Method for attendance authentication of meeting and system therefor

Abstract

The present invention relates to a method and system for reliably determining whether a user has actually arrived and authenticating attendance at a meeting by utilizing location information and sensor data. It determines in real-time whether a user has entered a geofence set based on the meeting location, and based on the result, generates an attendance authentication notification and sends it to other users or pre-designated devices. The geofence can be dynamically adjusted to reflect environmental factors such as terrain, building structure, weather, and traffic conditions, as well as the user's means of transportation, reliability, and health status, and an artificial intelligence model learns and corrects these factors to improve authentication accuracy.

Description

Technology Field

[0001] The present invention relates to a location information-based offline meeting attendance authentication system and method, and more specifically, to a geofence-based meeting authentication technology that uses a user's location information to verify whether the user has actually arrived at the meeting place, notifies other users of the result in real time, and simultaneously analyzes the attendance history to calculate the reliability of the user and the meeting.

[0002] The present invention relates to an AI-based location-adaptive attendance authentication system that dynamically adjusts the geofence radius by reflecting variables such as user characteristics, movement environment, and health status, predicts the probability of the user's attendance through an AI model, and notifies a guardian or a third-party terminal when an abnormal condition is detected. Background Technology

[0003] This invention originated from a technical background aimed at verifying attendance at offline meetings more objectively and reliably, amidst the current trend of widespread use of location-based technologies. Existing meeting platforms and attendance management systems often struggle to accurately determine whether a user has actually arrived at the meeting location, and most rely on methods that judge attendance based solely on simple GPS signals or time input. Consequently, the inability to clearly distinguish between actual attendees and non-attendees has led to a problem that undermines the credibility of the meetings.

[0004] Furthermore, as these systems failed to reflect individual characteristics such as users' means of transportation, health status, or travel speed, there were instances where users with mobility limitations were unfairly disadvantaged. In addition, the lack of features to evaluate reliability based on a user's history of repeated tardiness or absences, or to analyze the actual operational activity of meetings, made it difficult to guarantee the transparency and sustainability of the overall service.

[0005] Therefore, the present invention proposes a new meeting attendance authentication technology that goes beyond simple location recognition, performs accurate arrival authentication through geofence settings and AI-based judgments that reflect the user's characteristics and movement context, strengthens social trust by sharing the results in real time, and can provide safety notifications to guardians or third parties when necessary. Prior art literature

[0006] (Registered Patent 1) 10-2533133 (2023.05.11) The problem to be solved

[0007] The present invention aims to resolve the error and reliability issues inherent in existing simple location-based attendance systems and to enable more accurate and fair authentication of attendance at offline meetings. Existing technology determines attendance based solely on GPS coordinates, failing to properly correct signal errors occurring in urban areas or inside buildings. Furthermore, because it does not consider user movement characteristics or environmental constraints, errors frequently occurred where actual attendees were recognized as absent.

[0008] Accordingly, the present invention aims to minimize false positives by dynamically setting geofences by reflecting various factors such as terrain, weather, traffic congestion, and physical characteristics, and by calculating the probability of attendance in real time through an artificial intelligence model that learns the user's movement patterns and reliability. In addition, if multiple users perform authentication at the same time, reliability is corrected through mutual authentication, and safety is enhanced by automatically notifying a guardian or a third party if the user's health condition or movement abnormality is detected.

[0009] Ultimately, the present invention aims to implement a new form of meeting attendance authentication technology that satisfies accuracy, fairness, reliability, and safety by integrating personalized geofence adjustment, AI-based judgment, mutual trust correction, and third-party notification functions. means of solving the problem

[0010] A method for authenticating meeting attendance to solve the above problems may include: (a) setting a geofence of a pre-configured area based on the meeting location; (b) determining in real time whether a user terminal has entered the geofence; (c) generating a notification authenticating the user's attendance at the meeting based on the determination; and (d) transmitting the generated meeting attendance authentication to other user terminals of the meeting.

[0011] In addition, in the above method, the geofence of step (a) may be characterized by dynamically changing the pre-set area by reflecting at least one of the topographical characteristics of the gathering location, weather conditions, the number of floors of the building, and geographical obstacles.

[0012] In addition, in the above method, the geofence of step (a) reflects at least one of user characteristics, reliability, means of transportation, and traffic congestion,

[0013] The above-mentioned pre-set area can be characterized by changing it on an individual user basis.

[0014] In addition, in the above method, the geofence of step (a) may be characterized by changing the pre-set area by the first artificial intelligence model.

[0015] In addition, in the above method, the pre-set area of ​​step (a) may be characterized as being an irregularly formed closed area.

