Risk avoidance support system, server device, information processing method, program, and monitored terminal

Abstract

This technology provides the ability to further improve the safety of those being monitored. [Solution] A danger avoidance support system, server device, information processing method, program, and a monitored terminal are provided, comprising a monitored terminal carried by the person being monitored and a server. The server acquires the current location information of the monitored terminal, acquires danger information associated with area and time, and transmits predetermined warning information to the monitored terminal when the current location meets notification conditions based on the risk level and reliability of the acquired danger information. The reliability level is based on information source attributes, etc., and the risk level can be corrected by environmental factors. The server updates through relearning, and the terminal outputs vibrations, etc., according to the danger level.

Description

Technical field

[0001] The present invention relates to a danger avoidance support system, a server device, an information processing method, a program, and a watched-over terminal for assisting a watched-over person in avoiding danger.

Background art

[0002] In recent years, various monitoring services have been proposed to ensure the safety of watched-over persons such as children and the elderly (see Patent Document 1). Some of these services use a positioning function such as GPS (Global Positioning System) to grasp the current position of the watched-over person and notify the protector or the like. In addition, terminals equipped with a geofence function for detecting entry into and exit from a specific area and notifying, and a function for transmitting an SOS signal in an emergency are also becoming popular.

[0003] Furthermore, in order to provide more advanced safety support, technologies for warning watched-over persons of dangerous places and situations in advance have also been studied. For example, a system is known that creates a danger area map based on past accident information and suspicious person information and prompts attention when a watched-over person approaches the area. There are also attempts to utilize information collected in real time (for example, danger reports from other users, traffic conditions, etc.).

[0004] However, the conventional technologies have problems such that the types and ranges of danger information collected are limited, and the evaluation of the reliability and urgency of the information is not sufficient. In addition, at present, an integrated mechanism for integrating information from various information sources, predicting danger accurately and in real time according to the situation of each watched-over person, and supporting specific avoidance actions has not yet been established. In particular, it has been difficult to effectively collect local and real-time danger information (for example, temporary construction work, damage to playground equipment, dangerous traffic conditions at a specific time) possessed by local residents and the like and utilize it for ensuring the safety of watched-over persons.

Prior art documents

Patent documents

[0005] [Patent Document 1] Strength Patent No. 10631156 [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] One of the objectives of this invention is to provide a technology that can further improve the safety of the person being monitored. [Means for solving the problem]

[0007] To solve the above problems, a hazard avoidance support system according to one aspect of the present invention is a hazard avoidance support system comprising a monitored terminal equipped with a location information transmission function carried by the person being monitored, and a server, wherein the server acquires the current location information of the monitored terminal, acquires hazard information associated with area and time, and transmits predetermined warning information to the monitored terminal when the current location satisfies notification conditions based on the risk level and reliability of the acquired hazard information. [Effects of the Invention]

[0008] According to one aspect of the present invention, it is possible to improve the safety of the person being monitored. [Brief explanation of the drawing]

[0009] [Figure 1] This is a schematic diagram showing the overall configuration of a hazard avoidance support system according to one embodiment of the present invention. [Figure 2] This block diagram shows an example of the server hardware configuration according to this embodiment. [Figure 3] This is a block diagram showing an example of the hardware configuration of the monitored terminal according to this embodiment. [Figure 4] This block diagram shows an example of the hardware configuration of the parental device according to this embodiment. [Figure 5] This figure shows an example of the functional block configuration of the server according to this embodiment. [Figure 6] This figure shows an example of the functional block configuration of the monitored terminal according to this embodiment. [Figure 7] This figure shows an example of the functional block configuration of the parental device according to this embodiment. [Figure 8] This is a data structure diagram showing an example of hazard information used in this embodiment. [Figure 9] This flowchart shows an example of the hazard avoidance support process according to this embodiment. [Figure 10] This figure illustrates an example of the confidence calculation process according to this embodiment. [Figure 11] This figure illustrates an example of the risk calculation process (correction due to environmental factors) according to this embodiment. [Figure 12A] This is a perspective view showing an example of the appearance of the monitored terminal according to this embodiment. [Figure 12B] This is a perspective view showing an example of the appearance of the monitored terminal according to this embodiment. [Figure 12C] This is a perspective view showing an example of the appearance of the monitored terminal according to this embodiment. [Figure 13] This is a screen transition diagram showing an example of warning information output on the display of the monitored terminal according to this embodiment. [Figure 14] This figure shows an example of the screen display (map screen) of the parent terminal according to this embodiment. [Figure 15] This figure shows an example of the screen display (chat screen) of the parent terminal according to this embodiment. [Figure 16] This figure shows an example of the screen display of a parent terminal according to this embodiment (danger information posting screen). [Figure 17] This is a sequence diagram illustrating the additional notification process to the parent terminal according to this embodiment. [Figure 18] This figure illustrates the anonymized storage process of location history according to this embodiment. [Figure 19] This is a conceptual diagram illustrating the personalization process of warning information according to this embodiment. [Figure 20]This is a diagram for explaining a specific example of notification condition determination processing according to the present embodiment.

Mode for Carrying Out the Invention

[0010] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In each figure, the same or corresponding components are denoted by the same reference numerals, and redundant descriptions will be omitted as appropriate. Also, each component described as a "part" in this specification indicates a functional configuration unit, and does not necessarily mean that it is configured as a physically separated entity. That is, each part can be realized by hardware, software, or a combination thereof.

[0011] [[ID=

[12] ](Overview of the entire system) FIG. 1 is a schematic diagram showing the overall configuration of a danger avoidance support system 1 according to an embodiment of the present invention. As shown in FIG. 1, the danger avoidance support system 1 includes a wardee terminal 10 carried by a wardee U (for example, a child or an elderly person), a server 20, a protector terminal 30 used by a protector P, and a community information input terminal 40 used by a member C of a regional community (for example, a regional resident, a school-related person, a PTA, a volunteer, etc.). These terminals and the server are connected to be communicable with each other via a network N (for example, the Internet, a mobile communication network, a Wi-Fi (registered trademark) network, etc.).

[0012] The wardee terminal 10 has a positioning function such as a GPS receiver, and transmits its own current position information to the server 20 periodically or in response to a request from the server 20. The wardee terminal 10 also has a function of notifying the wardee U of the warning information received from the server 20 by sound, light, vibration, screen display, etc. An external appearance example of the wardee terminal 10 is shown in FIGS. 12A, 12B, and 12C. The wardee terminal 10 shown in FIG. 12A has a small-sized housing 10a suitable for portability, and can include a circular display 10b for displaying information and a circular button 10c operable by the wardee U on its surface. The monitored terminal 10 shown in Figure 12B similarly has a small housing 10a, and its surface is equipped with a rectangular display 10b for displaying information and a circular button 10c located below it. The monitored terminal 10 shown in Figure 12C is an example configuration in which only a circular display 10b is provided on the surface of the housing 10a, and physical buttons 10c are omitted. In this case, the display 10b has a touch panel function, and the same functions as buttons can be achieved by operating on the screen. These displays 10b may show the time, battery level, communication status, and warning information from the server 20 (e.g., danger icons, simple messages, changes in background color, etc.). Buttons 10c may be used, for example, to send an SOS signal, send a pre-set message to a guardian, or turn the device on or off. The material and shape of the housing 10a, and the shape, size, and arrangement of the displays 10b and buttons 10c are not limited to the illustrated examples and can be appropriately changed according to the age group of the person being monitored and the usage scenario. For example, a user-friendly design and colors may be adopted for children, while for the elderly, larger text and icons may be displayed, and button operation may be made simpler.

[0013] Server 20 acquires the current location information from the monitored terminal 10 and collects and integrates risk information associated with area (geographic range) and time (time of occurrence, validity period, etc.) from various sources (e.g., official information providers, guardian terminal 30, community information input terminal 40, etc.). Server 20 then evaluates and calculates the risk level and reliability of the collected risk information and sends warning information to the monitored terminal 10 if it determines that the current location of the monitored terminal 10 meets predetermined notification conditions (e.g., approaching or entering a high-risk and high-reliability risk area). This warning information may include instructions for alternative safe routes for the monitored person to avoid danger, as described in Appendix 8. Furthermore, as described in Appendix 9, the content of the warning information may be adjusted based on at least one of the following: the monitored person's age, pre-set characteristics (e.g., personality traits, specific phobias, etc.), or the current monitoring situation (e.g., whether they are alone or accompanied by a guardian, etc.).

[0014] The guardian terminal 30 is, for example, a smartphone, tablet, or personal computer, and is used by the guardian P to check information provided by the server 20 (for example, the current location and movement history of the person being monitored U, warning information sent from the server 20 to the monitored terminal 10, information on surrounding dangerous areas, etc.) on a map, send and receive messages (voice or text) with the monitored terminal 10, and provide the server 20 with information on dangers that the guardian P has recognized. The guardian terminal 30 also has the function of receiving emergency notifications from the server 20 regarding the person being monitored U (for example, additional notifications as described in Appendix 6) and notifying the guardian P.

