risk of flooding, allowing them to operate on routes with lower flood risk.

[0985] Terminal (on-board system for autonomous vehicles)

[0986] The in-vehicle system automatically selects the route for the autonomous vehicle based on the proposed route, and the route information is communicated to passengers via the in-vehicle display and audio system.

[0987] Application of Emotion Engine

[0988] server

[0989] Using an emotion engine, the system analyzes passengers' facial expressions using an in-car camera and recognizes their emotional state in real time. If a passenger is panicking, it will customize the notification content to match their emotion and display a reassuring message.

[0990] Terminal (on-board system for autonomous vehicles)

[0991] Customized notifications based on emotion recognition are displayed on in-car displays to encourage passengers to take appropriate action.

[0992] Feedback and model improvement

[0993] User

[0994] After the evacuation or flood occurs, feedback is provided on emotional state and actual evacuation behavior through a feedback form.

[0995] server

[0996] The feedback data will be analyzed and the results will be reflected in improvements to forecasting models and evacuation plans. Sentiment data will also be analyzed at the same time to improve the quality of notifications and plans for the next disaster response.

[0997] Examples of specific examples and prompts

[0998] Example scene:

[0999] When an autonomous vehicle is operating in heavy rain, the on-board system detects a flood risk area, analyzes the passengers' facial expressions, and selects an appropriate evacuation route. The in-car display shows, "A flood warning has been issued. Please remain calm and follow a safe route."

[1000] Example prompt sentence:

[1001] "Heavy rain warnings are issued and autonomous vehicles detect flood risk areas. If passengers start to panic, choose safe routes and display reassuring messages."

[1002] In this way, this invention is a system that combines highly accurate flood risk prediction with passenger emotion recognition, encouraging swift and appropriate evacuation behavior and ensuring safety. In particular, the introduction of an emotion engine enables flexible responses based on passenger reactions, further enhancing the effectiveness of disaster prevention measures.

[1003] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1004] Step 1:

[1005] The server retrieves river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases. The API key and parameters for the data to be retrieved are used as input. This outputs the latest meteorological and topographical data.

[1006] Step 2:

[1007] The server converts the acquired data into a unified format. The acquired river water level data, precipitation data, and topographical data are used as input. The data conversion process complements missing data and filters noise, preparing the data in a format that can be input into the generative AI model. The output is preprocessed data in a unified format.

[1008] Step 3:

[1009] The server inputs the preprocessed data in a unified format into a generative AI model to predict flood risk. The preprocessed data set is used as input. The generative AI model uses a deep learning algorithm to analyze past flood data, which outputs a numerical value for the risk of flooding within the next 48 hours.

[1010] Step 4:

[1011] The server performs risk assessment based on the predicted flood risk. Flood risk prediction data from the generative AI model is used as input. The risk assessment process quantifies the risk level of a specific area and visually displays it in color on a map. The output is risk level information on the map.

[1012] Step 5:

[1013] The server notifies local government disaster prevention personnel and residents of the risk assessment results. The inputs are the risk assessment results and notification information (email address, phone number, etc.). The notification process sends information including the risk level for a specific area and evacuation instructions via email, SMS, or a dedicated app. The output is a notification message sent to local government officials and residents.

[1014] Step 6:

[1015] The terminal (smartphone or personal computer of local government officials or residents) receives the notified information and checks the risk level and evacuation plan. The received notification message is used as input. The risk level and evacuation route are displayed on the terminal, urging the user to take prompt action. The output is display information that serves as a guide for the user's actions.

[1016] Step 7:

[1017] The server analyzes the passenger's facial expressions using the in-car camera and recognizes their emotional state in real time. It uses image data acquired from the in-car camera as input. It uses an emotion recognition engine to analyze the facial expression data and determine the passenger's emotional state. The output is passenger's emotional state information.

[1018] Step 8:

[1019] The server customizes the notification content based on the passenger's emotional state. The inputs are the emotional state information and the risk assessment results. In the notification customization process, if the passenger is in a panic state, the message is changed to one that gives a sense of security. The output is the customized notification message.

[1020] Step 9:

[1021] The server proposes safe routes to the autonomous vehicle to avoid areas with high flood risk. It uses the risk assessment results as input. It calculates a safe path using a route selection algorithm and sends it to the autonomous vehicle. The output is the route information for the autonomous vehicle.

[1022] Step 10:

[1023] The terminal (the autonomous vehicle's on-board system) sets the autonomous driving route based on the proposed route. Route information from the server is used as input. The route information is notified to passengers via the in-car display and audio system, ensuring safe driving. The output is the autonomous vehicle's driving route.

[1024] Step 11:

[1025] After an evacuation or flood occurs, users provide feedback about their emotional state and actual evacuation behavior through a feedback form. The evacuation experience and emotional state are used as input. The feedback data is recorded and sent to the server. The output is the feedback data sent to the server.

[1026] Step 12:

[1027] The server analyzes the collected feedback data to help improve the forecasting model and evacuation plan. It uses the feedback data as input. It uses analytical algorithms to analyze the data and incorporate it into the model and plan. The output is an improved forecasting model and evacuation plan.

[1028] 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.

[1029] 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.

[1030] 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.

[1031] [Third embodiment]

[1032] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1033] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[1034] 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).

[1035] 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.

[1036] 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.

[1037] 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).

[1038] 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.

[1039] 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.

[1040] 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.

[1041] 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.

[1042] 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.

[1043] 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."

[1044] The present invention is a system that predicts river flood risk with high accuracy and provides risk information to local governments and residents in real time. This system operates in cooperation with a server, terminals, and users.

[1045] Specific Embodiments of the System

[1046] Data collection

[1047] server

[1048] River water level data, precipitation data, and topographical data are obtained from meteorological databases, which are collected in real time through collaboration with meteorological agencies, satellite databases, and on-site observation stations.

[1049] Converting acquired data into a unified format to facilitate subsequent analysis and processing, for example, integrating water level, precipitation, and topographic data from different sources into a consistent format.

[1050] Data analysis

[1051] server

[1052] The data converted into a unified format is then fed into a generative AI model, which uses advanced machine learning algorithms such as deep learning to combine historical and current flood data to predict flood risk.

[1053] Data preprocessing is performed to remove noise from each data and filter out outliers, ensuring high quality of input data for the predictive model.

[1054] Risk Assessment and Notification

[1055] server

[1056] The output of the generative AI model is used to assess the level of flood risk. This quantifies the flood risk for a specific area and displays it according to the risk level. This display is then displayed on a map as a color-coded risk map.

[1057] The system notifies local government disaster prevention officials and residents of risk assessment results in real time, providing information quickly via email, SMS, and a dedicated app.

[1058] Evacuation Planning and Response

[1059] server

[1060] Based on the results of the risk assessment, specific evacuation plans and infrastructure operation plans are generated, which are customized for each local government.

[1061] Devices (smartphones and computers of local government employees and residents)

[1062] Check the received flood risk information and suggested evacuation plans. For example, use a smartphone app to intuitively understand evacuation locations and routes on a map.

[1063] User

[1064] After an evacuation or flood occurs, feedback on actual impacts and actions is sent to the system, collecting data that can be used to improve forecasting models and response plans.

[1065] Specific examples

[1066] Example 1: Real-time flood forecasting and notification

[1067] scene

[1068] In a situation where precipitation is increasing rapidly and river water levels are rising rapidly, the server retrieves the latest data from the weather database.

[1069] server

[1070] Using a generative AI model, the risk of flooding within the next 48 hours is predicted and specific urban areas are identified as being at high risk.

[1071] Device (local government employee's smartphone)

[1072] Receive notifications from the server and check detailed risk maps and evacuation plans.

[1073] User

[1074] Based on risk information, evacuation orders are issued to residents, who then begin evacuating promptly in accordance with the instructions.

[1075] Example 2: Feedback and Improvement

[1076] scene

[1077] It was reported that after the flood occurred, residents were able to evacuate safely and ensure their safety, and the actual impact was low.

[1078] User

[1079] Feedback is sent to the system about the actual impact and evacuation actions.

[1080] server

[1081] Feedback data is collected and analyzed, and reflected in improvements to prediction models and proposed algorithms.

[1082] In this way, the present invention provides a consistent system that covers everything from flood risk prediction to notification and countermeasure proposals, enabling real-time, highly accurate flood response. In particular, the series of processes, including data collection, analysis using generative AI models, risk assessment, notification, and feedback, work together to achieve a rapid and appropriate response.

[1083] The processing flow will be explained below.

[1084] Step 1: Data collection

[1085] server

[1086] It connects to meteorological databases, satellite databases, and local observation stations to obtain the latest river water level, precipitation, and topographical data. Specifically, it uses an API key to send requests to the Meteorological Bureau's services and downloads the data in real time.

[1087] The acquired data is stored in a database and the different data formats from each data source are converted into a unified format, for example, water level data is unified to meters and precipitation data is unified to millimeters.

[1088] Step 2: Data Preprocessing

[1089] server

[1090] The acquired data is subjected to noise removal and outlier filtering. For example, abnormally high or low values ​​are detected and excluded. In addition, if there is missing data, it is supplemented based on past data or surrounding data.

[1091] The data is formatted as time series data or geographic information system (GIS) data, making it suitable for input into generative AI models.

[1092] Step 3: Flood risk forecasting

[1093] server

[1094] The formatted data is then fed into a generative AI model, which uses deep learning algorithms to predict flood risk by combining historical flood data with current weather data to calculate the probability of a flood occurring within the next 48 hours.

[1095] The forecast results are quantified and the flood risk level is evaluated. For example, the flood risk level is displayed on a three-level scale: "low," "medium," or "high."

[1096] Step 4: Risk assessment and visualization

[1097] server

[1098] The predicted flood risk level is visually displayed on a map, with high-risk areas color-coded to provide an intuitive understanding of the level of risk to a particular area.

[1099] Make a list of potentially affected facilities (hospitals, schools, evacuation centers, etc.) and identify where action is needed.

[1100] Step 5: Notification

[1101] server

[1102] Notify local government disaster prevention officials and residents of flood risk information by sending emails and SMS containing the risk level of areas requiring action and recommended action plans (evacuation locations and evacuation routes).

[1103] Devices (smartphones and computers of local government employees and residents)

[1104] Residents can check the received notifications and understand the specific measures to take. Residents can use the smartphone app to check real-time risk information and evacuation plans.

[1105] Step 6: Implement your evacuation plan

[1106] User

[1107] Evacuations are carried out promptly based on the notified flood risk information and proposed evacuation plans. Local government officials issue evacuation orders, prepare evacuation shelters, and direct traffic. Residents also follow the evacuation routes provided to them and evacuate to safe locations.

[1108] Step 7: Gather feedback

[1109] User

[1110] After a flood occurs, feedback is provided to the system regarding the actual impact and evacuation behavior, including the time required for evacuation, the effectiveness of the evacuation plan, and the extent of the damage.

[1111] server

[1112] The collected feedback data will be analyzed to improve forecasting models and evacuation plans, allowing for more accurate and rapid responses the next time floods occur.

[1113] Example 1

[1114] 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."

[1115] To predict river flood risks in real time with high accuracy based on meteorological data, promptly and appropriately notify local government disaster prevention officials and residents of risk information, and provide appropriate evacuation plans.

[1116] 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.

[1117] In this invention, the server includes means for acquiring river water level data, precipitation data, and topographical data from a weather database, means for converting the acquired data into a unified format, means for preprocessing the converted data by removing noise and filtering outliers, means for inputting the preprocessed data into a generative AI model to predict flood risk, means for conducting a risk assessment based on the predicted flood risk and visually displaying the risk level on a map, means for notifying disaster prevention personnel in the local government and residents of the results of the risk assessment, means for proposing evacuation plans based on the notified risk assessment results, and means for collecting feedback from each local government and residents and using it to improve the accuracy of the model. This enables highly accurate real-time prediction of river flood risk and the provision of prompt and appropriate risk notifications and evacuation plans.