[0016] In addition, in the above method, step (b) may be characterized by determining in real time whether the user terminal has entered the geofence area by using at least one of a location sensor, an accelerometer, and a gyroscope.

[0017] In addition, in the above method, step (b) may be characterized by calculating the probability of a user attending a meeting in real time by a second artificial intelligence model that reflects at least one of the user's characteristics, reliability, means of transportation, and traffic congestion.

[0018] In addition, in the above method, step (c) may be characterized by determining the plurality of users as a companion group when the plurality of user terminals belonging to the same group generate a notification within a preset time interval.

[0019] In addition, in the above method, the companion group may be characterized by having its reliability corrected and calculated.

[0020] In addition, the above method may further include the step of transmitting the generated meeting attendance authentication to one or more pre-configured terminals.

[0021] In addition, in the above method, the one or more pre-configured terminals may be characterized as being terminals registered as at least one of a family member, an acquaintance, and a guardian.

[0022] In addition, in the above method, step (d) may be characterized by transmitting a meeting attendance authentication containing the details of the abnormality to another user terminal of the meeting if there is an abnormality in the user's characteristics.

[0023] In addition, in the above method, the user characteristics may be characterized by including at least one of the following: the user's physical condition, health information, whether a mobility aid is used, disability information, age, movement speed, and history of abnormal movement.

[0024] In addition, the above method may be characterized in that the reliability is calculated based on at least one of the following: user-specific arrival time history, meeting attendance history, meeting tardiness history, meeting cancellation history, and meeting non-attendance history.

[0025] Additionally, in a meeting attendance authentication system comprising a central processing unit and a memory, the central processing unit executes instructions for executing a meeting attendance authentication method stored in the memory, wherein the method may include: (a) setting a geofence of a pre-set area based on the meeting location; (b) determining in real time whether a user terminal has entered the geofence area; (c) generating a notification to authenticate the user's attendance at the meeting according to the determination; and (d) transmitting the generated meeting attendance authentication to another user terminal of the meeting. Effects of the invention

[0026] According to the present invention, the shape and radius of a geofence can be dynamically adjusted by comprehensively analyzing the user's location information and surrounding environment data, thereby effectively improving problems such as signal reflection, misrecognition in densely populated urban areas, and location errors inside buildings that occurred in existing location-based systems. In addition, by reflecting individual characteristics such as the user's means of transportation, physical condition, age, and traffic conditions, the problem of unfairly excluding the attendance of users who are slow or have mobility limitations can be prevented.

[0027] Furthermore, by utilizing an artificial intelligence model to learn users' past movement history and attendance patterns and calculating the probability of attending a meeting in real time, intelligent arrival determination beyond simple coordinate-based verification becomes possible. When multiple users perform authentication within the same geofence within a certain period, the authenticity of the meeting and the reliability of joint participation can be enhanced by correcting the trustworthiness of each user through mutual authentication.

[0028] Furthermore, if abnormalities in a user's health information or movement are detected, third parties such as guardians or family members are immediately notified, ensuring personal safety and enabling emergency response. These features go beyond simple location tracking to realize a next-generation offline gathering authentication ecosystem that considers both social trust and personal safety.

[0029] Consequently, the present invention provides an intelligent location-based meeting authentication technology that equally satisfies accuracy, fairness, reliability, and safety, thereby fundamentally resolving the technical limitations of existing meeting platforms and simultaneously improving user experience and service reliability. Brief explanation of the drawing

[0030] FIG. 1 schematically illustrates the overall flow of a method for authenticating attendance at a meeting according to one embodiment of the present invention. FIG. 2 illustrates the overall sequence of a meeting attendance authentication method according to one embodiment of the present invention. FIG. 3 schematically illustrates the configuration of a geofence setting step according to one embodiment of the present invention. FIG. 4 illustrates the overall sequence of the real-time entry determination step according to one embodiment of the present invention. FIG. 5 schematically illustrates the overall flow of the companion group determination step according to one embodiment of the present invention. FIG. 6 schematically illustrates the overall flow of the meeting attendance authentication transmission step according to one embodiment of the present invention. FIG. 7 schematically illustrates the flow of a third-party transmission process to a meeting attendance authentication transmission step according to an embodiment of the present invention. Specific details for implementing the invention

[0031] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In assigning reference numerals to the components of each drawing, the same components are given the same reference numeral whenever possible, even if they are shown in different drawings. Furthermore, in describing these embodiments, if it is determined that a detailed description of related known components or functions could obscure the essence of the present invention, such detailed description may be omitted.