[0015] The community information input terminal 40 is, for example, a smartphone, tablet, or personal computer, and is used by a member C of a local community to input and provide information about dangers they have noticed (for example, information about suspicious persons, information about dangerous locations, information about accidents, etc.) to the server 20, along with location information and time information.

[0016] (Description of hardware configuration) Next, an example of the hardware configuration of the server 20, the monitored terminal 10, and the guardian terminal 30 according to this embodiment will be described.

[0017] Figure 2 is a block diagram showing an example of the hardware configuration of server 20. Server 20 can be configured as a general-purpose computer system equipped with, for example, a CPU (Central Processing Unit) 201, RAM (Random Access Memory) 202, ROM (Read Only Memory) 203, auxiliary storage device 204 (e.g., HDD or SSD), communication interface 205, input device 206 (e.g., keyboard, mouse), and output device 207 (e.g., display). These components are interconnected via a bus 208. The CPU 201 loads programs stored in ROM 203 and auxiliary storage device 204 into RAM 202 and executes them, thereby realizing the various functional units of server 20 described later.

[0018] Figure 3 is a block diagram showing an example of the hardware configuration of the monitored terminal 10. The monitored terminal 10 includes, for example, a CPU 101, RAM 102, ROM 103, non-volatile memory 104 (e.g., flash memory), communication module 105 (e.g., a mobile communication module such as LTE (Long Term Evolution) or 5G, a Wi-Fi module, a Bluetooth® module, etc.), GPS receiver 106, various sensors 107 (e.g., an accelerometer, a gyroscope, an illuminance sensor, a temperature sensor, etc.), an input unit 108 (e.g., a button 10c shown in Figure 12, a touch panel (which may be integrated with the display 10b), etc.), an output unit 109 (e.g., a display 10b shown in Figure 12, a speaker (not shown), an LED (not shown), a vibrator (not shown), etc.), and a battery 110. The monitored terminal 10 may also be equipped with a USB Type-C® terminal (not shown) as a charging terminal. Furthermore, it is desirable that the housing 10a has water-resistant, dustproof, and shockproof properties. The CPU 101 loads programs stored in the ROM 103 and non-volatile memory 104 into the RAM 102 and executes them, thereby realizing the various functions of the monitored terminal 10, which will be described later.

[0019] Figure 4 is a block diagram showing an example of the hardware configuration of the parent terminal 30. The parent terminal 30 is typically a smartphone and includes a CPU 301, RAM 302, ROM 303, non-volatile memory 304, communication module 305 (compatible with Wi-Fi and mobile communication networks), GPS receiver 306, various sensors 307, touch panel display 308, camera 309, microphone 310, speaker 311, vibrator 312, etc. The CPU 301 realizes the various functional parts of the parent terminal 30, which will be described later, by executing the parent application (corresponding to the program described in Appendix 13) stored in the non-volatile memory 304, etc. The community information input terminal 40 can also have a hardware configuration similar to that of the parent terminal 30.

[0020] (Explanation of functional block configuration) Next, the functional block configurations of the server 20, the monitored terminal 10, and the guardian terminal 30 according to this embodiment will be described.

[0021] Figure 5 shows an example of the functional block configuration of server 20. Server 20 includes, as functional blocks, a location information acquisition unit 210, a hazard information acquisition unit 220, a risk level calculation unit 230, a reliability level calculation unit 240, a notification condition determination unit 250, a warning information transmission unit 260, a learning unit 270, and a location history management unit 280. These functional units are realized by the CPU 201 of server 20 executing a predetermined program.

[0022] The location information acquisition unit 210 periodically or as needed acquires the current location information (latitude, longitude, positioning time, etc.) from the monitored terminal 10 via the network N. Positioning methods that can be used individually or in combination include GNSS (GPS, Michibiki QZSS, BeiDou, Galileo, SBAS, etc.), assisted GPS (A-GPS), cloud positioning based on location information of Wi-Fi access points and mobile phone base stations, and motion positioning (inertial positioning) using acceleration sensors, etc.

[0023] The hazard information acquisition unit 220 acquires hazard information from various sources, which is associated with area (geographic range) and time (time of occurrence, validity period, etc.). Examples of information sources include publicly available information from official information providers such as the police and local governments (hereinafter referred to as "official information sources") (e.g., definitive information on suspicious persons, traffic accident information (location, date and time, type, etc.), warnings and evacuation information regarding natural disasters (typhoons, heavy rain, earthquakes, etc.) provided through websites and APIs), parent terminals 30 (more localized and real-time information on dangers that parents P have seen or heard in their daily lives or obtained from other parents or schools via the danger information posting unit 340 described later, such as sightings of suspicious persons in specific locations, reports of damage to playground equipment, and information on frequent occurrences of dangerous traffic behavior, posted as text, audio, images, or videos), and community information input terminals 40 (extensive or ongoing information on safety that is systematically collected or recognized by school personnel, PTAs, local volunteer groups, etc., such as areas of concern on school routes, temporary traffic restrictions and congestion information associated with local events, sharing of past near-miss incidents, and information on areas where bullying or exclusion is suspected).

[0024] The "hazard information associated with area and time" acquired by the hazard information acquisition unit 220 may include, in addition to information with a relatively wide area and a predetermined validity period provided by the official information sources and parent terminals mentioned above, the following information may also be included. For example, sensor data (e.g., acceleration sensor, gyroscope, microphone, illuminance sensor, temperature sensor, air quality sensor, etc.) detected by various sensors installed in the monitored terminal 10 itself that suggest potential dangers in the immediate surrounding environment of the monitored person (e.g., acceleration patterns that detect falls or collisions, sound patterns that detect screams or loud noises, sudden changes in illumination, detection of harmful gases, etc.) can also be acquired by the server 20 as hazard information associated with the location of the monitored terminal at the moment of detection as the "area" (specified as a point or extremely small area) and the detection time as the "time". In this case, the monitored terminal 10 transmits the detected sensor data to the server 20, and the hazard information acquisition unit 220 of the server 20 receives it and acquires the hazard information. Furthermore, real-time warning information shared directly or indirectly between other nearby monitored terminals and community information input terminals (for example, "information on suspicious person sightings nearby" or "information on recent accidents" shared via P2P communication using Bluetooth®, Wi-Fi Direct®, etc., or ad-hoc networks) is also associated with the location or target area and time from which the information was transmitted. Server 20 can acquire this information as "danger information" by relaying it or by receiving and collecting it indirectly from other terminals. Such localized and highly real-time information is also evaluated for risk level and reliability by Server 20 and used in deciding whether to send warning information.

[0025] In addition, the hazard information acquisition unit 220 may acquire incident reports and log information (including the content of the detected hazard, location, time, related sensor data, and details of the issued warning) transmitted from the monitored terminal 10 after the terminal autonomously detects a hazard and issues a warning. Even if the monitored terminal 10 locally determines a hazard and issues a warning based on a hazard profile or judgment model previously distributed from the server 20, the result is reported to the server 20, which then acquires this report as new "hazard information associated with area and time." This acquired information can be used by the server 20 to re-evaluate the risk level and reliability, by the learning unit 270 to learn and update the hazard profile and judgment model, or to notify the guardian terminal 30.

[0026] Hazard information may include information on suspicious persons, traffic accidents, natural disasters, environmental hazards (e.g., road collapses, damage to playground equipment), and other information that may threaten the safety of those being monitored. The hazard information acquisition unit 220 collects this information and stores it in a hazard information database (for example, built on the auxiliary storage device 204) described later, in an integrated manner that distinguishes the characteristics of each information source and the nature of the information.

[0027] The risk calculation unit 230 calculates the degree of potential danger (risk level) to the person being monitored for each piece of hazard information acquired by the hazard information acquisition unit 220, taking into consideration its content, type, scope of impact, past occurrence frequency, and related environmental factors. The risk level can be expressed, for example, as a normalized continuous value within a range of 0.0 (no danger) to 1.0 (extremely high danger), or categorized as multiple graded danger levels such as "low," "medium," and "high." When calculating the risk level, features such as the type of hazard information (e.g., stalking incidents by suspicious individuals posing a high direct threat to life and physical safety are high risk, environmental hazard information such as a simple streetlamp malfunction is medium risk, and local event information at a warning level is low risk), the severity of the reported hazard (e.g., incidents where actual damage has occurred are high risk, while mere eyewitness accounts or concerns are medium risk), the size of the area indicated by the hazard information and its geographical characteristics (e.g., narrow alleys with poor visibility, sparsely populated areas, and areas with few places to take refuge are higher risk), the frequency and statistical data of similar incidents in the past in the area, the time of day (e.g., specific time periods such as nighttime, early morning, and school hours are higher risk), the day of the week (e.g., busy downtown areas on weekend nights are higher risk), and the weather (related to the correction for environmental factors described later) can be used as features.