[1118] A "weather database" is a database for storing and managing weather information, and provides data on weather such as temperature, precipitation, wind speed, wind direction, and atmospheric pressure.

[1119] "River water level data" refers to data that indicates the water level at a specific point in a river, and is used to predict river conditions and flood risk.

[1120] "Precipitation data" is data that numerically represents the amount of precipitation, such as rain or snow, that fell at a specific location within a certain period of time.

[1121] "Topographic data" refers to data that expresses geographical features and the shape of the terrain as numerical values ​​or images, and includes information such as the height of the earth's surface, undulations, and land use.

[1122] "Unified format" means converting data obtained from different sources into a consistent format and making it conform to a common standard.

[1123] "Noise removal" is a process for removing unnecessary information and outliers in data analysis and processing to improve data accuracy.

[1124] "Outlier filtering" is a process of detecting, removing, or correcting values ​​that are outside the normal range or that are unreasonable from a business perspective, contained in a data set.

[1125] "Preprocessing" refers to the preparatory steps taken to prepare data for analysis, including noise removal and filtering of outliers.

[1126] A "generative AI model" is an algorithmic model that uses machine learning and deep learning to make predictions and classifications from data.

[1127] "Risk assessment" is the evaluation and judgment of the degree and impact of predicted risks in numerical and visual form.

[1128] A "risk level" is an indicator that shows the extent to which a particular risk exists, and is usually expressed as a number or color.

[1129] "Notifying" refers to the act of transmitting specific information to a designated recipient, and can be done via email, SMS, a dedicated app, etc.

[1130] An "evacuation plan" is a plan that outlines specific procedures and routes for evacuation in the event of a danger.

[1131] "Feedback" refers to information that collects evaluations and comments regarding the operation and results of the system and is used to help improve it in the future.

[1132] The present invention is implemented as a system that predicts river flood risk with high accuracy and provides risk information to local governments and residents in real time. This system operates in cooperation with a server, terminals, and users.

[1133] Data collection

[1134] server

[1135] The server uses APIs to obtain river water level data, precipitation data, and topographical data in real time from meteorological databases, satellite databases, and local observation stations. Specifically, the server calls the meteorological database API to obtain the latest precipitation data.

[1136] Data integration

[1137] server

[1138] It converts acquired data of different formats into a unified format, generating a consistent dataset that allows subsequent processing to be performed efficiently and effectively.

[1139] Data Preprocessing

[1140] server

[1141] The server then removes noise and filters out outliers from the data converted into a unified format. Specifically, it complements missing values, normalizes the data, and detects and removes abnormally high water level data.

[1142] Flood risk prediction

[1143] server

[1144] The preprocessed data is then fed into a generative AI model, which uses advanced machine learning algorithms such as deep learning to combine historical and current flood data to predict flood risk. For example, future rainfall and current water level data are fed into the model to predict the risk of flooding within 48 hours.

[1145] Risk Assessment and Notification

[1146] server

[1147] The output of the generated AI model is analyzed to assess the level of flood risk, which then calculates a risk index for a specific area and generates a risk map that displays different colors on a map according to the risk level.

[1148] Local government disaster prevention officials and residents will be notified of the risk assessment results in real time via email, SMS, and a dedicated app, along with a risk map.

[1149] Generate evacuation plans

[1150] server

[1151] Based on the flood risk assessment results, the server generates specific evacuation plans customized for each municipality, for example, providing residents in high-risk areas with a list of evacuation sites and optimal evacuation routes.

[1152] Real-time notifications and displays

[1153] Terminal

[1154] The devices of local government officials and residents receive the risk information and evacuation plans sent from the server. For example, the smartphones of local government officials receive emergency notifications and display detailed risk maps and evacuation routes on a dedicated app.

[1155] Collecting and analyzing feedback

[1156] User

[1157] Users provide feedback to the system about the actual impacts and actions they take after a flood, for example, residents reporting on the congestion of evacuation shelters and the actual flooding situation.

[1158] server

[1159] The server analyzes the collected feedback data and uses it to improve the prediction model and response plan. For example, based on feedback that an evacuation site was overcrowded, the server can improve the evacuation plan to suggest a different evacuation site for the next time.

[1160] Prompt Sentence Examples

[1161] "Collect real-time river water level and precipitation data, use generative AI models to predict flood risk, communicate risk assessment results to city officials and residents, and develop evacuation plans."

[1162] The system covers everything from data collection to risk prediction, notification, evacuation plan generation, and feedback collection, enabling highly accurate prediction of river flood risks and supporting rapid response in real time.

[1163] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1164] Step 1: Data collection

[1165] server

[1166] Input: weather database, satellite database, data requests from local observation stations

[1167] The server uses an API to obtain river water level data, precipitation data, and topographical data in real time from meteorological databases, satellite databases, and local observation stations.

[1168] Output: Acquired data (river water level data, precipitation data, topographical data)

[1169] Step 2: Integrate the data

[1170] server

[1171] Input: Acquired river water level data, precipitation data, topographical data

[1172] The server converts the acquired data into a unified format and generates a consistent data set. The server converts and maps the data to harmonize the information obtained from different data sources.

[1173] Output: Uniform format dataset

[1174] Step 3: Preprocessing the data

[1175] server

[1176] Input: Uniform format dataset

[1177] The server then performs noise removal and outlier filtering on the data converted into a unified format, for example, by filling in missing values, normalizing the data, and detecting and excluding abnormally high water level data.

[1178] Output: Preprocessed data

[1179] Step 4: Predict flood risk

[1180] server

[1181] Input: Preprocessed data

[1182] The preprocessed data is input into a generative AI model to predict flood risk. The generative AI model uses advanced machine learning algorithms such as deep learning to combine past flood data with current data to calculate future flood risk. For example, it can predict the likelihood of a flood occurring based on rainfall data within the next 48 hours and current water level data.

[1183] Output: Predicted flood risk data

[1184] Step 5: Risk assessment and notification

[1185] server

[1186] Input: Predicted flood risk data

[1187] The output of the generated AI model is analyzed to assess the level of flood risk. The server calculates a risk index for a specific area and generates a color-coded risk map based on the risk level. The server also sends real-time risk information via email, SMS, or a dedicated app to notify local government disaster prevention officials and residents of the results of the risk assessment.

[1188] Output: Risk assessment report and risk map, notification message

[1189] Step 6: Generate an evacuation plan

[1190] server

[1191] Input: Risk Assessment Report

[1192] Based on the flood risk assessment results, the server generates specific evacuation plans customized for each local government. For example, it provides a list of evacuation sites and optimal evacuation routes for residents in high-risk areas. The evacuation plans are displayed in conjunction with risk maps.

[1193] Output: Evacuation plan

[1194] Step 7: Real-time notifications and displays

[1195] Terminal

[1196] Input: Risk assessment report, risk map, draft evacuation plan

[1197] The devices of local government officials and residents receive the risk information and evacuation plans sent from the server. For example, the smartphones of local government officials receive emergency notifications and intuitively display detailed risk maps and evacuation routes on a dedicated app.

[1198] Output: Risk information and evacuation plan displayed on the terminal

[1199] Step 8: Collect and analyze feedback

[1200] User

[1201] Input: User feedback on the actual evacuation situation and impact

[1202] Users provide feedback to the system about the actual impacts and actions they take after a flood, for example reporting on the congestion of evacuation shelters and the actual flooding situation.

[1203] Output: Feedback data sent

[1204] server

[1205] Input: Collected feedback data

[1206] The server collects and analyzes the feedback data and uses it to improve the prediction model and response plan. For example, based on feedback that an evacuation site was overcrowded, it will suggest a different evacuation site for next time.

[1207] Output: Improved forecast model and evacuation plan

[1208] In this way, this system predicts river flood risks with high accuracy and in real time through step-by-step processing, providing appropriate information and supporting evacuation planning.

[1209] (Application example 1)

[1210] 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."

[1211] In recent years, climate change has led to an increase in extreme weather events, increasing the risk of river flooding. However, existing systems have difficulty accurately predicting flood risk in real time and responding quickly. Furthermore, evacuation plans and evacuation route suggestions are not provided properly, which can lead to problems in ensuring the safety of residents. Additionally, there is a lack of a feedback function to continuously improve the prediction model based on collected data.

[1212] 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.

[1213] In this invention, the server includes: means for acquiring river water level data, precipitation data, and topographical data from a weather database; means for converting the acquired data into a unified format; and means for inputting the converted data into a generative AI model to predict flood risk. This enables real-time, highly accurate prediction of flood risk and prompt response. The server also includes means for performing a risk assessment based on the predicted flood risk and visually displaying the risk level on a map; and means for notifying local government disaster prevention personnel and residents of the risk assessment results. This enables risk information to be provided promptly to residents and local governments. The server also includes means for proposing an evacuation plan and an optimal evacuation route based on the notified risk assessment results, and means for displaying evacuation sites in a specific area based on the user's location information. The server also includes means for collecting feedback from local governments and residents and using it to improve the accuracy of the model, thereby ensuring the safety of residents and continuously improving the prediction model.

[1214] "Weather database" refers to a database that collects, stores, and provides weather information, including river water level data, precipitation data, and weather forecast data.

[1215] "River water level data" refers to data that indicates the height of the water level in a river, and is information necessary for assessing flood risk.

[1216] "Precipitation data" refers to data that indicates the amount of precipitation over a certain period of time in a specific area, and is an important factor in predicting the risk of flooding.

[1217] "Topographical data" refers to data containing information about the topography of an area, which is useful for planning the extent of flood impacts and evacuation routes.

[1218] A "unified format" is a consistent data format for converting data obtained from different data sources into a format that is easy to analyze.

[1219] A "generative AI model" is an artificial intelligence model used to predict flood risk based on historical and current data, and specifically includes deep learning models.

[1220] "Risk assessment" is the process of quantifying and assessing the flood risk of a specific area based on the output of a generative AI model.

[1221] "Visually displaying" means presenting the assessed risk to the user in a visual format such as a map or graph.

[1222] "Notifying" means informing users of the assessed flood risk and related information via email, SMS, smartphone apps, etc.

[1223] An "evacuation plan" is a set of instructions and route guidance designed to enable residents to evacuate safely when the risk of flooding increases.

[1224] The "optimal evacuation route" is a route that suggests the safest and quickest evacuation route based on the user's current location.

[1225] "Feedback" refers to residents and local governments sending information back to the system about actual evacuation behavior and the impact of flooding, which is used to improve the accuracy of the model.

[1226] The system proposed in this invention aims to predict river flood risk in real time and provide prompt and appropriate information to residents and local governments. To achieve this, it is necessary for the server, terminals, and users to work together.

[1227] server

[1228] Data collection

[1229] The server obtains river water level data, precipitation data, and topographical data from a meteorological database, which uses a database that collects, stores, and provides meteorological information.

[1230] Converting acquired data into a unified format: Data obtained from different data sources is converted into a unified data format to make it easier to analyze.

[1231] Data analysis

[1232] The data converted into a unified format is then input into a generative AI model, an artificial intelligence model used to predict flood risk based on past and current data. A deep learning model is typically used.

[1233] Generative AI models are implemented using advanced machine learning frameworks such as TensorFlow and Keras, which denoise data and filter outliers to generate high-quality input data.

[1234] Risk Assessment and Notification

[1235] The level of flood risk is assessed based on the output of the generative AI model. The risk assessment quantifies the flood risk of a specific area and displays it according to the risk level.

[1236] The results of the risk assessment are visually displayed on a map and are notified in real time to local government disaster prevention officials and residents via email, SMS, and a dedicated smartphone app.

[1237] Evacuation Planning and Response

[1238] Based on the risk assessment results, the system proposes specific evacuation plans and optimal evacuation routes. Based on the user's location information, it displays evacuation sites in specific areas.

[1239] Evacuation plans can be customized to suit the needs of each local government.