[0032] Terms such as "comprising," "having," and "consisting of" as used in this specification do not exclude the addition of other elements unless otherwise specified. Furthermore, even when a component is expressed in the singular, it may be interpreted as including the plural unless otherwise noted.

[0033] In describing the components of the present invention, terms such as "first," "second," etc. are used to distinguish the components from other components and do not limit the order, number, or importance of the components.

[0034] When describing the positional relationship of components, if two or more components are described using terms such as "connection," "combination," or "connection," they may be directly connected, or they may be connected with an intermediate component interposed. In this case, the inclusion of an intermediate component should be interpreted as falling within the scope of the present invention.

[0035] When describing the flow of time, such as a method of operation or production, even if terms such as "after," "following," or "next" are used, a non-continuous flow may be included unless "immediately" or "directly" is specified.

[0036] Furthermore, the functional blocks shown in the drawings are merely examples of possible implementations, and various modifications are possible without departing from the spirit and scope of the invention. The functional blocks may be implemented in hardware, software, or a combination thereof, and the invention is not limited to specific configurations.

[0037] The purpose, technical configuration, and resulting effects of the present invention will be more clearly understood through the attached drawings, and embodiments of the present invention will be described in detail below. The embodiments disclosed in this specification should not be interpreted as limiting the scope of the present invention, and it should be understood that various applications and modifications are possible.

[0038] In the left area of ​​Fig. 1, a closed boundary encircling the meeting place is displayed in an irregular shape on the map; this boundary is designed to reflect topographical factors such as the arrangement of surrounding buildings, alley shapes, and radio wave reflection zones. A reference marker indicating the meeting place is placed inside the boundary, and multiple user location markers are distributed both inside and outside the boundary. The markers inside the boundary signify users whose arrival has been authenticated, and changes in the marker's intensity and saturation may also suggest additional states based on biometric signals collected from the terminal or sensor conditions. For example, a marker indicating a stable stay within the boundary signifies a normal arrival requiring no separate action, while a marker within the boundary but with a high biometric signal intensity indicates that a burden has been detected during the movement and stay process, allowing for fine-tuning of the subsequent notification policy. The markers outside the boundary signify users who have not yet arrived, and the color tone and outline shape of the markers can be used to intuitively distinguish the level of arrival estimated by the server. One user is assigned a high probability of arrival as they are determined to be on foot, another user has their prediction conservatively adjusted due to a history of multiple past absences, and yet another user is assigned a medium probability reflecting route variability and potential delays due to being within the traffic congestion zone. These judgments are calculated by considering contextual information such as current location and direction of movement, stability of travel speed, participation patterns in past gatherings, weather and road conditions on the day, and wireless signal quality.

[0039] The summary panel on the right side of the drawing, representing this, serves as a legend area that briefly explains the meaning of each user indicator, displaying in text the characteristic judgment criteria referenced by the system, such as the normal arrival of users within the boundary, specific signals like the heart rate response of internal users, the high probability of arrival of external users, the tendency of external users to be absent, and the passage of external users through congested areas.

[0040] In particular, the boundaries of geofences on the drawings are not limited to circles or regular polygons but are represented as irregular closed lines that follow the contours of terrain features. This is intended to compensate for the influence that real-world physical boundaries, such as building clusters, walls, underground entrances, and elevated structures, have on positioning. In complex spaces where indoor and outdoor environments are mixed, auxiliary indicators such as near-field wireless identifiers, wireless signal strength, and periodic scan results are added to finely adjust the width and shape of the boundaries. Corrections based on individual mobility characteristics are also applied; for users utilizing mobility aids or those with slow movement speeds, the allowable boundary width is set slightly wider, while for users with consistently stable movement paths, stricter dwell criteria are applied to determine boundary entry. These corrections operate as technical measures to ensure the fair recognition of actual arrivals, rather than intending arbitrary discrimination that is disadvantageous or advantageous to individual users, and are performed within a limited scope based on user consent and privacy policies.

[0041] In the diagram, users within the boundary are interpreted as having their arrival authentication already verified by the server, and the server notifies other terminals in the same group of this fact without delay. When notifying, group determination may be performed in parallel to form a mutual verification relationship if multiple authentications accumulate within a short period of time at the same location, and the results are displayed intensively on the operation screen or host screen. Conversely, for users outside the boundary, the availability of a stay signal and location reliability criteria are checked together to avoid making hasty judgments based solely on prediction results; if movement is stagnant for a long time or the biometric signal shows an abnormal range, safety guidance is provided first to the user terminal, and a brief notification may be transmitted to a pre-registered third-party terminal.