[0028] Based on these features, the risk level for each piece of hazard information can be quantitatively calculated using a rule-based scoring model designed in advance by an expert (for example, assigning weights to each feature to assign a score, and determining the risk level based on the total score), or a predictive model trained using machine learning with a vast amount of past hazard information and subsequent actual accident / incident occurrences and the extent of damage as training data (for example, logistic regression models, support vector machines, decision tree ensembles (random forests, gradient boosting trees, etc.), neural networks, etc.). When using a machine learning model, the server 20 trains the model using these features and their corresponding past hazard event occurrences (or risk level labels assigned by experts) as training data, and updates the model periodically or as needed by the learning unit 270 described later. The risk level and reliability level calculated here are not only used directly by the server 20 when deciding whether to send a warning in real time, but also serve as important basic information for generating and updating the aforementioned "risk profile" and "decision model" that are distributed to the monitored terminal 10. In other words, the results of the risk assessment and reliability assessment by the server 20 are indirectly incorporated into these profiles and models.

[0029] Furthermore, as described in Appendix 3, the risk assessment unit 230 can correct the calculated risk level using at least one environmental factor in the area, such as rainfall, illuminance (brightness), or temperature. For example, in the case of nighttime (illuminance below a predetermined lux) or bad weather (e.g., when rainfall is 5 mm / h or more, or when the temperature is below freezing), the risk level is adjusted to be higher even if the hazard information is the same. This correction may be performed by multiplying the basic risk level by a correction coefficient (e.g., 1.5 times for nighttime and rainy weather, 1.1 times for simply cloudy and somewhat dark weather, etc., which can be set in stages) or by adding it. Alternatively, these environmental factors themselves may be added as part of the input features of the machine learning model for calculating the risk level described above, and the model may be configured to automatically learn the contribution of the environmental factors and reflect them in the risk level.

[0030] This environmental information can be obtained in real time or periodically from weather information services and other sources. This allows for a more accurate assessment of potential environmental risks, such as poor visibility and deteriorating road conditions during rainy weather, poor visibility at night, and increased activity of suspicious individuals, thereby improving the accuracy of risk prediction, especially under these conditions.

[0031] Official sources may include, for example, definitive information on suspicious individuals, traffic accidents, and natural disaster warnings disclosed by the police or local authorities. Parent terminals 30 may post more localized and real-time information about dangers that parents P have directly witnessed or heard in their daily lives (e.g., sightings of suspicious individuals in specific locations, damage to playground equipment, frequent occurrences of dangerous traffic behavior, etc.) as text, audio, images, or videos. Community information input terminals 40 may provide broader or ongoing safety information that has been systematically collected or recognized by school officials, PTAs, local volunteer groups, etc. (e.g., areas of concern on school routes, temporary traffic restrictions and congestion information related to local events, past near-miss incidents, etc.). The danger information acquisition unit 220 can handle these diverse types of information in an integrated manner while distinguishing between them.

[0032] The reliability calculation unit 240 calculates the degree of certainty (reliability) of each piece of hazard information acquired by the hazard information acquisition unit 220. The reliability can be expressed as a continuous value between, for example, 0.0 (very low reliability) and 1.0 (very high reliability). As described in Appendix 2, the reliability calculation unit 240 can calculate the reliability based on at least two of the following: the source attribute of the hazard information, the degree of reporting overlap, the specificity of the reported content, and the passage of time. The reliability calculation unit 240 calculates the degree of certainty (reliability) of each piece of risk information acquired by the risk information acquisition unit 220. The method for calculating or evaluating this reliability can take various forms depending on the nature of the information and the source of the information. For example, in addition to a detailed method that calculates a dynamic score by comprehensively evaluating multiple elements such as source attribute, degree of report overlap, specificity of report content, and passage of time, it is also possible to use a simpler method or apply a prior evaluation of a specific source. Specifically, for risk information acquired from specific sources whose reliability has been certified in advance by the server administrator, etc. as being extremely high, such as official organizations like the police, fire departments, and local governments, or information providers whose reliability is guaranteed based on contracts, the reliability can be uniformly set to a predetermined high fixed value (for example, 1.0 or a similar value), or the reliability can be categorized into several predefined stages (e.g., "high reliability," "normal reliability," "requires verification," etc.) and assigned to one of the stages according to the type and characteristics of the source of information. These processes are also included in the scope of reliability evaluation and calculation in this invention. In this way, server 20 performs some form of reliability assessment on each piece of risk information, depending on its content and source.

[0033] Specifically, as an information source attribute, a base confidence level is set according to the type of information source. For example, information from official sources has an initial confidence level of 1.0, information from authenticated community information input terminals 40 such as schools and PTAs has a confidence level of 0.8, and information from general parent terminals 30 has a confidence level of 0.6. Furthermore, the system can learn from each information provider's past posting history (e.g., whether their past information provision was highly accurate, how often they were pointed out as misinformation, etc.), dynamically adjust the confidence score for each information provider, and incorporate this into the evaluation of the information source attribute.

[0034] Next, regarding the degree of overlap in reporting, if the same or similar risk information (for example, determined by AI-based similarity of text content, or proximity of location information and time of occurrence) is reported multiple times (e.g., three or more times) from different sources within a short period (e.g., within the last hour), the reliability of that information is considered high, and the score is increased (e.g., the maximum reliability of each report is taken, or an additive adjustment is made according to the number of reports and the diversity of sources).

[0035] Furthermore, the specificity of the report is evaluated based on whether the hazardous situation is described in detail, whether the type of hazard is clearly identified, whether multimedia information that enhances objectivity, such as photos, videos, and audio, is included, and whether the location and time of occurrence are accurately indicated. The higher the level of specificity, the higher the reliability.

[0036] Furthermore, as time progresses, the likelihood that the information reflects the current situation decreases as the time elapsed since it was reported decreases, and the confidence score is, for example, exponentially reduced.

[0037] The confidence calculation unit 240 may use these elements as features and calculate confidence based on a predefined set of rules or formulas, or it may use a machine learning model (for example, logistic regression, decision trees, support vector machines, or ensemble models thereof) that has been trained on previously collected hazard information and the accuracy of that information (e.g., whether an accident actually occurred, whether multiple corroborations were obtained, whether it was a false report, etc.) as training data to calculate an overall confidence score. In this case, the AI ​​can learn the content of the collected hazard information (e.g., the credibility and sentiment polarity of keywords extracted from text information using natural language processing techniques), the reporter's past behavior patterns (e.g., the correlation between the frequency of reports in a particular area and past reliability evaluations), and the spatiotemporal relationship with other information, and use implicit evaluation criteria that are difficult for humans to explicitly define as rules to calculate confidence.

[0038] The notification condition determination unit 250 compares the current location information of the monitored terminal 10 acquired by the location information acquisition unit 210 with the danger information acquired by the danger information acquisition unit 220, for which the risk level and reliability level have been calculated by the risk level calculation unit 230 and the reliability level calculation unit 240, respectively. The notification condition determination unit 250 then determines whether the current location of the monitored terminal 10 satisfies predetermined notification conditions. For example, as described in Appendix 10, the notification conditions can be set to be "satisfied when the risk level is equal to or greater than a first predetermined threshold, and the reliability level is equal to or greater than a second predetermined threshold." However, this is just one example of notification conditions, and the present invention is not limited to such specific logical conditions. These first and second predetermined thresholds may be fixed values, or they may be dynamically adjusted based on the age of the person being monitored, a characteristic profile set in advance by the guardian (e.g., changing the threshold according to the level of sensitivity to danger), the current situation of the person being monitored (e.g., setting the threshold lower when the person is alone, and higher when the person is with an adult), the time of day (e.g., lowering the threshold at night), the weather (e.g., lowering the threshold during bad weather), etc.

[0039] Figure 20 is a diagram illustrating a specific example of the notification condition determination process. It shows an example where the distance D between the current location 2001 of the person being monitored and the dangerous area 2002 (with risk level R and confidence level S) is evaluated, and the notification condition is determined to be met if D is within a predetermined distance (for example, within a radius of X meters from the boundary of the dangerous area, or when entering the dangerous area), and R ≥ threshold R1 and S ≥ threshold S1. Alternatively, the notification condition determination unit 250 can combine multiple determination logics, such as considering both risk level (R) and confidence level (S), and determining that the notification condition is met if the overall danger score calculated by (R × S) exceeds a predetermined threshold Th1. It can also apply different determination logics and thresholds depending on the type of danger information (e.g., in the case of highly urgent danger types that directly affect life, notifications may be issued even with lower risk levels or confidence levels). Furthermore, a configuration is also conceivable in which the notification condition determination unit 250 makes a more sophisticated situational judgment based on a combination of conditions, such as when the person being monitored enters an area where danger reports have frequently occurred in the past (a high-risk potential area learned by the learning unit 270) at a specific time of day (e.g., twilight), or when movement that deviates significantly from the normal behavior pattern is detected (e.g., rapid movement in a direction not usually taken for a predetermined period of time or longer), and issues a warning.