[1240] Terminal

[1241] The terminal (smartphone or PC) functions as a device for checking the received flood risk information and the proposed evacuation plan. For example, using a smartphone app, users can intuitively check evacuation locations and evacuation routes on a map.

[1242] User

[1243] Users act quickly based on the risk information provided. After an evacuation or flood occurs, they provide feedback to the system about the actual impact and actions taken. This feedback gathers data that can be used to improve predictive models and response plans.

[1244] Specific examples

[1245] Real-time flood forecasting and notifications

[1246] When residents press the "Check current flood risk" button, the latest flood risk map is retrieved from the server and displayed on the screen. When residents select "Show nearest evacuation site," the smartphone app identifies the user's current location and displays the optimal evacuation site and route on the map.

[1247] Prompt Sentence Examples

[1248] 1. "Check your current flood risk"

[1249] 2. "Display the nearest evacuation site"

[1250] Through these functions, this system aims to ensure the safety of residents and provide rapid evacuation support, while also improving the accuracy of the model.

[1251] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1252] Step 1:

[1253] The server retrieves river water level data, precipitation data, and topography data from the weather database. This includes specific operations such as collecting data in real time using APIs. The input data includes information on water level, precipitation, and topography, and the output is raw data retrieved in bulk.

[1254] Step 2:

[1255] The server converts the acquired data into a unified format. Specifically, it converts data in different formats into a standardized format and processes it to make it easier to analyze. The input data is raw data, and the output data is data converted into a unified format. This process includes data normalization and unit unification.

[1256] Step 3:

[1257] The server inputs the converted data into a unified format into a generative AI model to predict flood risk. The generative AI model is implemented using TensorFlow and Keras, and predicts risk based on past and current data. The input data is standardized data, and the output data is the predicted flood risk level. This step also removes noise from the data and filters out outliers.

[1258] Step 4:

[1259] The server performs a risk assessment based on the output of the generative AI model, quantifying and assessing the flood risk of a specific area. The resulting risk level is visually displayed on a map. The input data is the predicted flood risk level, and the output data is a visually displayed risk map. Specifically, the risk level is displayed color-coded on the map application.

[1260] Step 5:

[1261] The server notifies local government disaster prevention personnel and residents of the risk assessment results. Specifically, notifications are sent via email, SMS, and a dedicated smartphone app. The input data is the risk assessment results, and the output data is the notification message. This process also includes prioritizing notifications and selecting recipients.

[1262] Step 6:

[1263] The server proposes an evacuation plan and optimal evacuation route based on the notified risk assessment results. It displays evacuation locations in a specific area based on the user's location information. The input data are the risk assessment results and the user's location information, and the output data are the evacuation plan and evacuation route. Specifically, it displays the optimal evacuation route on a map application.

[1264] Step 7:

[1265] The device checks the received flood risk information and proposed evacuation plan and presents it to the user. The user uses a smartphone app to intuitively check evacuation locations and routes on a map. The input data is the notified risk information and evacuation plan, and the output data is the displayed evacuation map.

[1266] Step 8:

[1267] Users send feedback to the system about the actual impacts and actions taken after an evacuation or flood occurs. The input data is feedback information about evacuation actions and impacts, and the output data is the data stored on the server that received it. In this step, the prediction model and response plan are improved based on the collected feedback.

[1268] In this way, each step works in coordination to predict river flood risks with high accuracy and enable appropriate responses in real time.

[1269] 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.

[1270] This invention combines a system that obtains river water level data, precipitation data, and topographical data from meteorological and satellite databases, and uses this data to predict flood risk with high accuracy, with an emotion engine that recognizes the user's emotions. This system operates in cooperation with a server, terminals, and users.

[1271] Specific Embodiments of the System

[1272] Data collection and analysis

[1273] server

[1274] Obtain the latest river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases. This includes obtaining data from the Meteorological Agency using an API key and analyzing image data from the satellite database.

[1275] The acquired data is converted into a unified format, missing data is imputed, and noise is filtered out, and the converted data is used as input for generative AI models.

[1276] Flood risk forecasting

[1277] server

[1278] After the pre-processing process is complete, the data is fed into a generative AI model, which uses deep learning algorithms to predict flood risk. The model takes into account historical flood data, current water levels, precipitation, and topographical data to calculate the risk of flooding within the next 48 hours.

[1279] The prediction results are quantified according to the risk level, and a risk assessment is carried out based on this, with the risk level of a specific area being visually displayed on a map using different colors.

[1280] Risk notification and evacuation planning

[1281] server

[1282] Flood risk information is sent to local government disaster prevention officials and residents via email, SMS, and a dedicated app, and includes specific risk levels and evacuation instructions.

[1283] Devices (smartphones and computers of local government employees and residents)

[1284] Residents can receive the information and check the risk level. Detailed evacuation plans and evacuation routes are also provided at the same time, allowing them to take action quickly.

[1285] Application of Emotion Engine

[1286] server

[1287] The emotion engine analyzes the user's emotional state when receiving and responding to notifications. For example, it analyzes emails, in-app reaction data, and text data obtained from feedback forms to recognize the user's emotions in real time.

[1288] The notification language and content are adjusted based on the user's emotional state. Specifically, if a user is in a panic despite the urgency of the situation, the notification language will be changed to one that helps the user remain calm.

[1289] Feedback and Improvements

[1290] User

[1291] After the evacuation and flood events, provide feedback on emotional states and actual evacuation behavior, for example, recording fear and difficulties felt during evacuation and the effectiveness of evacuation routes.

[1292] server

[1293] The collected feedback data will be analyzed to improve forecasting models and evacuation plans, and sentiment data will also be analyzed to improve the quality of notifications and plans for the next disaster response.

[1294] Specific examples

[1295] Example 1: Flood risk prediction and emotional response

[1296] scene

[1297] A heavy rain warning is issued, causing river water levels to rise rapidly. The server retrieves the latest water level and precipitation data from the weather database and uses a deep learning model to predict flood risk.

[1298] server

[1299] Determine whether a particular area is at high risk and assess the risk level.

[1300] The emotion engine analyzed that some users had panicked during previous evacuations, and this time the notification sent a message encouraging those users to remain calm and act quickly.

[1301] Device (resident's smartphone)

[1302] Residents receive notifications, reassuring instructions along with evacuation routes, and take action quickly.

[1303] Example 2: Feedback collection and system improvement

[1304] scene

[1305] Flooding occurs and residents are evacuated. After evacuation, residents provide feedback on their feelings and the evacuation situation.

[1306] User

[1307] In the feedback form, participants provided specific feedback such as, "I panicked during the evacuation, but the notification message was very helpful."

[1308] server

[1309] The feedback and sentiment data will be analyzed and reflected in response measures for the next flood, and the sentiment engine algorithm will be improved to provide more effective risk notifications and evacuation plans.

[1310] In this way, the present invention is a system that combines highly accurate flood risk prediction with user emotional responses, encouraging prompt and appropriate evacuation behavior and ensuring safety. In particular, the introduction of an emotion engine enables flexible responses based on user reactions, further enhancing the effectiveness of disaster prevention measures.

[1311] The processing flow will be explained below.

[1312] Step 1: Data collection

[1313] server

[1314] Connect to meteorological and satellite databases to get real-time river water level data, precipitation data, and topographic data. Use an API key to request data from the meteorological bureau and download the latest data.

[1315] The acquired data is stored in an internal database and data from different data sources is converted into a unified format, for example, water level data is standardized to meters and precipitation data is converted to millimeters.

[1316] Step 2: Data Preprocessing

[1317] server

[1318] The acquired data is subjected to noise removal and outlier filtering. For example, abnormally high or low measurement values ​​are detected and excluded. In addition, if there is missing data, it is supplemented based on past data or surrounding data.

[1319] The data is formatted as time series data so that it can be input into a generative AI model.

[1320] Step 3: Flood risk forecasting

[1321] server

[1322] The formatted data is then fed into a generative AI model, which uses deep learning algorithms to predict flood risk. The generative AI model calculates the probability of flooding within the next 48 hours based on historical flood data, current water levels, precipitation, and topographical data.

[1323] The forecast results are quantified and the flood risk level for each area is evaluated. For example, flood risk levels are displayed on a three-level scale: "low," "medium," and "high."

[1324] Step 4: Risk assessment and visualization

[1325] server

[1326] The predicted flood risk level is visually displayed on a map, with high-risk areas color-coded to provide an intuitive understanding of the level of risk to a particular area.

[1327] Make a list of potentially affected facilities (hospitals, schools, evacuation centers, etc.) and identify where action is needed.

[1328] Step 5: Risk notification and sentiment analysis

[1329] server

[1330] Flood risk information is sent to local government disaster prevention officials and residents via email, SMS, and a dedicated app, and includes specific risk levels and evacuation instructions.

[1331] The emotion engine collects user reaction data when receiving notifications and analyzes their emotional state. For example, it recognizes emotions based on the user's immediate reaction when opening a notification or text input from a feedback form.

[1332] Step 6: Evacuation planning and emotional response

[1333] server

[1334] The content and presentation of notification messages are adjusted based on the user's emotional state analyzed by the emotion engine. For example, if a user is in a panic despite the urgency of the situation, a message encouraging them to remain calm will be sent.

[1335] Devices (smartphones and computers of local government employees and residents)

[1336] Along with the risk information provided, residents can confirm appropriate evacuation plans and routes. For example, residents can visually check specific evacuation locations and routes via a smartphone app.

[1337] Step 7: Evacuation implementation and feedback

[1338] User

[1339] Evacuate promptly based on the notified flood risk information and the proposed evacuation plan. After completing the evacuation, provide feedback on the emotions and evacuation behavior during the evacuation. For example, enter information such as "the fear and difficulty felt during the evacuation" and "the effectiveness of the evacuation route" in the feedback form.

[1340] server

[1341] The collected feedback data will be analyzed to improve prediction models and evacuation plans. In particular, analyzing emotion data at the same time will improve the quality of notifications and plans for the next disaster response.

[1342] As a result, this invention provides a system that integrates highly accurate flood risk prediction with flexible responses based on user emotions, encouraging prompt and appropriate evacuation behavior. The introduction of an emotion engine makes it possible to realize notifications and evacuation plans based on user reactions, further enhancing the effectiveness of disaster prevention measures.

[1343] Example 2

[1344] 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."

[1345] Current flood risk prediction systems often lack real-time capabilities and accuracy, resulting in delays in evacuation notifications and evacuation plan proposals to residents. These systems also lack the ability to respond to users' emotional states, which can lead to panic among residents in emergencies. As a result, appropriate evacuation behavior is not possible, and safety is not ensured. Furthermore, there is also the issue that feedback is not fully utilized, resulting in the time required to improve the models.

[1346] 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.

[1347] In this invention, the server includes a means for acquiring river water level data, precipitation data, and topographical data from a meteorological database and a satellite database; a means for converting the data into a unified format, filling in missing data, and performing noise filtering; and a means for inputting the converted data into a generative AI model to predict flood risk. This enables real-time and highly accurate prediction of flood risk. Furthermore, risk assessments are performed and displayed visually on a map, and an emotion recognition engine is used to analyze the user's emotional state and adjust notification content to encourage appropriate evacuation behavior. Feedback data can be collected and used to improve the accuracy of the model, further enhancing the effectiveness of future disaster responses.

[1348] A "weather database" is a data storage system for collecting, storing, and providing weather information. It is primarily managed and operated by meteorological agencies and related organizations.

[1349] A "satellite database" is a system that stores data collected from artificial satellites for the purpose of Earth observation. It provides a variety of information, including topographical and meteorological data.

[1350] "River water level data" is numerical information measuring the water surface height of a specific river. It is an important element in flood prediction.

[1351] "Precipitation data" is data that numerically indicates the amount of precipitation (rain, snow, fog, etc.) that fell in a specific area within a certain period of time. It is used for weather forecasting and flood prediction.

[1352] "Terrain data" is data that contains information about the shape, height, and structure of the Earth's surface. It is primarily used for mapping and environmental modeling.