[0040] The reliability level is also expressed as a continuous value between, for example, 0.0 and 1.0. Regarding source attribute, for example, the initial reliability level can be set to 1.0 for official information, 0.8 for verified communities, and 0.6 for general guardians, and can be dynamically adjusted based on each information provider's past posting history (e.g., number of times providing highly accurate information, number of times misinformation has been pointed out, etc.). Regarding the degree of overlapping reports, if similar danger information (e.g., determined by keyword matching or AI-based content similarity judgment) is reported from multiple different sources in the same area (e.g., within a 50m radius) within a short period of time (e.g., within 1 hour), the reliability level of that group of reports will be increased synergistically (e.g., by taking the maximum reliability level of each report, or by adding adjustments according to the number of reports). Regarding the specificity of the report content, reliability will be increased if multimedia information such as photos, videos, and audio is included, or if the situation of danger, location, time, and characteristics of the suspicious person are described in detail and objectively.

[0041] Regarding the passage of time, the reliability is processed to decrease exponentially with the time elapsed since the report. These elements can be used as features, and a machine learning model (e.g., a classification model that has learned patterns of information that were evaluated as "accurate" in past feedback) can be used to calculate an overall reliability score. In this case, the AI ​​can learn the content of the collected risk information (e.g., keywords and sentiment polarity extracted from text information using natural language processing techniques), the reporter's past behavior patterns (e.g., frequency of reports in a specific area and past reliability evaluations), and the spatiotemporal relationship with other information, and build implicit evaluation criteria to use in calculating reliability.

[0042] The warning information transmission unit 260 transmits predetermined warning information to the monitored terminal 10 when the notification condition determination unit 250 determines that the notification conditions are met. The predetermined warning information includes information that at least suggests the presence or type of danger to the monitored person, and proactively prompts them to pay attention or take some kind of avoidance action. This includes icons or simple text messages indicating specific types of danger (e.g., suspicious person, traffic accident warning, construction, falling rock warning, etc.), warning sounds according to the danger level, flashing lights such as LEDs, vibration patterns, etc. The warning information may include the type of danger, the danger level, recommended avoidance actions (e.g., leave the area, move in a specific direction, temporarily evacuate to a nearby safe place (e.g., convenience store, police box, etc.), contact a guardian, etc.), and instructions for alternative safe routes as described in Appendix 8. The warning information transmission unit 260 transmits this warning information proactively before the monitored person actually encounters danger or enters a dangerous area. Furthermore, the warning information transmission unit 260 can simultaneously transmit the same warning information, a more detailed explanation of the situation (such as the content of the danger information and an overview of the basis for evaluating reliability and risk), and a request for action to the parent terminal 30.

[0043] Furthermore, the "predetermined warning information" transmitted from the server 20 to the monitored terminal 10 is not limited to direct warning displays or warning sound instructions, but may include information that allows the monitored terminal 10 to autonomously judge danger and issue a warning, such as a "danger profile" that defines the type and pattern of danger and the corresponding initial risk level and countermeasures, or a "judgment model" for identifying dangerous situations from sensor data, its parameters, rule sets, etc. In this case, if the server 20 transmits (distributes) these danger profiles and judgment models to the monitored terminal 10, and the monitored terminal 10 autonomously generates and outputs a warning using them, it can be interpreted that the server 20 has effectively transmitted the information that forms the basis of the warning. Based on the learning results from the learning unit 270 described later, the server 20 periodically or as needed updates these danger profiles and judgment models and transmits them to the monitored terminal 10.

[0044] The aforementioned alternative safe route instructions are generated, for example, by the risk level calculation unit 230 of the server 20 or a dedicated route search engine (not shown) searching in real time for the safest detour route, which avoids dangerous areas and takes into account other known and predicted danger levels at that time, based on the current location of the monitored terminal 10, the destination if set, and the calculated and corrected risk level map (or dangerous area information). In doing so, a unique safety scoring system is assigned to each route segment, which not only avoids danger but also considers factors such as the presence or absence of sidewalks, the number and brightness of streetlights, the amount of pedestrian traffic (including variations depending on the time of day and day of the week), the safety of intersections (presence or absence of traffic lights, visibility, etc.), and the proximity of evacuation facilities (e.g., children's emergency homes, convenience stores, train stations, etc.), and an overall optimal route or multiple recommended route candidates are selected. The maximum allowable detour distance and travel time are also considered to ensure that the detour is not too long for the monitored person. The generated alternative safety route instructions are presented to the monitored device 10 in an intuitive interface that children can easily understand, such as simple direction-of-travel icons on the display 10b (e.g., an arrow icon indicating to turn right at the next corner, an icon indicating to go straight), an LED (if available) that indicates the direction of travel with a flashing light pattern, or short voice guidance such as "Turn right at the next corner" or "Keep going straight." For the guardian device 30, the dangerous area and the recommended alternative safety route are clearly displayed in different colors on its map screen 1400, allowing the guardian to confirm their appropriateness and, if necessary, provide additional verbal instructions to the monitored person U.

[0045] (Personalization of warning information) As described in Appendix 9, the warning information transmission unit 260 can adjust the content of the warning information to be transmitted according to the characteristics of the person being monitored U. Figure 19 is a conceptual diagram illustrating the personalization process of warning information. The server 20 stores the person being monitored profile 1901 (for example, age, gender, language setting (e.g., Japanese, English), allergy information (e.g., issuing particularly strong warnings in areas where bees are a concern), phobias of specific things (e.g., dislike of dark places, dislike of loud noises), personality traits (e.g., keywords selected by the guardian from multiple options or entered in free text, such as 'cautious,' 'curious and prone to approaching dangerous places,' 'easily panicked,' etc.), and understanding level according to the stage of intellectual development (selected by the guardian), etc.) which has been registered in advance from the various settings units 350 of the guardian terminal 30. Furthermore, the system estimates the current monitoring status 1902 (e.g., walking, cycling, vehicle, etc.) from information from various sensors 107 of the monitored terminal 10 (e.g., means of transportation estimated from the accelerometer (walking, cycling, vehicle, etc.)) and the proximity status between the monitored terminal 10 and the guardian terminal 30 (e.g., distance estimation using Bluetooth®, etc.).

[0046] The warning information transmission unit 260, based on these monitored person profiles 1901 and the current monitored situation 1902, adjusts the wording of the warning message 1903 (e.g., for toddlers, "There's a dog, let's go that way!"; for elementary school children, "There's a suspicious person. Avoid the alley ahead and take the main road," using language and message length appropriate to age and comprehension level. When an adult is present, direct messages to the child are suppressed, and detailed information is sent to the guardian), the intensity of the warning 1904 (e.g., for children set as "easily frightened," the warning volume is kept low, or a reassuring melody is used. For children who "don't like loud noises," vibration is used as the primary method. The volume, vibration strength, and display color vividness are adjusted according to the danger level). Personalized warning information 1906 is generated by adjusting factors such as the degree to which the information is presented (e.g., gradually changing the level of information), the type of information presented 1905 (for example, for children who are set as "easily frightened" or "prone to panic," more reassuring words (e.g., "It's okay, this is the safe way") are added, and specific evacuation locations (e.g., "The nearby XX supermarket") are actively presented. On the other hand, for children who are "curious," consider adding a short reason why that action is necessary (e.g., "It's dangerous here because bicycles suddenly dart out, so let's go through the park"). If there is a danger related to allergy information (e.g., a beehive warning area), a warning is given along with an icon that suggests it).

[0047] In the future, it is conceivable that the AI ​​will automatically adjust and optimize the most effective and acceptable way of expressing guidance, the timing of presentation, and the amount of information provided, based on the monitored person's past behavioral logs, physiological responses to specific warnings and guidance (if wearable sensors can detect changes in heart rate, etc.), and behavioral changes (for example, whether they acted as instructed or showed discomfort to a specific warning sound, learned indirectly through sensor information on the device and feedback from parents).

[0048] As described in Appendix 6, after the warning information transmission unit 260 transmits warning information to the monitored terminal 10, if the location of the monitored terminal 10 does not change by a predetermined distance or more within a predetermined time (i.e., if there is a possibility that the monitored person U has not taken appropriate evasive action in response to the warning), it may send an additional notification to the guardian terminal 30 (for example, an alert indicating a high level of urgency, or a message prompting contact with the monitored person U).

[0049] As described in Appendix 4, the learning unit 270 updates the risk level or reliability of hazard information by relearning at predetermined time intervals. Specifically, in addition to regular relearning at predetermined time intervals (e.g., once a week, once a month), it adaptively performs relearning when manually triggered at any time by the system administrator, when the amount of newly collected hazard information exceeds a certain amount (e.g., 1000 items), when a certain percentage of feedback from parents (e.g., evaluation information and comments provided from the parent app such as "This warning was helpful," "This was a false alarm," or "This area was actually safe") is accumulated, or when the system's hazard prediction accuracy is determined to have fallen below a predetermined value in monitoring indicators (e.g., recall and precision in a validation set based on past data).