[1353] "Missing data" refers to the absence of required data in a dataset. Also known as missing values.

[1354] "Noise filtering" is a process to remove unnecessary information (noise) from data. It is performed to improve the quality of data.

[1355] A "generative AI model" is a computational model designed to generate new data or information using artificial intelligence, often involving deep learning algorithms.

[1356] "Flood risk" refers to the possibility of flooding occurring in a particular area and the extent of its impact. It is used in risk assessment.

[1357] An "emotion recognition engine" is software that analyzes a user's emotional state from input such as text data and voice data, enabling the system to respond based on the user's reaction.

[1358] An "evacuation plan" is a plan that shows routes and methods for safe evacuation in the event of a disaster. It is provided to ensure the safety of residents.

[1359] "Feedback" refers to opinions, impressions, and evaluation information provided by system users. It is used to improve the system.

[1360] A "municipal government" is a local government or related institution that administers a particular area and ensures the welfare and safety of its residents.

[1361] "Residents" refers to people who live in a specific area. They are often users of the system.

[1362] A "deep learning model" is a machine learning model consisting of a multi-layered neural network. It is used for advanced data analysis and prediction.

[1363] The present invention is a system that predicts flood risk using data acquired from a meteorological database and a satellite database, and provides notifications that correspond to the user's emotional state. A specific embodiment of this system is described below.

[1364] Data collection and analysis

[1365] server

[1366] The server obtains the latest river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases. APIs are used to obtain data from the meteorological bureau and satellite databases. For example, API requests are sent using the Python requests library. The obtained data is stored in a database management system such as PostgreSQL or MySQL.

[1367] Data Preprocessing

[1368] server

[1369] The server converts the acquired data into a unified format, imputes missing data, and performs noise filtering. Python's pandas library is used to unify different data formats and impute missing values. Statistical methods such as median are used to improve data quality during this process. After preprocessing, the data is formatted as input for the generative AI model and converted into NumPy arrays or TensorFlow tensors.

[1370] Flood risk forecasting

[1371] server

[1372] The server inputs the preprocessed data into the generative AI model. A deep learning model is built using TensorFlow and PyTorch, and the data is used for prediction. The model takes into account past flood data, current water levels, precipitation, and topographical data to predict flood risk within the next 48 hours. The prediction results are quantified based on risk levels, and the risk level for specific areas is visually displayed on a map. Map libraries such as Google Maps API and Leaflet are used to display the map.

[1373] Risk notification and evacuation planning

[1374] server

[1375] The server then notifies local government disaster prevention officials and residents of flood risk information via email, SMS, or a dedicated app. For example, AWS SNS (Simple Notification Service) can be used to send SMS notifications.

[1376] Terminal

[1377] The device receives the information and checks the risk level. A push notification is sent to the smartphone app, and by tapping the notification, detailed information is displayed. Evacuation plans and evacuation routes are also presented, and the optimal evacuation route can be displayed using the Google Maps API.

[1378] Emotion recognition and notification adjustment

[1379] server

[1380] The server uses an emotion engine to analyze the emotional state of users receiving notifications in real time. For example, it uses the Natural Language Toolkit (NLTK) and BERT models to analyze notification emails and app response data. Based on the analysis results, it adjusts the wording and content of notifications. For example, it sends an encouraging message to a panicked user to help them stay calm.

[1381] Feedback and System Improvement

[1382] User

[1383] Users provide feedback after an evacuation or flood occurs. For example, they can write in the app's feedback form that they felt scared during the evacuation but found the notifications helpful.

[1384] server

[1385] The server analyzes the collected feedback data to help improve the predictive models and evacuation plans. It uses data mining techniques and sentiment analysis algorithms to analyze the feedback, which will enable it to provide more effective risk notifications and evacuation plans in the next disaster response.

[1386] Specific examples

[1387] The following is a specific example of how this system works.

[1388] Scene 1: Processing in the event of a heavy rain warning

[1389] The server uses an API to retrieve the latest precipitation and river water level data from the weather database.

[1390] Convert the acquired data into a unified format and fill in any missing data.

[1391] The preprocessed data is input into a generative AI model to predict flood risk.

[1392] High-risk areas are identified and marked in red on the map.

[1393] Scene 2: Notifications and Emotion Recognition

[1394] The server sends an SMS to users who live in high-risk areas, for example, "Your area is at high risk of flooding. Please begin evacuation."

[1395] The user views the SMS on their smartphone and checks the evacuation route.

[1396] The server uses an emotion engine to analyze the user's reaction and detect whether they are in a panic state.

[1397] The server sends a calming message to panicked users, for example, "Remain calm and evacuate. You are safe."

[1398] Based on these examples, the system can ensure the safety of residents by providing highly accurate predictions of flood risk and prompt evacuation notifications that respond to users' emotions.

[1399] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1400] Step 1: Data collection

[1401] server

[1402] The server retrieves river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases, and sends API requests using the Python requests library.

[1403] Input: API request (weather database and satellite database)

[1404] Output: Acquired river water level data, precipitation data, topographical data

[1405] The acquired data is stored in a database management system such as PostgreSQL or MySQL.

[1406] Step 2: Data Preprocessing

[1407] server

[1408] The data retrieved by the server is converted into a unified format. The data is cleaned and formatted consistently using the Python pandas library.

[1409] The server imputes missing data and performs noise filtering, for example by imputing missing values ​​with the median to ensure data consistency.

[1410] Input: Acquired river water level data, precipitation data, topographical data

[1411] Output: Preprocessed data (unified format, missing data imputation, noise filtering)

[1412] The preprocessed data is converted into NumPy arrays or TensorFlow tensors.

[1413] Step 3: Flood risk forecasting

[1414] server

[1415] The server inputs the preprocessed data into a generative AI model, which uses a deep learning model built using TensorFlow or PyTorch.

[1416] The model takes into account historical flood data, current water levels, rainfall and topographical data to predict flood risk within the next 48 hours.

[1417] Input: Preprocessed data (NumPy arrays or TensorFlow tensors)

[1418] Output: Flood risk rating for each region (ranging from 0 to 1)

[1419] High-risk areas are color-coded on a map, and the map is displayed using Google Maps API and Leaflet.

[1420] Step 4: Risk notification

[1421] server

[1422] The server notifies local government disaster prevention officials and residents of flood risk information via email, SMS, and a dedicated app.

[1423] Send SMS notifications using AWS SNS (Simple Notification Service).

[1424] Input: Flood risk assessment results, notification messages based on risk level

[1425] Output: Notifications (email, SMS) sent to local government disaster prevention officials and residents

[1426] Terminal

[1427] The device receives the information and checks the risk level. A push notification is sent to the smartphone app, displaying detailed information.

[1428] The device will present the user with an evacuation plan and evacuation route, using the Google Maps API to display safe evacuation routes.

[1429] Input: Risk notification message, evacuation plan information

[1430] Output: Notifications and evacuation routes displayed to the user

[1431] Step 5: Emotion recognition and notification adjustment

[1432] server

[1433] The server uses an emotion engine to analyze the emotional state of the user receiving the notification information in real time. It analyzes the text data using NLTK (Natural Language Toolkit) and the BERT model.

[1434] Tailor the wording and content of notifications based on the user's emotional state: if a user is panicking, send them a notification that helps them stay calm.

[1435] Input: User emotion data (text, reaction data)

[1436] Output: Adjusted notification message

[1437] Example: A message encouraging people to remain calm and evacuate.

[1438] Step 6: Gather feedback and improve the system

[1439] User

[1440] After an evacuation or flood occurs, users provide feedback on their emotional state and actual evacuation behavior. For example, they can write in the in-app feedback form, "I was able to make appropriate decisions quickly when evacuating."

[1441] Input: Feedback form input (text data)

[1442] Output: Feedback data provided

[1443] server

[1444] The server analyzes the collected feedback data to help improve prediction models and evacuation plans. It uses data mining techniques and sentiment analysis algorithms to analyze the feedback.

[1445] Input: Feedback data (emotional information, behavioral records)

[1446] Output: Improved models and evacuation plans

[1447] Improved emotion engine algorithms will provide more effective risk notifications and evacuation plans for the next flood.

[1448] This allows the system to achieve highly accurate flood risk prediction and user response at each step, promoting safe evacuation behavior.

[1449] (Application example 2)

[1450] 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."

[1451] Conventional flood risk prediction systems primarily use meteorological and satellite data to predict flood risks and notify local governments and residents. However, they do not take into account the emotional state of passengers, particularly in autonomous vehicles, which can lead to confusion and anxiety caused by emotions such as panic. Furthermore, they lack the ability to automatically avoid areas with a high risk of flooding, which can make it difficult to fully ensure passenger safety. A system that can effectively solve these issues is needed.

[1452] 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.

[1453] In this invention, the server includes means for acquiring river water level data, precipitation data, and topographical data from a meteorological database and a satellite database, means for converting the acquired data into a unified format, and means for inputting the converted data into a generative AI model to predict flood risk, thereby enabling highly accurate flood risk prediction and risk assessment.

[1454] The server also includes a means for performing a risk assessment based on the predicted flood risk and visually displaying the risk level on a map, a means for notifying local government disaster prevention personnel and residents of the risk assessment results, a means for proposing an evacuation plan based on the notified risk assessment results, a means for collecting feedback from each local government and residents and using it to improve the accuracy of the model, a means for recognizing the emotional state of passengers, a means for customizing the content of the notification based on the emotional state, and a means for the autonomous vehicle to automatically select a route that avoids the flood risk. This enables flexible notifications according to the emotional state of passengers and enables the autonomous vehicle to select a safe route.

[1455] A "weather database" is a database that includes various weather data (e.g., precipitation data, temperature data, wind speed data, etc.).

[1456] A "satellite database" is a database that includes topographical data and water level data obtained from artificial satellites.

[1457] "River water level data" is data that indicates information on the water level in a specific river basin.

[1458] "Precipitation data" is data that indicates information about the amount of precipitation in a specific area.

[1459] "Topography data" refers to data that indicates information such as the topography and altitude of a specific area.

[1460] A "unified format" is a format for converting data obtained from different data sources into a consistent format.

[1461] A "generative AI model" is a machine learning model that makes predictions and classifications based on data.

[1462] "Flood risk" is an indicator of the likelihood of flooding occurring in a particular area within a certain period of time.

[1463] "Risk assessment" is the process of determining danger based on acquired data and indicating the risk level using numbers, colors, etc.

[1464] A "local government disaster prevention officer" is an official responsible for disaster prevention measures in a local government.

[1465] "Residents" refers to people who live in a particular area.

[1466] An "evacuation plan" is a plan for safely evacuating in the event of a disaster such as a flood.

[1467] "Feedback" refers to information that collects opinions and impressions from users and is used to improve the system.

[1468] "Passenger emotional state" is an index that indicates the emotional state that passengers are feeling (e.g., relief, anxiety, panic, etc.).

[1469] "Customizing notification content" means changing the content of a signal or message depending on the emotional state of the recipient.

[1470] An "autonomously selected route" is a safe route that an autonomous vehicle selects to avoid hazards such as flooding.

[1471] This invention is a system for predicting flood risks and safely operating autonomous vehicles. This system obtains river water level data, precipitation data, and topographical data from meteorological and satellite databases, and predicts flood risks using a generative AI model based on this data. Furthermore, the system is characterized by its ability to recognize passenger emotions and customize notification content.

[1472] Data collection and analysis

[1473] server

[1474] Obtain the latest river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases. This includes obtaining data from the Meteorological Agency using an API key and analyzing image data from the satellite database.

[1475] The acquired data is converted into a unified format, missing data is imputed, and noise is filtered out, and the converted data is used as input for generative AI models.

[1476] Flood risk forecasting and risk assessment

[1477] server

[1478] After the pre-processing process is complete, the data is fed into a generative AI model, which uses deep learning algorithms to predict flood risk. The model takes into account historical flood data, current water levels, precipitation, and topographical data to calculate the risk of flooding within the next 48 hours.