[0050] The training data used includes feedback from parents, information on actual accidents and incidents (for example, the results of spatiotemporally comparing publicly available news articles and statistical data with hazard information within the system to determine the accuracy of predictions), and logs of whether the person being monitored actually took evasive action in response to a specific warning (for example, anonymized and statistical data on the percentage of people who followed the recommended evasive route after a warning). The learning unit 270 uses this training data to update (retrain) the entire set of parameters of the machine learning models used by the risk level calculation unit 230 and the confidence level calculation unit 240, or to perform additional learning on existing trained models using only new training data (incremental learning or online learning), or to revise the model structure itself and replace it with a higher-performance model, or to optimize the weighting and thresholds of each element in the rule-based scoring logic based on the accumulated data. This continuously improves the accuracy of risk level and confidence level calculations and allows the system to adapt to changing environments and trends in new hazards.

[0051] The location history management unit 280 manages the location history of the person being monitored U, as described in Appendix 7. In particular, from the standpoint of protecting privacy, the collected location history can be anonymized after a predetermined period of time by coarsening the coordinates (for example, converting detailed latitude and longitude information into a wider-area mesh code or information at the city / town level) and stored in a form that does not allow for the identification of individuals. The anonymized location history data can be used, for example, for trend analysis of new danger areas or as training data for AI models.

[0052] Figure 6 shows an example of the functional block configuration of the monitored terminal 10. The monitored terminal 10 may include, as functional blocks, a location information transmission unit 120, a warning information reception unit 130, a warning output unit 140, a time display unit 150, and a battery status display unit 160. These functional units are realized by the CPU 101 of the monitored terminal 10 executing a predetermined program.

[0053] The location information transmission unit 120 periodically transmits the current location information of the monitored terminal 10, which has been determined using a GPS receiver 106 or the like, to the server 20 via the communication module 105, or in response to a request from the server 20.

[0054] The warning information receiving unit 130 receives warning information transmitted from the server 20 via the communication module 105.

[0055] The warning output unit 140 notifies the person being monitored U of the warning information received by the warning information receiving unit 130 using the output unit 109, particularly the display 10b, speaker, vibrator, etc.

[0056] As described in Appendix 5 and Appendix 14, the warning output unit 140 can output warning information as different vibration patterns or color displays depending on the degree of danger included in the received warning information. For example, as shown in Figure 13, the display 10b normally displays the normal display 1301 (e.g., the time and battery level), but when it receives warning information from the server 20, it displays a warning display 1302 according to the degree of danger. In this warning display 1302, the background color of the display 10b can be changed in stages according to the degree of danger (e.g., blue normally, yellow for attention level 1 (low risk), orange for attention level 2 (medium risk), red for warning level (high risk), etc.), or an animated warning icon 1302a that intuitively indicates the type of danger (e.g., suspicious person, traffic accident, natural disaster, etc.) and the degree of danger can be displayed.

[0057] In addition, the device can simultaneously output different vibration patterns 1303 from the vibrator of the output unit 109 according to the level of danger (for example, a short, single vibration for low danger, repeated short vibrations for moderate danger, and a long, continuous strong vibration for high danger, designed to attract the attention of the person being monitored without causing excessive anxiety), or output warning sounds (such as sounds whose tone, tempo, and volume change according to the level of danger, or specific sound effects that suggest the type of danger) from the speaker. If the display on the display 10b or the voice warnings are undesirable due to school rules, etc., the voice warnings can be turned off and the device can switch to notifying warning information only through vibration and / or display on the display 10b (for example, a change in color display, an icon display, or flashing of an LED if connected), by remote settings from the various settings unit 350 of the parent terminal 30 or by operating a specific button 10c on the monitored terminal 10 (for example, by long-pressing button 10c or setting silent mode by a specific combination of operations). This makes it possible to effectively alert those being monitored even in environments where it is inappropriate to emit sounds or displays, such as schools, libraries, and public transportation.

[0058] The time display unit 150 acquires the current time and displays it on the display 10b. The time information may be acquired from a GPS signal, or from a server 20 or an NTP server via the communication module 105.

[0059] The battery status display unit 160 detects the remaining charge of the battery 110 and displays the remaining charge information on the display 10b. For example, it can be displayed as an icon or a percentage.

[0060] Figure 7 shows an example of the functional block configuration of the parent terminal 30. The parent terminal 30 includes an information display unit 310, a message sending / receiving unit 320, a notification receiving unit 330, a danger information posting unit 340, and various setting units 350 as functional blocks of the parent application (which is realized by the execution of the program described in Appendix 13 by the CPU 301). (Comment: We clarified that the parent app handles programmatic claims and that functional blocks are the result of that.)

[0061] The information display unit 310 displays various information received from the server 20 (for example, the current location of the person being monitored U, their movement history, warning information sent from the server 20 to the monitored terminal 10, and information on surrounding dangerous areas) on the touch panel display 308. For example, as shown in Figure 14, the map screen 1400 can display the current location icon 1410 of the person being monitored U, their movement history 1420 showing past movement routes, pre-registered notification spots 1430 such as home and school, and dangerous areas 1440 recognized by the server 20 (for example, displayed in different colors according to the degree of risk). In addition, if a warning is sent from the server 20 to the monitored terminal 10, the content of the warning (for example, the type of danger, recommended avoidance route, etc.) can also be displayed on the map screen 1400 or on a separate screen. The guardian P can check the details of areas of interest by scrolling and zooming on the map screen 1400.

[0062] The message sending / receiving unit 320 sends and receives voice messages and text messages with the monitored terminal 10. Figure 15 shows an example of the talk screen 1500 of the guardian terminal 30. The guardian P can record a voice message using the microphone 310 and send it to the monitored terminal 10 via the server 20 by operating the send button 1510a (sent voice 1510), or enter a text message in the text input area 1520a and send it by operating the send button 1520b (sent text 1520). Voice messages received from the monitored terminal 10 (received voice 1530) can be played back on the speaker 311 by operating the play button 1530a, and the content converted into text by the voice recognition function of the server 20 (transcribed text 1540) may also be displayed. Read information 1550 (for example, a "read" mark or playback time) indicating whether the message has been played back on the monitored terminal 10 may also be displayed.

[0063] The notification receiving unit 330 receives various notifications concerning the person being monitored U (for example, arrival / departure notifications to notification spots, notifications of deviation from the activity range, warnings of approaching dangerous areas, low battery notifications, and additional notifications as described in Appendix 6) transmitted from the server 20 in real time, and notifies the guardian P by a pop-up display on the touch panel display 308, sound (notification sound), or vibration of the vibrator 312. The guardian P can transition to the relevant detailed information screen (for example, the map screen 1400 or the chat screen 1500) by performing operations such as tapping the notification.

[0064] The Danger Information Posting Unit 340 provides a user interface for parents P to send danger information they have discovered or recognized to the server 20. Figure 16 shows an example of the Danger Information Posting Screen 1600. Parent P can specify the type of danger 1610 (e.g., suspicious person, traffic accident, dangerous location, environmental problem, etc.) using a pull-down menu or icon selection, specify the location where the danger occurred by tapping on the map (location input 1620), or use the current location obtained by the GPS receiver 306. In addition, they can input the date and time of occurrence 1630, a detailed description of the situation 1640 as text, attach photos or videos taken with the camera 309 (not shown), or attach voice memos recorded with the microphone 310, and send it to the server 20 by operating the post button 1650.

[0065] The various settings unit 350 provides a user interface for the guardian P to make various settings related to this system (for example, registering and editing the profile of the person being monitored (age, gender, personality traits, allergy information, etc.), registering and editing notification spots (home, school, cram school, etc.) on the map, deciding whether or not to receive various notifications (danger warnings, arrival and departure notifications, etc.), setting notification sounds and vibration patterns, setting the operating mode of the monitored terminal 10 (e.g., silent mode such as turning off warning sounds in school, vibration only, or display off), and selecting the frequency of location information updates (e.g., battery priority mode, frequency priority mode), etc.).

[0066] (Explanation of data structure) Figure 8 is a data structure diagram showing an example of hazard information that can be stored in the hazard information database of server 20 in this embodiment. As shown in Figure 8, a hazard information record 800 may include fields such as: hazard information ID 801, reporting date and time 802, source type 803 (e.g., official, parent, community A, monitored terminal sensor, P2P information, etc.), reporter ID 804 (may be anonymized), hazard area information 805 (e.g., latitude and longitude, radius, polygon shape, if detected by a monitored terminal sensor, point information indicating the detection location or a very narrow area, if P2P information indicates a specific location or route section, etc.), predicted hazard occurrence time period 806 (the detection time itself in the case of real-time detection), hazard type 807 (e.g., suspicious person, traffic accident, environmental hazard, fall detection, sudden acceleration / deceleration detection, loud noise detection, air quality abnormality, etc.), hazard details 808 (text, links to photos / audio files, sensor data values ​​or patterns, etc.), calculated risk level 809, calculated confidence level 810, and related environmental factors 811 (e.g., weather, illuminance, temperature, etc.). This information is collected and organized by the risk information acquisition unit 220, and the risk level calculation unit 230 and the confidence level calculation unit 240 calculate and update the risk level 809 and confidence level 810.

[0067] (Explanation of the processing flow) Next, an example of the processing flow in the hazard avoidance support system 1 according to this embodiment will be described. Figure 9 is a flowchart showing an example of the hazard avoidance support processing performed by the server 20.