[1479] The prediction results are quantified according to the risk level, and a risk assessment is carried out based on this, with the risk level of a specific area being visually displayed on a map using different colors.

[1480] Risk notification and evacuation planning

[1481] server

[1482] Flood risk information is sent to local government disaster prevention officials and residents via email, SMS, and a dedicated app, and includes specific risk levels and evacuation instructions.

[1483] Devices (smartphones and personal computers of local government employees and residents)

[1484] Residents can receive the information and check the risk level. Detailed evacuation plans and evacuation routes are also provided at the same time, allowing them to take action quickly.

[1485] Route selection for autonomous vehicles

[1486] server

[1487] This system suggests safe routes for autonomous vehicles to avoid areas at high risk of flooding, allowing them to operate on routes with lower flood risk.

[1488] Terminal (on-board system for autonomous vehicles)

[1489] The in-vehicle system automatically selects the route for the autonomous vehicle based on the proposed route, and the route information is communicated to passengers via the in-vehicle display and audio system.

[1490] Application of Emotion Engine

[1491] server

[1492] Using an emotion engine, the system analyzes passengers' facial expressions using an in-car camera and recognizes their emotional state in real time. If a passenger is panicking, it will customize the notification content to match their emotion and display a reassuring message.

[1493] Terminal (on-board system for autonomous vehicles)

[1494] Customized notifications based on emotion recognition are displayed on in-car displays to encourage passengers to take appropriate action.

[1495] Feedback and model improvement

[1496] User

[1497] After the evacuation or flood occurs, feedback is provided on emotional state and actual evacuation behavior through a feedback form.

[1498] server

[1499] The feedback data will be analyzed and the results will be reflected in improvements to forecasting models and evacuation plans. Sentiment data will also be analyzed at the same time to improve the quality of notifications and plans for the next disaster response.

[1500] Examples of specific examples and prompts

[1501] Example scene:

[1502] When an autonomous vehicle is operating in heavy rain, the on-board system detects a flood risk area, analyzes the passengers' facial expressions, and selects an appropriate evacuation route. The in-car display shows, "A flood warning has been issued. Please remain calm and follow a safe route."

[1503] Example prompt sentence:

[1504] "Heavy rain warnings are issued and autonomous vehicles detect flood risk areas. If passengers start to panic, choose safe routes and display reassuring messages."

[1505] In this way, this invention is a system that combines highly accurate flood risk prediction with passenger emotion recognition, encouraging swift and appropriate evacuation behavior and ensuring safety. In particular, the introduction of an emotion engine enables flexible responses based on passenger reactions, further enhancing the effectiveness of disaster prevention measures.

[1506] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1507] Step 1:

[1508] The server retrieves river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases. The API key and parameters for the data to be retrieved are used as input. This outputs the latest meteorological and topographical data.

[1509] Step 2:

[1510] The server converts the acquired data into a unified format. The acquired river water level data, precipitation data, and topographical data are used as input. The data conversion process complements missing data and filters noise, preparing the data in a format that can be input into the generative AI model. The output is preprocessed data in a unified format.

[1511] Step 3:

[1512] The server inputs the preprocessed data in a unified format into a generative AI model to predict flood risk. The preprocessed data set is used as input. The generative AI model uses a deep learning algorithm to analyze past flood data, which outputs a numerical value for the risk of flooding within the next 48 hours.

[1513] Step 4:

[1514] The server performs risk assessment based on the predicted flood risk. Flood risk prediction data from the generative AI model is used as input. The risk assessment process quantifies the risk level of a specific area and visually displays it in color on a map. The output is risk level information on the map.

[1515] Step 5:

[1516] The server notifies local government disaster prevention personnel and residents of the risk assessment results. The inputs are the risk assessment results and notification information (email address, phone number, etc.). The notification process sends information including the risk level for a specific area and evacuation instructions via email, SMS, or a dedicated app. The output is a notification message sent to local government officials and residents.

[1517] Step 6:

[1518] The terminal (smartphone or personal computer of local government officials or residents) receives the notified information and checks the risk level and evacuation plan. The received notification message is used as input. The risk level and evacuation route are displayed on the terminal, urging the user to take prompt action. The output is display information that serves as a guide for the user's actions.

[1519] Step 7:

[1520] The server analyzes the passenger's facial expressions using the in-car camera and recognizes their emotional state in real time. It uses image data acquired from the in-car camera as input. It uses an emotion recognition engine to analyze the facial expression data and determine the passenger's emotional state. The output is passenger's emotional state information.

[1521] Step 8:

[1522] The server customizes the notification content based on the passenger's emotional state. The inputs are the emotional state information and the risk assessment results. In the notification customization process, if the passenger is in a panic state, the message is changed to one that gives a sense of security. The output is the customized notification message.

[1523] Step 9:

[1524] The server proposes safe routes to the autonomous vehicle to avoid areas with high flood risk. It uses the risk assessment results as input. It calculates a safe path using a route selection algorithm and sends it to the autonomous vehicle. The output is the route information for the autonomous vehicle.

[1525] Step 10:

[1526] The terminal (the autonomous vehicle's on-board system) sets the autonomous driving route based on the proposed route. Route information from the server is used as input. The route information is notified to passengers via the in-car display and audio system, ensuring safe driving. The output is the autonomous vehicle's driving route.

[1527] Step 11:

[1528] After an evacuation or flood occurs, users provide feedback about their emotional state and actual evacuation behavior through a feedback form. The evacuation experience and emotional state are used as input. The feedback data is recorded and sent to the server. The output is the feedback data sent to the server.

[1529] Step 12:

[1530] The server analyzes the collected feedback data to help improve the forecasting model and evacuation plan. It uses the feedback data as input. It uses analytical algorithms to analyze the data and incorporate it into the model and plan. The output is an improved forecasting model and evacuation plan.

[1531] 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.

[1532] 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.

[1533] 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.

[1534] [Fourth embodiment]

[1535] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1536] 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.

[1537] 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).

[1538] 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.

[1539] 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.

[1540] 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).

[1541] 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.

[1542] 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.

[1543] 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.

[1544] 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.

[1545] 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.

[1546] 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.

[1547] 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."

[1548] The present invention is a system that predicts river flood risk with high accuracy and provides risk information to local governments and residents in real time. This system operates in cooperation with a server, terminals, and users.

[1549] Specific Embodiments of the System

[1550] Data collection

[1551] server

[1552] River water level data, precipitation data, and topographical data are obtained from meteorological databases, which are collected in real time through collaboration with meteorological agencies, satellite databases, and on-site observation stations.

[1553] Converting acquired data into a unified format to facilitate subsequent analysis and processing, for example, integrating water level, precipitation, and topographic data from different sources into a consistent format.

[1554] Data analysis

[1555] server

[1556] The data converted into a unified format is then fed into a generative AI model, which uses advanced machine learning algorithms such as deep learning to combine historical and current flood data to predict flood risk.

[1557] Data preprocessing is performed to remove noise from each data and filter out outliers, ensuring high quality of input data for the predictive model.

[1558] Risk Assessment and Notification

[1559] server

[1560] The output of the generative AI model is used to assess the level of flood risk. This quantifies the flood risk for a specific area and displays it according to the risk level. This display is then displayed on a map as a color-coded risk map.

[1561] The system notifies local government disaster prevention officials and residents of risk assessment results in real time, providing information quickly via email, SMS, and a dedicated app.

[1562] Evacuation Planning and Response

[1563] server

[1564] Based on the results of the risk assessment, specific evacuation plans and infrastructure operation plans are generated, which are customized for each local government.

[1565] Devices (smartphones and computers of local government employees and residents)

[1566] Check the received flood risk information and suggested evacuation plans. For example, use a smartphone app to intuitively understand evacuation locations and routes on a map.

[1567] User

[1568] After an evacuation or flood occurs, feedback on actual impacts and actions is sent to the system, collecting data that can be used to improve forecasting models and response plans.

[1569] Specific examples

[1570] Example 1: Real-time flood forecasting and notification

[1571] scene

[1572] In a situation where precipitation is increasing rapidly and river water levels are rising rapidly, the server retrieves the latest data from the weather database.

[1573] server

[1574] Using a generative AI model, the risk of flooding within the next 48 hours is predicted and specific urban areas are identified as being at high risk.

[1575] Device (local government employee's smartphone)

[1576] Receive notifications from the server and check detailed risk maps and evacuation plans.

[1577] User

[1578] Based on risk information, evacuation orders are issued to residents, who then begin evacuating promptly in accordance with the instructions.

[1579] Example 2: Feedback and Improvement

[1580] scene

[1581] It was reported that after the flood occurred, residents were able to evacuate safely and ensure their safety, and the actual impact was low.

[1582] User

[1583] Feedback is sent to the system about the actual impact and evacuation actions.

[1584] server

[1585] Feedback data is collected and analyzed, and reflected in improvements to prediction models and proposed algorithms.

[1586] In this way, the present invention provides a consistent system that covers everything from flood risk prediction to notification and countermeasure proposals, enabling real-time, highly accurate flood response. In particular, the series of processes, including data collection, analysis using generative AI models, risk assessment, notification, and feedback, work together to achieve a rapid and appropriate response.

[1587] The processing flow will be explained below.

[1588] Step 1: Data collection

[1589] server

[1590] It connects to meteorological databases, satellite databases, and local observation stations to obtain the latest river water level, precipitation, and topographical data. Specifically, it uses an API key to send requests to the Meteorological Bureau's services and downloads the data in real time.

[1591] The acquired data is stored in a database and the different data formats from each data source are converted into a unified format, for example, water level data is unified to meters and precipitation data is unified to millimeters.

[1592] Step 2: Data Preprocessing

[1593] server

[1594] The acquired data is subjected to noise removal and outlier filtering. For example, abnormally high or low values ​​are detected and excluded. In addition, if there is missing data, it is supplemented based on past data or surrounding data.

[1595] The data is formatted as time series data or geographic information system (GIS) data, making it suitable for input into generative AI models.

[1596] Step 3: Flood risk forecasting

[1597] server

[1598] The formatted data is then fed into a generative AI model, which uses deep learning algorithms to predict flood risk by combining historical flood data with current weather data to calculate the probability of a flood occurring within the next 48 hours.

[1599] The forecast results are quantified and the flood risk level is evaluated. For example, the flood risk level is displayed on a three-level scale: "low," "medium," or "high."

[1600] Step 4: Risk assessment and visualization

[1601] server

[1602] The predicted flood risk level is visually displayed on a map, with high-risk areas color-coded to provide an intuitive understanding of the level of risk to a particular area.

[1603] Make a list of potentially affected facilities (hospitals, schools, evacuation centers, etc.) and identify where action is needed.

[1604] Step 5: Notification

[1605] server

[1606] Notify local government disaster prevention officials and residents of flood risk information by sending emails and SMS containing the risk level of areas requiring action and recommended action plans (evacuation locations and evacuation routes).

[1607] Devices (smartphones and computers of local government employees and residents)

[1608] Residents can check the received notifications and understand the specific measures to take. Residents can use the smartphone app to check real-time risk information and evacuation plans.

[1609] Step 6: Implement your evacuation plan

[1610] User

[1611] Evacuations are carried out promptly based on the notified flood risk information and proposed evacuation plans. Local government officials issue evacuation orders, prepare evacuation shelters, and direct traffic. Residents also follow the evacuation routes provided to them and evacuate to safe locations.

[1612] Step 7: Gather feedback

[1613] User

[1614] After a flood occurs, feedback is provided to the system regarding the actual impact and evacuation behavior, including the time required for evacuation, the effectiveness of the evacuation plan, and the extent of the damage.

[1615] server

[1616] The collected feedback data will be analyzed to improve forecasting models and evacuation plans, allowing for more accurate and rapid responses the next time floods occur.

[1617] Example 1

[1618] 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."