[0068] First, the location information acquisition unit 210 of the server 20 acquires the current location information from the monitored terminal 10 (step S901).

[0069] Next, the hazard information acquisition unit 220 acquires hazard information associated with area and time from official information sources, parent terminals 30, community information input terminals 40, etc., and updates the hazard information database (step S902).

[0070] Next, the confidence calculation unit 240 calculates (or recalculates) the confidence level of the risk information acquired or updated in step S902 (step S903). The confidence level is calculated considering factors such as those shown in Figure 10. Figure 10 is a diagram illustrating an example of the confidence level calculation process, showing how information source attributes 1001, report overlap 1002, specificity of report content 1003, and the passage of time 1004 become inputs to the machine learning model 1005 and rule-based logic, and how a confidence score 1006 is output.

[0071] Next, the risk calculation unit 230 calculates (or recalculates) the risk level for the hazard information acquired or updated in step S902 (step S904). In calculating the risk level, as shown in Figure 11, a correction 1103 is made to the basic risk assessment 1101 (based on hazard type, past occurrence frequency, etc.) by environmental factors 1102 (e.g., rainfall, illuminance, temperature, etc.) to calculate the final risk score 1104.

[0072] Subsequently, the notification condition determination unit 250 compares the current location of the monitored terminal 10 obtained in step S901 with each piece of risk information having a reliability level calculated in step S903 and a risk level calculated in step S904, and determines whether the current location of the monitored terminal 10 satisfies the predetermined notification conditions (step S905).

[0073] If the notification conditions are not met (NO in step S905), the process returns to step S901 and waits for the next location information acquisition.

[0074] On the other hand, if it is determined that the notification conditions are met (YES in step S905), the warning information transmission unit 260 transmits predetermined warning information to the monitored terminal 10 (step S906). At this time, a notification is also sent to the guardian terminal 30 if necessary. As shown in Figure 13, the monitored terminal 10 performs processing based on the received warning information, such as transitioning from the normal display 1301 to the warning display 1302 (color display or icon display 1302a according to the degree of danger) and outputting a vibration pattern 1303.

[0075] Furthermore, the server 20 can monitor the actions of the monitored terminal 10 after the warning information has been sent. Figure 17 is a sequence diagram illustrating the process of sending an additional notification to the guardian terminal. After the server 20 sends warning information to the monitored terminal 10 (sequence S1701), it monitors the location information that is periodically sent from the monitored terminal 10 (sequence S1702). If, after a predetermined time T1 has elapsed since the warning was sent (sequence S1703), the change in the location of the monitored terminal 10 is less than a predetermined distance D1 (YES in sequence S1704), the server 20 determines that the monitored person U may not have been able to avoid danger and sends an additional emergency notification to the guardian terminal 30 (sequence S1705).

[0076] Furthermore, the learning unit 270 uses the collected hazard information, the calculated risk level and confidence level, the transmitted warning information, and any feedback thereon to retrain the machine learning model at predetermined time intervals and update the risk level and confidence level calculation logic (not shown, corresponding to Appendix 4).

[0077] The location history management unit 280, as shown in Figure 18, converts the original detailed coordinate data 1801 of the location history of the monitored terminal 10 into coarser coordinate data 1802 (for example, by meshing or aggregating to representative points) after a predetermined period of time has elapsed, and then performs an anonymization process 1803 to prevent the identification of individuals, before saving it to the anonymized location history database 1804 (corresponding to Appendix 7).

[0078] (Examples of variations and applications, etc.) The present invention is not limited to the embodiments described above, and various modifications are possible without departing from its spirit. For example, in addition to the above, other sources of hazard information could also be used, such as information from sensors mounted on vehicles (e.g., sudden brake detection, road surface condition sensors) or the results of analyzing real-time video collected by drones, etc.

[0079] The algorithms for calculating risk and reliability, the specific thresholds for notification conditions, and the content and presentation methods of warning information (for example, more interactive guidance or specific navigation to a safe evacuation location) can also be modified as appropriate to the extent that the objectives of the present invention can be achieved. The monitored terminal 10 may also have a function to play voice messages based on warning information received from the server 20 (for example, voice recorded by the guardian or voice synthesized by the server from text information). In this case, it can be linked to the operation of the button 10c on the monitored terminal 10 (for example, play with a click, record a response message with a long press).

[0080] The monitored device may be implemented as a smartphone application or provided as a dedicated wearable device. Furthermore, the server functionality may be implemented as a single server device or as a cloud computing system distributed across multiple server devices.

[0081] Although this embodiment primarily describes the use of the invention for monitoring children, it can be applied to a variety of people and uses, such as monitoring the elderly, providing safety support for people walking alone at night, providing real-time safety information to foreign tourists, or managing the safety of workers.

[0082] Furthermore, the collected and analyzed hazard information and risk maps could be used in collaboration with local governments and police to develop regional crime prevention plans, traffic safety measures, and urban planning. This would not only provide support for individual risk avoidance but also contribute to creating safer communities.

[0083] This invention contributes to improving the functionality of computer systems (particularly servers, monitored terminals, and guardian terminals). Specifically, it enables advanced information processing on the server by collecting and integrating fragmented and unreliable risk information from diverse sources in real time, objectively evaluating the reliability and urgency of that information using machine learning, and further predicting spatiotemporal risk occurrence. This makes it possible to detect potential risks early and accurately evaluate them, which was previously difficult, and improves the risk detection capability of the information processing system. In addition, the system architecture for efficiently performing this computationally intensive analysis and prediction (e.g., stream processing, distributed processing, optimization of machine learning models, etc.) contributes to the efficient use of computer processing power and improved responsiveness.

[0084] The monitored terminal provides a user interface that outputs complex danger information and risk levels received from the server in a format that is intuitively understandable and encourages action for the monitored person, such as a child (e.g., different vibration patterns depending on the level of danger, and color and icon displays on the display 10b). This helps the monitored person recognize the situation, supports quick danger avoidance actions, and improves usability and accessibility. In particular, in emergencies, concise and clear information transmission helps prevent errors and reduces the mental burden on the user. The content displayed on the display 10b (time, battery level, warning icons, etc.) is optimized considering the visibility and ease of understanding for children.

[0085] On parental devices, a dedicated application provides a user interface that displays the current location and movement history of the person being monitored, information on surrounding dangers, and warnings from the server in an integrated and visually easy-to-understand manner on a map. This allows parents to quickly grasp complex information and make accurate judgments about the situation. In addition, a simple function for posting danger information and a message sending / receiving function encourage parental participation in the system, contributing to more efficient information gathering and smoother communication. These UI improvements reduce the operational burden on parents and enhance the continuity and effectiveness of monitoring activities.

[0086] Furthermore, although this embodiment describes a configuration in which the server 20 performs major intelligent processing (such as evaluating risk information, predicting risks, and generating guidance), the present invention is not limited to this. For example, if the monitored terminal 10 or guardian terminal 30 has a certain level of processing power, it is also possible to configure it as a distributed processing system that applies the concepts of edge computing or fog computing, in which some risk evaluation processing and customization processing of warning displays according to the situation are performed on the terminal side based on simplified risk information and judgment models (for example, lightweight machine learning models that detect specific risk patterns) distributed from the server 20. Even in this case, the server 20 still plays an important role in broad information collection and integration, learning and management of advanced AI models, and overall system management. Alternatively, it is conceivable to provide an auxiliary information transmission route in which some risk information (especially those of high urgency) is directly shared between nearby monitored terminals or guardian terminals in a P2P (Peer-to-Peer) manner without going through the server 20. These modifications can also be included within the scope of the technical idea of ​​the present invention.

[0087] Furthermore, while this embodiment has primarily described a configuration in which the server 20 performs major intelligent processing (such as evaluating risk information, predicting risks, and generating guidance), the present invention is not limited to this. For example, if the monitored terminal 10 or guardian terminal 30 has a certain level of processing power (for example, if it is equipped with a more powerful CPU or a dedicated AI chip), it is also possible to configure it as a distributed processing system that applies the concepts of edge computing or fog computing, in which the terminal performs some risk evaluation processing (e.g., immediate risk assessment in combination with local sensor information), customized warning display processing according to the situation, or short-term movement prediction and simple route selection based on that, based on simplified risk information (e.g., a set of unevaluated raw information of the surrounding area) or a general-purpose judgment model or its parameters (e.g., a lightweight machine learning model or rule set for detecting specific risk patterns) distributed from the server 20.