[1619] To predict river flood risks in real time with high accuracy based on meteorological data, promptly and appropriately notify local government disaster prevention officials and residents of risk information, and provide appropriate evacuation plans.

[1620] 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.

[1621] In this invention, the server includes means for acquiring river water level data, precipitation data, and topographical data from a weather database, means for converting the acquired data into a unified format, means for preprocessing the converted data by removing noise and filtering outliers, means for inputting the preprocessed data into a generative AI model to predict flood risk, means for conducting a risk assessment based on the predicted flood risk and visually displaying the risk level on a map, means for notifying disaster prevention personnel in the local government and residents of the results of the risk assessment, means for proposing evacuation plans based on the notified risk assessment results, and means for collecting feedback from each local government and residents and using it to improve the accuracy of the model. This enables highly accurate real-time prediction of river flood risk and the provision of prompt and appropriate risk notifications and evacuation plans.

[1622] A "weather database" is a database for storing and managing weather information, and provides data on weather such as temperature, precipitation, wind speed, wind direction, and atmospheric pressure.

[1623] "River water level data" refers to data that indicates the water level at a specific point in a river, and is used to predict river conditions and flood risk.

[1624] "Precipitation data" is data that numerically represents the amount of precipitation, such as rain or snow, that fell at a specific location within a certain period of time.

[1625] "Topographic data" refers to data that expresses geographical features and the shape of the terrain as numerical values ​​or images, and includes information such as the height of the earth's surface, undulations, and land use.

[1626] "Unified format" means converting data obtained from different sources into a consistent format and making it conform to a common standard.

[1627] "Noise removal" is a process for removing unnecessary information and outliers in data analysis and processing to improve data accuracy.

[1628] "Outlier filtering" is a process of detecting, removing, or correcting values ​​that are outside the normal range or that are unreasonable from a business perspective, contained in a data set.

[1629] "Preprocessing" refers to the preparatory steps taken to prepare data for analysis, including noise removal and filtering of outliers.

[1630] A "generative AI model" is an algorithmic model that uses machine learning and deep learning to make predictions and classifications from data.

[1631] "Risk assessment" is the evaluation and judgment of the degree and impact of predicted risks in numerical and visual form.

[1632] A "risk level" is an indicator that shows the extent to which a particular risk exists, and is usually expressed as a number or color.

[1633] "Notifying" refers to the act of transmitting specific information to a designated recipient, and can be done via email, SMS, a dedicated app, etc.

[1634] An "evacuation plan" is a plan that outlines specific procedures and routes for evacuation in the event of a danger.

[1635] "Feedback" refers to information that collects evaluations and comments regarding the operation and results of the system and is used to help improve it in the future.

[1636] The present invention is implemented as a system that predicts river flood risk with high accuracy and provides risk information to local governments and residents in real time. This system operates in cooperation with a server, terminals, and users.

[1637] Data collection

[1638] server

[1639] The server uses APIs to obtain river water level data, precipitation data, and topographical data in real time from meteorological databases, satellite databases, and local observation stations. Specifically, the server calls the meteorological database API to obtain the latest precipitation data.

[1640] Data integration

[1641] server

[1642] It converts acquired data of different formats into a unified format, generating a consistent dataset that allows subsequent processing to be performed efficiently and effectively.

[1643] Data Preprocessing

[1644] server

[1645] The server then removes noise and filters out outliers from the data converted into a unified format. Specifically, it complements missing values, normalizes the data, and detects and removes abnormally high water level data.

[1646] Flood risk prediction

[1647] server

[1648] The preprocessed data is then fed into a generative AI model, which uses advanced machine learning algorithms such as deep learning to combine historical and current flood data to predict flood risk. For example, future rainfall and current water level data are fed into the model to predict the risk of flooding within 48 hours.

[1649] Risk Assessment and Notification

[1650] server

[1651] The output of the generated AI model is analyzed to assess the level of flood risk, which then calculates a risk index for a specific area and generates a risk map that displays different colors on a map according to the risk level.

[1652] Local government disaster prevention officials and residents will be notified of the risk assessment results in real time via email, SMS, and a dedicated app, along with a risk map.

[1653] Generate evacuation plans

[1654] server

[1655] Based on the flood risk assessment results, the server generates specific evacuation plans customized for each municipality, for example, providing residents in high-risk areas with a list of evacuation sites and optimal evacuation routes.

[1656] Real-time notifications and displays

[1657] Terminal

[1658] The devices of local government officials and residents receive the risk information and evacuation plans sent from the server. For example, the smartphones of local government officials receive emergency notifications and display detailed risk maps and evacuation routes on a dedicated app.

[1659] Collecting and analyzing feedback

[1660] User

[1661] Users provide feedback to the system about the actual impacts and actions they take after a flood, for example, residents reporting on the congestion of evacuation shelters and the actual flooding situation.

[1662] server

[1663] The server analyzes the collected feedback data and uses it to improve the prediction model and response plan. For example, based on feedback that an evacuation site was overcrowded, the server can improve the evacuation plan to suggest a different evacuation site for the next time.

[1664] Prompt Sentence Examples

[1665] "Collect real-time river water level and precipitation data, use generative AI models to predict flood risk, communicate risk assessment results to city officials and residents, and develop evacuation plans."

[1666] The system covers everything from data collection to risk prediction, notification, evacuation plan generation, and feedback collection, enabling highly accurate prediction of river flood risks and supporting rapid response in real time.

[1667] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1668] Step 1: Data collection

[1669] server

[1670] Input: weather database, satellite database, data requests from local observation stations

[1671] The server uses an API to obtain river water level data, precipitation data, and topographical data in real time from meteorological databases, satellite databases, and local observation stations.

[1672] Output: Acquired data (river water level data, precipitation data, topographical data)

[1673] Step 2: Integrate the data

[1674] server

[1675] Input: Acquired river water level data, precipitation data, topographical data

[1676] The server converts the acquired data into a unified format and generates a consistent data set. The server converts and maps the data to harmonize the information obtained from different data sources.

[1677] Output: Uniform format dataset

[1678] Step 3: Preprocessing the data

[1679] server

[1680] Input: Uniform format dataset

[1681] The server then performs noise removal and outlier filtering on the data converted into a unified format, for example, by filling in missing values, normalizing the data, and detecting and excluding abnormally high water level data.

[1682] Output: Preprocessed data

[1683] Step 4: Predict flood risk

[1684] server

[1685] Input: Preprocessed data

[1686] The preprocessed data is input into a generative AI model to predict flood risk. The generative AI model uses advanced machine learning algorithms such as deep learning to combine past flood data with current data to calculate future flood risk. For example, it can predict the likelihood of a flood occurring based on rainfall data within the next 48 hours and current water level data.

[1687] Output: Predicted flood risk data

[1688] Step 5: Risk assessment and notification

[1689] server

[1690] Input: Predicted flood risk data

[1691] The output of the generated AI model is analyzed to assess the level of flood risk. The server calculates a risk index for a specific area and generates a color-coded risk map based on the risk level. The server also sends real-time risk information via email, SMS, or a dedicated app to notify local government disaster prevention officials and residents of the results of the risk assessment.

[1692] Output: Risk assessment report and risk map, notification message

[1693] Step 6: Generate an evacuation plan

[1694] server

[1695] Input: Risk Assessment Report

[1696] Based on the flood risk assessment results, the server generates specific evacuation plans customized for each local government. For example, it provides a list of evacuation sites and optimal evacuation routes for residents in high-risk areas. The evacuation plans are displayed in conjunction with risk maps.

[1697] Output: Evacuation plan

[1698] Step 7: Real-time notifications and displays

[1699] Terminal

[1700] Input: Risk assessment report, risk map, draft evacuation plan

[1701] The devices of local government officials and residents receive the risk information and evacuation plans sent from the server. For example, the smartphones of local government officials receive emergency notifications and intuitively display detailed risk maps and evacuation routes on a dedicated app.

[1702] Output: Risk information and evacuation plan displayed on the terminal

[1703] Step 8: Collect and analyze feedback

[1704] User

[1705] Input: User feedback on the actual evacuation situation and impact

[1706] Users provide feedback to the system about the actual impacts and actions they take after a flood, for example reporting on the congestion of evacuation shelters and the actual flooding situation.

[1707] Output: Feedback data sent

[1708] server

[1709] Input: Collected feedback data

[1710] The server collects and analyzes the feedback data and uses it to improve the prediction model and response plan. For example, based on feedback that an evacuation site was overcrowded, it will suggest a different evacuation site for next time.

[1711] Output: Improved forecast model and evacuation plan

[1712] In this way, this system predicts river flood risks with high accuracy and in real time through step-by-step processing, providing appropriate information and supporting evacuation planning.

[1713] (Application example 1)

[1714] 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."

[1715] In recent years, climate change has led to an increase in extreme weather events, increasing the risk of river flooding. However, existing systems have difficulty accurately predicting flood risk in real time and responding quickly. Furthermore, evacuation plans and evacuation route suggestions are not provided properly, which can lead to problems in ensuring the safety of residents. Additionally, there is a lack of a feedback function to continuously improve the prediction model based on collected data.

[1716] 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.

[1717] In this invention, the server includes: means for acquiring river water level data, precipitation data, and topographical data from a weather database; means for converting the acquired data into a unified format; and means for inputting the converted data into a generative AI model to predict flood risk. This enables real-time, highly accurate prediction of flood risk and prompt response. The server also includes means for performing a risk assessment based on the predicted flood risk and visually displaying the risk level on a map; and means for notifying local government disaster prevention personnel and residents of the risk assessment results. This enables risk information to be provided promptly to residents and local governments. The server also includes means for proposing an evacuation plan and an optimal evacuation route based on the notified risk assessment results, and means for displaying evacuation sites in a specific area based on the user's location information. The server also includes means for collecting feedback from local governments and residents and using it to improve the accuracy of the model, thereby ensuring the safety of residents and continuously improving the prediction model.

[1718] "Weather database" refers to a database that collects, stores, and provides weather information, including river water level data, precipitation data, and weather forecast data.

[1719] "River water level data" refers to data that indicates the height of the water level in a river, and is information necessary for assessing flood risk.

[1720] "Precipitation data" refers to data that indicates the amount of precipitation over a certain period of time in a specific area, and is an important factor in predicting the risk of flooding.

[1721] "Topographical data" refers to data containing information about the topography of an area, which is useful for planning the extent of flood impacts and evacuation routes.

[1722] A "unified format" is a consistent data format for converting data obtained from different data sources into a format that is easy to analyze.

[1723] A "generative AI model" is an artificial intelligence model used to predict flood risk based on historical and current data, and specifically includes deep learning models.

[1724] "Risk assessment" is the process of quantifying and assessing the flood risk of a specific area based on the output of a generative AI model.

[1725] "Visually displaying" means presenting the assessed risk to the user in a visual format such as a map or graph.

[1726] "Notifying" means informing users of the assessed flood risk and related information via email, SMS, smartphone apps, etc.

[1727] An "evacuation plan" is a set of instructions and route guidance designed to enable residents to evacuate safely when the risk of flooding increases.

[1728] The "optimal evacuation route" is a route that suggests the safest and quickest evacuation route based on the user's current location.

[1729] "Feedback" refers to residents and local governments sending information back to the system about actual evacuation behavior and the impact of flooding, which is used to improve the accuracy of the model.

[1730] The system proposed in this invention aims to predict river flood risk in real time and provide prompt and appropriate information to residents and local governments. To achieve this, it is necessary for the server, terminals, and users to work together.

[1731] server

[1732] Data collection

[1733] The server obtains river water level data, precipitation data, and topographical data from a meteorological database, which uses a database that collects, stores, and provides meteorological information.

[1734] Converting acquired data into a unified format: Data obtained from different data sources is converted into a unified data format to make it easier to analyze.

[1735] Data analysis

[1736] The data converted into a unified format is then input into a generative AI model, an artificial intelligence model used to predict flood risk based on past and current data. A deep learning model is typically used.