[0088] In this case, the server 20 still plays an important role in broad information gathering and integration, learning, managing, and distributing advanced AI models, overall system control, and complex analysis that cannot be processed on the terminal side. Alternatively, it is conceivable to provide an auxiliary information transmission path in which some danger information (especially information that is highly urgent and meaningful only among nearby terminals, such as "a suspicious person is talking to a child right around the corner"—ultra-localized, ultra-real-time information) is directly shared and warned about between nearby monitored terminals and guardian terminals via a P2P (Peer-to-Peer) ad-hoc network using Bluetooth®, Wi-Fi Direct®, or other short-range wireless communication technologies, without going through the server 20. These variations can also be included within the scope of the technical concept of the present invention and can contribute to improving system responsiveness, reducing communication load, and providing limited functionality in offline situations. Even if the information is shared and warned about via these P2P-like ad-hoc networks, for example, if one of the terminals that initially generates or receives the information transmits it to the server 20 (after anonymizing or summarizing it as necessary), the server 20 can "acquire" the "risk information associated with area and time" and use it to make warning decisions for a wider range of other monitored individuals, as well as to record and analyze the information. Thus, not only information directly collected by the server, but also information indirectly acquired via monitored terminals or other terminals can be included in the methods of acquiring risk information in this invention. In such a distributed processing system, even when the monitored terminal 10 autonomously performs some risk assessment processing and warning generation using judgment models and risk profiles distributed from the server 20, the server 20 still plays the role of (a) a centralized intelligence hub that collects and analyzes a wide range of risk information, evaluates the degree of risk and reliability of that information, and generates and updates the judgment models and risk profiles; (b) a role in transmitting (distributing) the generated and updated judgment models and risk profiles (which substantially constitute the basis or logical framework of "predetermined warning information") to the monitored terminal; and (c) a role in acquiring post-event reports from the monitored terminal regarding risk detection events and warning activations (which are also a form of "risk information associated with area and time"), thereby contributing to the learning of the entire system and the improvement of risk avoidance capabilities. Therefore, such a configuration can also be included in the basic framework of the present invention in that the server is involved in acquiring risk information, evaluation based on the degree of risk and reliability (including inclusion in models, etc.), and transmitting warning information (or the information underlying it).

[0089] The components of each embodiment described herein can be combined as appropriate, to the extent that they do not conflict with each other.

[0090] [General tasks] One of the objectives of this invention is to provide a technology that can further improve the safety of the person being monitored.

[0091] Issues corresponding to [Appendix 1] One of the objectives of this invention is to reduce the risk of the person being monitored encountering dangerous situations and to improve safety. [Note 1] A risk avoidance support system comprising a monitored terminal equipped with a location information transmission function carried by the person being monitored, and a server, wherein the server acquires the current location information of the monitored terminal, acquires risk information associated with area and time, and transmits predetermined warning information to the monitored terminal when the current location meets notification conditions based on the risk level and reliability of the acquired risk information. (Effects of Appendix 1) By evaluating various types of risk information based on their risk level and reliability, and sending warning information when the monitored person's current location meets the notification conditions, the risk of the monitored person encountering dangerous situations can be reduced. This makes it possible to improve the safety of the monitored person.

[0092] Issues corresponding to [Appendix 2] One of the objectives of this invention is to evaluate the reliability of risk information more objectively and from multiple perspectives. [Note 2] A risk avoidance support system as described in Note 1, wherein the reliability is calculated based on at least two of the following: the attributes of the source of risk information, the degree of overlap in reports, the specificity of the reported content, and the passage of time. (Effects of Appendix 2) By calculating the reliability of risk information based on multiple objective indicators such as the attributes of the information source, the degree of overlap in reports, the specificity of the content, and the passage of time, warning decisions can be made based on more reliable information, contributing to the reduction of false alarms and the improvement of the accuracy of warnings.

[0093] Issues corresponding to [Appendix 3] One of the objectives of this invention is to evaluate the degree of risk of danger more realistically by taking environmental factors into consideration. [Note 3] A hazard avoidance support system as described in Note 1, wherein the calculation of the risk level is corrected using at least one of the rainfall, illuminance, or temperature in the area in question. (Effects of Appendix 3) By considering that environmental factors such as rainfall, darkness, and abnormal temperatures affect the occurrence and severity of hazards, and by adjusting the risk level, it becomes possible to make more situation-appropriate and realistic hazard assessments, thereby improving the accuracy of warnings against potential risks that tend to be overlooked.

[0094] Issues corresponding to [Appendix 4] One of the objectives of this invention is to continuously improve the accuracy of hazard information assessment. [Note 4] A hazard avoidance support system as described in Note 1, wherein the server updates the risk level or reliability level of the hazard information by relearning at predetermined time intervals. (Effects of Appendix 4) By retraining and updating the risk level and reliability calculation logic (e.g., machine learning model) based on new hazard information and feedback, the system can adapt to changing environments and new hazard trends, and continuously maintain and improve the accuracy of its hazard assessment.

[0095] Issues corresponding to [Appendix 5] One of the objectives of this invention is to communicate the degree of danger to the person being monitored in a more intuitive way. [Note 5] A hazard avoidance support system as described in Note 1, wherein the monitored terminal outputs the warning information as different vibration patterns or color displays according to the degree of danger. (Effects according to Appendix 5) By outputting warning information using vibration intensity and rhythm that changes according to the degree of danger, or by displaying a color (e.g., green, yellow, red), the degree of danger can be conveyed intuitively and effectively, especially to children and those being monitored who have difficulty instantly understanding textual information, thereby prompting attention and behavioral change.

[0096] Issues corresponding to [Appendix 6] One of the objectives of this invention is to enable prompt intervention by guardians when the person being monitored does not respond to a warning. [Note 6] A risk avoidance support system as described in Note 1, wherein the server sends an additional notification to the guardian terminal if the position of the monitored terminal does not change by a predetermined distance or more within a predetermined time after the transmission of the warning information. (Effects of Appendix 6) If a person under guardianship has been warned but has not taken or is unable to take action to avoid a dangerous situation, promptly notifying the guardian of the situation will encourage more direct intervention by the guardian, such as making a phone call or rushing to the scene, thereby increasing the certainty of ensuring the safety of the person under guardianship.

[0097] Issues corresponding to [Appendix 7] One of the objectives of this invention is to safely manage and utilize location information while respecting the privacy of the person being monitored. [Note 7] The risk avoidance support system described in Note 1, wherein the server stores the location history of the person being monitored after a predetermined period of time has elapsed, by coarsening the coordinates and storing it anonymously. (Effects of Appendix 7) Instead of retaining the detailed location history of the person being monitored for a long period of time, the granularity of the coordinates is reduced after a certain period of time, and the data is stored anonymously so that individuals cannot be identified. This reduces the risk of privacy infringement while enabling its use for statistical analysis and as training data for AI.

[0098] Issues corresponding to [Appendix 8] One of the objectives of the present invention is to include specific avoidance measures in the warning information provided to the person being monitored. [Note 8] A hazard avoidance support system as described in Note 1, wherein the warning information includes instructions for an alternative safe route for the person being monitored to avoid danger. (Effects of Appendix 8) By not only warning of danger but also providing specific alternative safe routes, it is possible to help those being monitored take quicker and more appropriate actions to avoid danger, thereby increasing the effectiveness of safety measures.

[0099] Issues corresponding to [Appendix 9] One of the objectives of this invention is to provide optimal warning information tailored to the situation of the person being monitored. [Note 9] A hazard avoidance support system as described in Note 1, wherein the warning information is adjusted based on at least one of the age of the person being monitored, pre-set characteristics, or the current situation of the person being monitored. (Effects of Appendix 9) By adjusting the content of the warning information (wording, amount of information, intensity of warning, etc.) according to the age and characteristics (personality, allergies, etc.) of the person being monitored and their current situation (whether they are alone or accompanied by an adult, etc.), the information can be conveyed in a way that is easier for the person being monitored to understand and accept, thereby enhancing the effectiveness of the warning.

[0100] Issues corresponding to [Appendix 10] One of the objectives of the present invention is to more specifically define the notification conditions for issuing a warning. [Note 10] A hazard avoidance support system as described in Note 1, wherein the notification condition is met when the risk level is equal to or greater than a first predetermined threshold and the reliability level is equal to or greater than a second predetermined threshold. (Effects of Appendix 10) By determining notification conditions based on specific thresholds that consider both the risk level and reliability of the hazard information, the need for warnings can be judged more accurately, the occurrence of unnecessary warnings can be suppressed, and the reliability of warnings for hazards that must not be overlooked can be increased.

[0101] Issues corresponding to [Appendix 11] One of the objectives of the present invention is to provide a server device that reduces the risk of the person being monitored encountering dangerous situations and improves safety. [Note 11] A server device that receives location information from a monitored terminal and transmits warning information based on danger information, comprising: A) a location acquisition unit that receives location information of the monitored terminal; B) a danger information acquisition unit that acquires danger information associated with area and time; C) a notification determination unit that determines whether the location information satisfies notification conditions based on the risk level and reliability of the danger information; and D) a notification unit that transmits warning information to the monitored terminal when it is determined that the notification conditions are met. (Effects of Appendix 11) By evaluating various risk information based on its risk level and reliability, and sending warning information when the monitored person's current location meets the notification conditions, the risk of the monitored person encountering dangerous situations can be reduced. This makes it possible to provide a server device that can improve the safety of the monitored person.