[1737] Generative AI models are implemented using advanced machine learning frameworks such as TensorFlow and Keras, which denoise data and filter outliers to generate high-quality input data.

[1738] Risk Assessment and Notification

[1739] The level of flood risk is assessed based on the output of the generative AI model. The risk assessment quantifies the flood risk of a specific area and displays it according to the risk level.

[1740] The results of the risk assessment are visually displayed on a map and are notified in real time to local government disaster prevention officials and residents via email, SMS, and a dedicated smartphone app.

[1741] Evacuation Planning and Response

[1742] Based on the risk assessment results, the system proposes specific evacuation plans and optimal evacuation routes. Based on the user's location information, it displays evacuation sites in specific areas.

[1743] Evacuation plans can be customized to suit the needs of each local government.

[1744] Terminal

[1745] The terminal (smartphone or PC) functions as a device for checking the received flood risk information and the proposed evacuation plan. For example, using a smartphone app, users can intuitively check evacuation locations and evacuation routes on a map.

[1746] User

[1747] Users act quickly based on the risk information provided. After an evacuation or flood occurs, they provide feedback to the system about the actual impact and actions taken. This feedback gathers data that can be used to improve predictive models and response plans.

[1748] Specific examples

[1749] Real-time flood forecasting and notifications

[1750] When residents press the "Check current flood risk" button, the latest flood risk map is retrieved from the server and displayed on the screen. When residents select "Show nearest evacuation site," the smartphone app identifies the user's current location and displays the optimal evacuation site and route on the map.

[1751] Prompt Sentence Examples

[1752] 1. "Check your current flood risk"

[1753] 2. "Display the nearest evacuation site"

[1754] Through these functions, this system aims to ensure the safety of residents and provide rapid evacuation support, while also improving the accuracy of the model.

[1755] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1756] Step 1:

[1757] The server retrieves river water level data, precipitation data, and topography data from the weather database. This includes specific operations such as collecting data in real time using APIs. The input data includes information on water level, precipitation, and topography, and the output is raw data retrieved in bulk.

[1758] Step 2:

[1759] The server converts the acquired data into a unified format. Specifically, it converts data in different formats into a standardized format and processes it to make it easier to analyze. The input data is raw data, and the output data is data converted into a unified format. This process includes data normalization and unit unification.

[1760] Step 3:

[1761] The server inputs the converted data into a unified format into a generative AI model to predict flood risk. The generative AI model is implemented using TensorFlow and Keras, and predicts risk based on past and current data. The input data is standardized data, and the output data is the predicted flood risk level. This step also removes noise from the data and filters out outliers.

[1762] Step 4:

[1763] The server performs a risk assessment based on the output of the generative AI model, quantifying and assessing the flood risk of a specific area. The resulting risk level is visually displayed on a map. The input data is the predicted flood risk level, and the output data is a visually displayed risk map. Specifically, the risk level is displayed color-coded on the map application.

[1764] Step 5:

[1765] The server notifies local government disaster prevention personnel and residents of the risk assessment results. Specifically, notifications are sent via email, SMS, and a dedicated smartphone app. The input data is the risk assessment results, and the output data is the notification message. This process also includes prioritizing notifications and selecting recipients.

[1766] Step 6:

[1767] The server proposes an evacuation plan and optimal evacuation route based on the notified risk assessment results. It displays evacuation locations in a specific area based on the user's location information. The input data are the risk assessment results and the user's location information, and the output data are the evacuation plan and evacuation route. Specifically, it displays the optimal evacuation route on a map application.

[1768] Step 7:

[1769] The device checks the received flood risk information and proposed evacuation plan and presents it to the user. The user uses a smartphone app to intuitively check evacuation locations and routes on a map. The input data is the notified risk information and evacuation plan, and the output data is the displayed evacuation map.

[1770] Step 8:

[1771] Users send feedback to the system about the actual impacts and actions taken after an evacuation or flood occurs. The input data is feedback information about evacuation actions and impacts, and the output data is the data stored on the server that received it. In this step, the prediction model and response plan are improved based on the collected feedback.

[1772] In this way, each step works in coordination to predict river flood risks with high accuracy and enable appropriate responses in real time.

[1773] 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.

[1774] This invention combines a system that obtains river water level data, precipitation data, and topographical data from meteorological and satellite databases, and uses this data to predict flood risk with high accuracy, with an emotion engine that recognizes the user's emotions. This system operates in cooperation with a server, terminals, and users.

[1775] Specific Embodiments of the System

[1776] Data collection and analysis

[1777] server

[1778] Obtain the latest river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases. This includes obtaining data from the Meteorological Agency using an API key and analyzing image data from the satellite database.

[1779] The acquired data is converted into a unified format, missing data is imputed, and noise is filtered out, and the converted data is used as input for generative AI models.

[1780] Flood risk forecasting

[1781] server

[1782] After the pre-processing process is complete, the data is fed into a generative AI model, which uses deep learning algorithms to predict flood risk. The model takes into account historical flood data, current water levels, precipitation, and topographical data to calculate the risk of flooding within the next 48 hours.

[1783] The prediction results are quantified according to the risk level, and a risk assessment is carried out based on this, with the risk level of a specific area being visually displayed on a map using different colors.

[1784] Risk notification and evacuation planning

[1785] server

[1786] Flood risk information is sent to local government disaster prevention officials and residents via email, SMS, and a dedicated app, and includes specific risk levels and evacuation instructions.

[1787] Devices (smartphones and computers of local government employees and residents)

[1788] Residents can receive the information and check the risk level. Detailed evacuation plans and evacuation routes are also provided at the same time, allowing them to take action quickly.

[1789] Application of Emotion Engine

[1790] server

[1791] The emotion engine analyzes the user's emotional state when receiving and responding to notifications. For example, it analyzes emails, in-app reaction data, and text data obtained from feedback forms to recognize the user's emotions in real time.

[1792] The notification language and content are adjusted based on the user's emotional state. Specifically, if a user is in a panic despite the urgency of the situation, the notification language will be changed to one that helps the user remain calm.

[1793] Feedback and Improvements

[1794] User

[1795] After the evacuation and flood events, provide feedback on emotional states and actual evacuation behavior, for example, recording fear and difficulties felt during evacuation and the effectiveness of evacuation routes.

[1796] server

[1797] The collected feedback data will be analyzed to improve forecasting models and evacuation plans, and sentiment data will also be analyzed to improve the quality of notifications and plans for the next disaster response.

[1798] Specific examples

[1799] Example 1: Flood risk prediction and emotional response

[1800] scene

[1801] A heavy rain warning is issued, causing river water levels to rise rapidly. The server retrieves the latest water level and precipitation data from the weather database and uses a deep learning model to predict flood risk.

[1802] server

[1803] Determine whether a particular area is at high risk and assess the risk level.

[1804] The emotion engine analyzed that some users had panicked during previous evacuations, and this time the notification sent a message encouraging those users to remain calm and act quickly.

[1805] Device (resident's smartphone)

[1806] Residents receive notifications, reassuring instructions along with evacuation routes, and take action quickly.

[1807] Example 2: Feedback collection and system improvement

[1808] scene

[1809] Flooding occurs and residents are evacuated. After evacuation, residents provide feedback on their feelings and the evacuation situation.

[1810] User

[1811] In the feedback form, participants provided specific feedback such as, "I panicked during the evacuation, but the notification message was very helpful."

[1812] server

[1813] The feedback and sentiment data will be analyzed and reflected in response measures for the next flood, and the sentiment engine algorithm will be improved to provide more effective risk notifications and evacuation plans.

[1814] In this way, the present invention is a system that combines highly accurate flood risk prediction with user emotional responses, encouraging prompt and appropriate evacuation behavior and ensuring safety. In particular, the introduction of an emotion engine enables flexible responses based on user reactions, further enhancing the effectiveness of disaster prevention measures.

[1815] The processing flow will be explained below.

[1816] Step 1: Data collection

[1817] server

[1818] Connect to meteorological and satellite databases to get real-time river water level data, precipitation data, and topographic data. Use an API key to request data from the meteorological bureau and download the latest data.

[1819] The acquired data is stored in an internal database and data from different data sources is converted into a unified format, for example, water level data is standardized to meters and precipitation data is converted to millimeters.

[1820] Step 2: Data Preprocessing

[1821] server

[1822] The acquired data is subjected to noise removal and outlier filtering. For example, abnormally high or low measurement values ​​are detected and excluded. In addition, if there is missing data, it is supplemented based on past data or surrounding data.

[1823] The data is formatted as time series data so that it can be input into a generative AI model.

[1824] Step 3: Flood risk forecasting

[1825] server

[1826] The formatted data is then fed into a generative AI model, which uses deep learning algorithms to predict flood risk. The generative AI model calculates the probability of flooding within the next 48 hours based on historical flood data, current water levels, precipitation, and topographical data.

[1827] The forecast results are quantified and the flood risk level for each area is evaluated. For example, flood risk levels are displayed on a three-level scale: "low," "medium," and "high."

[1828] Step 4: Risk assessment and visualization

[1829] server

[1830] The predicted flood risk level is visually displayed on a map, with high-risk areas color-coded to provide an intuitive understanding of the level of risk to a particular area.

[1831] Make a list of potentially affected facilities (hospitals, schools, evacuation centers, etc.) and identify where action is needed.

[1832] Step 5: Risk notification and sentiment analysis

[1833] server

[1834] Flood risk information is sent to local government disaster prevention officials and residents via email, SMS, and a dedicated app, and includes specific risk levels and evacuation instructions.

[1835] The emotion engine collects user reaction data when receiving notifications and analyzes their emotional state. For example, it recognizes emotions based on the user's immediate reaction when opening a notification or text input from a feedback form.

[1836] Step 6: Evacuation planning and emotional response

[1837] server

[1838] The content and presentation of notification messages are adjusted based on the user's emotional state analyzed by the emotion engine. For example, if a user is in a panic despite the urgency of the situation, a message encouraging them to remain calm will be sent.

[1839] Devices (smartphones and computers of local government employees and residents)

[1840] Along with the risk information provided, residents can confirm appropriate evacuation plans and routes. For example, residents can visually check specific evacuation locations and routes via a smartphone app.

[1841] Step 7: Evacuation implementation and feedback

[1842] User

[1843] Evacuate promptly based on the notified flood risk information and the proposed evacuation plan. After completing the evacuation, provide feedback on the emotions and evacuation behavior during the evacuation. For example, enter information such as "the fear and difficulty felt during the evacuation" and "the effectiveness of the evacuation route" in the feedback form.

[1844] server

[1845] The collected feedback data will be analyzed to improve prediction models and evacuation plans. In particular, analyzing emotion data at the same time will improve the quality of notifications and plans for the next disaster response.

[1846] As a result, this invention provides a system that integrates highly accurate flood risk prediction with flexible responses based on user emotions, encouraging prompt and appropriate evacuation behavior. The introduction of an emotion engine makes it possible to realize notifications and evacuation plans based on user reactions, further enhancing the effectiveness of disaster prevention measures.

[1847] Example 2

[1848] 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."

[1849] Current flood risk prediction systems often lack real-time capabilities and accuracy, resulting in delays in evacuation notifications and evacuation plan proposals to residents. These systems also lack the ability to respond to users' emotional states, which can lead to panic among residents in emergencies. As a result, appropriate evacuation behavior is not possible, and safety is not ensured. Furthermore, there is also the issue that feedback is not fully utilized, resulting in the time required to improve the models.

[1850] 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.

[1851] In this invention, the server includes a means for acquiring river water level data, precipitation data, and topographical data from a meteorological database and a satellite database; a means for converting the data into a unified format, filling in missing data, and performing noise filtering; and a means for inputting the converted data into a generative AI model to predict flood risk. This enables real-time and highly accurate prediction of flood risk. Furthermore, risk assessments are performed and displayed visually on a map, and an emotion recognition engine is used to analyze the user's emotional state and adjust notification content to encourage appropriate evacuation behavior. Feedback data can be collected and used to improve the accuracy of the model, further enhancing the effectiveness of future disaster responses.