[0102] Issues corresponding to [Appendix 12] One of the objectives of the present invention is to provide an information processing method that reduces the risk of the person being monitored encountering dangerous situations and improves safety. [Note 12] An information processing method to be performed by a processor, comprising: (1) acquiring location information from a monitored terminal; (2) acquiring danger information associated with area and time; (3) determining whether the acquired location information satisfies notification conditions based on the risk level and reliability of the danger information; and (4) transmitting warning information to the monitored terminal if it is determined that the notification conditions are met. (Effects of Appendix 12) By evaluating various risk information based on its risk level and reliability, and by performing the step of sending warning information when the monitored person's current location meets the notification conditions, the risk of the monitored person encountering dangerous situations can be reduced. This provides an information processing method that can improve the safety of the monitored person.

[0103] Issues corresponding to [Appendix 13] One of the objectives of the present invention is to provide a program that helps guardians effectively understand the situation of the person they are looking after and take appropriate action as needed. [Note 13] A program for causing the processor of a guardian terminal to perform the following steps: receiving monitoring information from a server that is communicably connected to the guardian terminal, which includes warning information about the person being monitored, the current location information of the person being monitored, and at least one of the information related to the danger information, transmitted by the server based on the current location and danger information of the monitored terminal; displaying the received monitoring information on the display of the guardian terminal; and transmitting a message to the monitored terminal or danger information to the server in response to user operation on the guardian terminal. (Effects of Appendix 13) By having the parent terminal's processor receive and display important information about the person being monitored (warnings, current location, danger information, etc.) from the server, and by enabling the parent to send messages and danger information, a program can be provided that efficiently and effectively supports the parent's monitoring activities and contributes to ensuring the safety of the person being monitored.

[0104] Issues corresponding to [Appendix 14] One of the objectives of the present invention is to provide a monitoring terminal that can more intuitively communicate the degree of danger to the person being monitored. [Note 14] A monitored terminal equipped with a location information transmission function and a warning output function, comprising: (i) a location transmission unit that transmits the current location to a server; and (ii) a warning output unit that outputs warning information received from the server, wherein the warning output unit performs different vibration patterns or color displays according to the degree of danger. (Effects of Appendix 14) By outputting warning information received from the server using vibration intensity, rhythm, or display color that changes according to the degree of danger, it is possible to provide a monitoring terminal that can intuitively and effectively convey the degree of danger, especially to children or those being monitored who have difficulty instantly understanding textual information, thereby prompting attention and behavioral change. [Explanation of symbols]

[0105] 1… Hazard avoidance support system 10…Monitored device 10a... enclosure 10b…Display 10c... button 20... Server 30…Parental device 301...CPU (Parental Control Device) 302...RAM (Parental Device) 303...ROM (Parental Control Unit) 304...Non-volatile memory (parental control device) 305...Communication module (parental device) 306…GPS receiver (parental device) 307... Various sensors (parental device) 308…Touchscreen display (parental device) 309...Camera (Parental device) 310... Microphone (Parental device) 311...Speaker (Parental Device) 312...Vibrator (Parental Control Device) 320...Message sending / receiving unit (parental device) 330... Notification receiving unit (parent / guardian terminal) 340... Danger Information Submission Department (Parent / Guardian Terminal) 350... Various settings section (parental device) 40…Community information input terminal 101, 201…CPU 102, 202... RAM 103, 203…ROM 104... Non-volatile memory 105...Communication module 106...GPS receiver 107... Various sensors 108...Input section 109...Output section 110... Battery 120...Location information transmission unit 130...Warning Information Receiving Unit 140...Warning output section 150...Time display section 160...Battery status display 204…Auxiliary storage device 205...Communication Interface 206...Input device 207…Output device 208... Bus 210...location information acquisition unit 220... Hazard Information Acquisition Department 230...Risk Calculation Unit 240... Confidence calculation unit 250...Notification condition determination section 260…Warning Information Transmission Unit 270…Learning Department 280...Location History Management Department 800… Hazard Information Record 801... Danger Information ID 802…Report date and time 803... Source type 804…Reporter ID 805... Dangerous Area Information 806... Predicted time period for danger 807… Hazard Classification 808...Details of the danger 809…Risk level 810... Reliability 811…Related environmental factors 1001…Attributes of the information source 1002… Degree of overlap in reports 1003…Specificity of the report content 1004... Passage of time 1005…Machine learning models 1006...Confidence score 1101...Basic Risk Assessment 1102…Environmental factors 1103... Correction 1104…Risk score 1301…Normal display 1302...Warning display 1302a…Warning icon 1303...Vibration Pattern 1400...Map screen (parent's device) 1410... Icon of the current location of person being monitored U 1420... Movement history 1430... Notification Spot 1440... Dangerous area 1500... Chat screen (parent's device) 1510... Transmitted voice 1510a... Send button (voice) 1520...Send text 1520a... Text input area 1520b... Send button (text) 1530... Received audio 1530a... Play button 1540...Transcript text 1550... Read information 1600... Danger Information Submission Screen (Parent / Guardian Device) 1610...Types of dangers 1620... Location input 1630... Date and time 1640...Detailed explanation of the situation 1650...Submit button 1801...Detailed coordinate data 1802...Roughened coordinate data 1803...Anonymization process 1804...Anonymized Location History Database 1901... Profile of the person being monitored 1902...Current status of those under surveillance 1903…Warning message expression 1904…Warning intensity 1905...Types of information to present 1906…Personalized Warning Information 2001...Current location of the person being monitored 2002…Dangerous Area N... Network U... Person being watched over P…Parent C... Members of the local community S901~S906...Step S1701~S1705...Sequence

Claims

1. A risk avoidance support system comprising a monitoring terminal equipped with a location information transmission function carried by the person being monitored, and a server, The aforementioned server, The current location information of the monitored terminal is acquired, Obtain danger information linked to area and time, Based on the risk level and reliability of the acquired danger information, if the current location meets the notification conditions, a predetermined warning information is sent to the monitored terminal. Hazard avoidance support system.

2. A hazard avoidance support system according to claim 1, The aforementioned reliability is calculated by a risk avoidance support system based on at least two of the following: the attributes of the source of risk information, the degree of overlap in reports, the specificity of the reported content, and the passage of time.

3. A hazard avoidance support system according to claim 1, The aforementioned risk avoidance support system calculates the risk level using at least one of the following: rainfall, illuminance, or temperature in the area in question.

4. A hazard avoidance support system according to claim 1, The server is a risk avoidance support system that updates the risk level or reliability of the risk information by relearning at predetermined time intervals.

5. A hazard avoidance support system according to claim 1, The monitored terminal is a danger avoidance support system that outputs the warning information as different vibration patterns or color displays according to the degree of danger.

6. A hazard avoidance support system according to claim 1, The server is a risk avoidance support system that sends an additional notification to the guardian's terminal if the location of the monitored terminal does not change by a predetermined distance or more within a predetermined time after the transmission of the warning information.

7. A hazard avoidance support system according to claim 1, The aforementioned server is a risk avoidance support system that stores the location history of the person being monitored after a predetermined period of time by coarsening the coordinates and anonymizing the data.

8. A hazard avoidance support system according to claim 1, The warning information includes instructions for an alternative safe route for the person being monitored to avoid danger, and is part of a hazard avoidance support system.

9. A hazard avoidance support system according to claim 1, The warning information is a risk avoidance support system in which the content is adjusted based on at least one of the age of the person being monitored, pre-set characteristics, or the current situation of the person being monitored.

10. A hazard avoidance support system according to claim 1, The aforementioned notification conditions are met when the risk level is equal to or greater than a first predetermined threshold and the reliability level is equal to or greater than a second predetermined threshold in a risk avoidance support system.

11. A server device that receives location information from a monitored terminal and transmits warning information based on danger information, A) Location acquisition unit that receives location information of the monitored terminal, B) Hazard information acquisition unit that acquires hazard information associated with area and time. C) A notification determination unit that determines whether the location information satisfies the notification conditions based on the risk level and reliability of the aforementioned danger information. D) A notification unit that sends warning information to the monitored terminal when it is determined that the notification conditions are met. A server device equipped with the following features.

12. A method of information processing performed by a processor, (1) A step of obtaining location information from the monitored device, (2) Steps to obtain hazard information associated with area and time, (3) A step in which the acquired location information is determined to meet the notification conditions based on the risk level and reliability of the said danger information, (4) A step in which warning information is sent to the monitored terminal when it is determined that the notification conditions are met. Information processing methods including

13. The processor in the parent device, The steps include receiving monitoring information from a server that is communicatively connected to the guardian terminal, which includes warning information about the person being monitored transmitted by the server based on the current location and danger information of the monitored terminal, the current location information of the person being monitored, and at least one of the information related to the danger information, The steps include: displaying the received monitoring information on the display of the guardian terminal; The steps include sending a message to the monitored device or danger information to the server in response to user operation on the guardian device, A program to execute.

14. A monitoring terminal equipped with a location information transmission function and a warning output function, (i) A location transmission unit that transmits the current location to the server, (ii) Warning output unit that outputs warning information received from the server, Equipped with, The aforementioned warning output unit is a monitored terminal that displays different vibration patterns or colors according to the degree of danger.

Images