[1852] A "weather database" is a data storage system for collecting, storing, and providing weather information. It is primarily managed and operated by meteorological agencies and related organizations.

[1853] A "satellite database" is a system that stores data collected from artificial satellites for the purpose of Earth observation. It provides a variety of information, including topographical and meteorological data.

[1854] "River water level data" is numerical information measuring the water surface height of a specific river. It is an important element in flood prediction.

[1855] "Precipitation data" is data that numerically indicates the amount of precipitation (rain, snow, fog, etc.) that fell in a specific area within a certain period of time. It is used for weather forecasting and flood prediction.

[1856] "Terrain data" is data that contains information about the shape, height, and structure of the Earth's surface. It is primarily used for mapping and environmental modeling.

[1857] "Missing data" refers to the absence of required data in a dataset. Also known as missing values.

[1858] "Noise filtering" is a process to remove unnecessary information (noise) from data. It is performed to improve the quality of data.

[1859] A "generative AI model" is a computational model designed to generate new data or information using artificial intelligence, often involving deep learning algorithms.

[1860] "Flood risk" refers to the possibility of flooding occurring in a particular area and the extent of its impact. It is used in risk assessment.

[1861] An "emotion recognition engine" is software that analyzes a user's emotional state from input such as text data and voice data, enabling the system to respond based on the user's reaction.

[1862] An "evacuation plan" is a plan that shows routes and methods for safe evacuation in the event of a disaster. It is provided to ensure the safety of residents.

[1863] "Feedback" refers to opinions, impressions, and evaluation information provided by system users. It is used to improve the system.

[1864] A "municipal government" is a local government or related institution that administers a particular area and ensures the welfare and safety of its residents.

[1865] "Residents" refers to people who live in a specific area. They are often users of the system.

[1866] A "deep learning model" is a machine learning model consisting of a multi-layered neural network. It is used for advanced data analysis and prediction.

[1867] The present invention is a system that predicts flood risk using data acquired from a meteorological database and a satellite database, and provides notifications that correspond to the user's emotional state. A specific embodiment of this system is described below.

[1868] Data collection and analysis

[1869] server

[1870] The server obtains the latest river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases. APIs are used to obtain data from the meteorological bureau and satellite databases. For example, API requests are sent using the Python requests library. The obtained data is stored in a database management system such as PostgreSQL or MySQL.

[1871] Data Preprocessing

[1872] server

[1873] The server converts the acquired data into a unified format, imputes missing data, and performs noise filtering. Python's pandas library is used to unify different data formats and impute missing values. Statistical methods such as median are used to improve data quality during this process. After preprocessing, the data is formatted as input for the generative AI model and converted into NumPy arrays or TensorFlow tensors.

[1874] Flood risk forecasting

[1875] server

[1876] The server inputs the preprocessed data into the generative AI model. A deep learning model is built using TensorFlow and PyTorch, and the data is used for prediction. The model takes into account past flood data, current water levels, precipitation, and topographical data to predict flood risk within the next 48 hours. The prediction results are quantified based on risk levels, and the risk level for specific areas is visually displayed on a map. Map libraries such as Google Maps API and Leaflet are used to display the map.

[1877] Risk notification and evacuation planning

[1878] server

[1879] The server then notifies local government disaster prevention officials and residents of flood risk information via email, SMS, or a dedicated app. For example, AWS SNS (Simple Notification Service) can be used to send SMS notifications.

[1880] Terminal

[1881] The device receives the information and checks the risk level. A push notification is sent to the smartphone app, and by tapping the notification, detailed information is displayed. Evacuation plans and evacuation routes are also presented, and the optimal evacuation route can be displayed using the Google Maps API.

[1882] Emotion recognition and notification adjustment

[1883] server

[1884] The server uses an emotion engine to analyze the emotional state of users receiving notifications in real time. For example, it uses the Natural Language Toolkit (NLTK) and BERT models to analyze notification emails and app response data. Based on the analysis results, it adjusts the wording and content of notifications. For example, it sends an encouraging message to a panicked user to help them stay calm.

[1885] Feedback and System Improvement

[1886] User

[1887] Users provide feedback after an evacuation or flood occurs. For example, they can write in the app's feedback form that they felt scared during the evacuation but found the notifications helpful.

[1888] server

[1889] The server analyzes the collected feedback data to help improve the predictive models and evacuation plans. It uses data mining techniques and sentiment analysis algorithms to analyze the feedback, which will enable it to provide more effective risk notifications and evacuation plans in the next disaster response.

[1890] Specific examples

[1891] The following is a specific example of how this system works.

[1892] Scene 1: Processing in the event of a heavy rain warning

[1893] The server uses an API to retrieve the latest precipitation and river water level data from the weather database.

[1894] Convert the acquired data into a unified format and fill in any missing data.

[1895] The preprocessed data is input into a generative AI model to predict flood risk.

[1896] High-risk areas are identified and marked in red on the map.

[1897] Scene 2: Notifications and Emotion Recognition

[1898] The server sends an SMS to users who live in high-risk areas, for example, "Your area is at high risk of flooding. Please begin evacuation."

[1899] The user views the SMS on their smartphone and checks the evacuation route.

[1900] The server uses an emotion engine to analyze the user's reaction and detect whether they are in a panic state.

[1901] The server sends a calming message to panicked users, for example, "Remain calm and evacuate. You are safe."

[1902] Based on these examples, the system can ensure the safety of residents by providing highly accurate predictions of flood risk and prompt evacuation notifications that respond to users' emotions.

[1903] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1904] Step 1: Data collection

[1905] server

[1906] The server retrieves river water level data, precipitation data, and topographical data in real time from meteorological and satellite databases, and sends API requests using the Python requests library.

[1907] Input: API request (weather database and satellite database)

[1908] Output: Acquired river water level data, precipitation data, topographical data

[1909] The acquired data is stored in a database management system such as PostgreSQL or MySQL.

[1910] Step 2: Data Preprocessing

[1911] server

[1912] The data retrieved by the server is converted into a unified format. The data is cleaned and formatted consistently using the Python pandas library.

[1913] The server imputes missing data and performs noise filtering, for example by imputing missing values ​​with the median to ensure data consistency.

[1914] Input: Acquired river water level data, precipitation data, topographical data

[1915] Output: Preprocessed data (unified format, missing data imputation, noise filtering)

[1916] The preprocessed data is converted into NumPy arrays or TensorFlow tensors.

[1917] Step 3: Flood risk forecasting

[1918] server

[1919] The server inputs the preprocessed data into a generative AI model, which uses a deep learning model built using TensorFlow or PyTorch.

[1920] The model takes into account historical flood data, current water levels, rainfall and topographical data to predict flood risk within the next 48 hours.

[1921] Input: Preprocessed data (NumPy arrays or TensorFlow tensors)

[1922] Output: Flood risk rating for each region (ranging from 0 to 1)

[1923] High-risk areas are color-coded on a map, and the map is displayed using Google Maps API and Leaflet.

[1924] Step 4: Risk notification

[1925] server

[1926] The server notifies local government disaster prevention officials and residents of flood risk information via email, SMS, and a dedicated app.

[1927] Send SMS notifications using AWS SNS (Simple Notification Service).

[1928] Input: Flood risk assessment results, notification messages based on risk level

[1929] Output: Notifications (email, SMS) sent to local government disaster prevention officials and residents

[1930] Terminal

[1931] The device receives the information and checks the risk level. A push notification is sent to the smartphone app, displaying detailed information.

[1932] The device will present the user with an evacuation plan and evacuation route, using the Google Maps API to display safe evacuation routes.

[1933] Input: Risk notification message, evacuation plan information

[1934] Output: Notifications and evacuation routes displayed to the user

[1935] Step 5: Emotion recognition and notification adjustment

[1936] server

[1937] The server uses an emotion engine to analyze the emotional state of the user receiving the notification information in real time. It analyzes the text data using NLTK (Natural Language Toolkit) and the BERT model.

[1938] Tailor the wording and content of notifications based on the user's emotional state: if a user is panicking, send them a notification that helps them stay calm.

[1939] Input: User emotion data (text, reaction data)

[1940] Output: Adjusted notification message

[1941] Example: A message encouraging people to remain calm and evacuate.

[1942] Step 6: Gather feedback and improve the system

[1943] User

[1944] After an evacuation or flood occurs, users provide feedback on their emotional state and actual evacuation behavior. For example, they can write in the in-app feedback form, "I was able to make appropriate decisions quickly when evacuating."

[1945] Input: Feedback form input (text data)

[1946] Output: Feedback data provided

[1947] server

[1948] The server analyzes the collected feedback data to help improve prediction models and evacuation plans. It uses data mining techniques and sentiment analysis algorithms to analyze the feedback.

[1949] Input: Feedback data (emotional information, behavioral records)

[1950] Output: Improved models and evacuation plans

[1951] Improved emotion engine algorithms will provide more effective risk notifications and evacuation plans for the next flood.

[1952] This allows the system to achieve highly accurate flood risk prediction and user response at each step, promoting safe evacuation behavior.

[1953] (Application example 2)

[1954] 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 robot 414 will be referred to as a "terminal."

[1955] Conventional flood risk prediction systems primarily use meteorological and satellite data to predict flood risks and notify local governments and residents. However, they do not take into account the emotional state of passengers, particularly in autonomous vehicles, which can lead to confusion and anxiety caused by emotions such as panic. Furthermore, they lack the ability to automatically avoid areas with a high risk of flooding, which can make it difficult to fully ensure passenger safety. A system that can effectively solve these issues is needed.

[1956] 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.

[1957] In this invention, the server includes means for acquiring river water level data, precipitation data, and topographical data from a meteorological database and a satellite database, means for converting the acquired data into a unified format, and means for inputting the converted data into a generative AI model to predict flood risk, thereby enabling highly accurate flood risk prediction and risk assessment.

[1958] The server also includes a means for performing a risk assessment based on the predicted flood risk and visually displaying the risk level on a map, a means for notifying local government disaster prevention personnel and residents of the risk assessment results, a means for proposing an evacuation plan based on the notified risk assessment results, a means for collecting feedback from each local government and residents and using it to improve the accuracy of the model, a means for recognizing the emotional state of passengers, a means for customizing the content of the notification based on the emotional state, and a means for the autonomous vehicle to automatically select a route that avoids the flood risk. This enables flexible notifications according to the emotional state of passengers and enables the autonomous vehicle to select a safe route.

[1959] A "weather database" is a database that includes various weather data (e.g., precipitation data, temperature data, wind speed data, etc.).

[1960] A "satellite database" is a database that includes topographical data and water level data obtained from artificial satellites.

[1961] "River water level data" is data that indicates information on the water level in a specific river basin.

[1962] "Precipitation data" is data that indicates information about the amount of precipitation in a specific area.

[1963] "Topography data" refers to data that indicates information such as the topography and altitude of a specific area.

[1964] A "unified format" is a format for converting data obtained from different data sources into a consistent format.

[1965] A "generative AI model" is a machine learning model that makes predictions and classifications based on data.

[1966] "Flood risk" is an indicator of the likelihood of flooding occurring in a particular area within a certain ...

Claims

1. means for acquiring river water level data, precipitation data, and topographical data from a meteorological database; means for converting the acquired data into a unified format; A means for inputting the converted data into a generative AI model to predict flood risk; A means for performing risk assessment based on the predicted flood risk and visually displaying the risk level on a map; A means of notifying the results of the risk assessment to local government disaster prevention personnel and residents; A means for proposing an evacuation plan based on the notified risk assessment result; A means of collecting feedback from local governments and residents to improve the accuracy of the model; A system including:

2. 10. The system of claim 1, further comprising means for obtaining data in real time through a meteorological database and a satellite database.

3. The system of claim 1 , further comprising means for using a deep learning model as a generative AI model for predicting and assessing flood risk.

Images