[0042] In the embodiment, the server performs a determination of dwelling within the boundary using map tiles and a gridded spatial representation, and the terminal filters out instantaneous positional jumps by estimating the movement state using sensor fusion signals such as acceleration, gyroscope, and geomagnetism. In sections with low location reliability, the average movement amount is calculated by correcting the set of recent coordinates as a central trend, and the structure may have a mechanism where entry candidate events are not immediately confirmed but are promoted to authentication after observing a short dwell time. The prediction model learns differences in time zones, days of the week, and meeting types from past meeting data, returning predictions with higher reliability for meetings of the same type; furthermore, for users with an accumulated tendency to be absent, the model is designed to ensure that potential changes are not missed by using a weighting method that reflects recent trends more significantly than simple cumulative deductions.

[0043] This configuration is designed so that the terminal's simple response, the server's reliability correction, boundary dynamic adjustment, and predictive updates do not conflict with each other, and contributes to fairly acknowledging actual arrivals without excessive misrecognition, even under real-world conditions such as geographical congestion, indoor environments, or the use of mobility aids.

[0044] In the geofence setting step (S100) of Fig. 2, the server retrieves the coordinates and location type of a meeting place registered at a certain time and calculates a pre-set area. At this time, the shape of the area is expressed as a closed boundary that considers geographical features on the map, building layout, entry and exit routes, and areas where radio waves can reflect. In environments where indoor and outdoor spaces are mixed, auxiliary indicators such as wireless proximity signals or network identifiers are used together. The server prepares the boundary in advance for the period prior to the start of the meeting schedule to prevent the background location service on the terminal side from repeating excessive scanning, and can limit coordinate precision or adjust the level of boundary exposure according to privacy protection settings. For example, in complex facilities such as a performance hall, sub-zones of the boundary following the movement route between floors are recorded together, and in spaces with few obstacles such as outdoor event venues, a simplified boundary based on the radio wave environment is applied.

[0045] In the following real-time entry judgment step (S200), location samples periodically acquired by the user terminal and movement status obtained from sensors are transmitted to the server or filtered locally, after which an inspection is performed. The server corrects the center of the recent set of locations, the direction of movement, and the presence or absence of a dwell signal, and can eliminate momentary location spikes by referring to the reliability of satellite signals, the status of available satellites, and the quality of the wireless network. When the terminal approaches the boundary, the judgment delay is reduced by temporarily increasing the scan cycle, but to suppress battery consumption, it may switch to a loose cycle when there is no screen activity or movement event. Users moving in vehicles may be classified as passing-type approach by reflecting the range of speed fluctuation and road type, and are promoted to boundary entry candidates when they switch to walking and a dwell signal is confirmed for a certain period.

[0046] Once entry into the boundary is confirmed, an attendance verification notification is generated during the meeting attendance verification notification step (S300). This notification includes the identifier of the authenticating subject, the meeting identifier, the time of authentication, and the key basis for determining the stay within the boundary. If the terminal is in an environment with strong vibration or noise, the notification is displayed in a combined form of voice and visual guidance; conversely, users with accessibility settings enabled may be provided with a simple confirmation interface featuring large fonts and a wide touch area. In cases involving pre-agreed user characteristics, such as health status or the use of mobility aids, authentication is completed with a single simple action to avoid requiring excessive operation; furthermore, if the environment allows a stay-based automatic verification mode, authentication can be confirmed without any separate operation.

[0047] Once the notification generation is complete, the meeting attendance authentication transmission (S400) sends the fact of arrival to other user terminals in the same meeting. To prevent unnecessary exposure of personal information among participants, the display level of this transmission is configured differently according to role; the organizer can check the status including aggregation and summary, while general participants receive only minimal arrival information. If multiple authentications occur at the same location within a short time interval, the server naturally forms a companionship relationship and stores it as evidence for mutual verification. In this process, the individual's trustworthiness will be adjusted not as a fixed value, but by weighting it based on recent trends.

[0048] In one embodiment, the server automatically proposes the width and shape of the boundary during the boundary setting phase by referring to recommended values ​​from a model trained on past event records. In locations with large positional errors, such as urban canyons or indoors, a dwell signal is required for determining the boundary's internal position; however, in environments with stable signals, such as large-scale outdoor events, candidate determination is performed based solely on movement direction and speed predictions even if the dwell signal is weak, and authentication is confirmed through short-term observation. Since areas with severe traffic congestion are estimated to have a high probability of delay, the timing of notification generation is adjusted to occur almost simultaneously with entry into the boundary. For types of gatherings where users frequently exhibited a pattern of being late, the notification timing is advanced slightly to strengthen pre-arrival guidance. If a user is stuck for a long time while moving or their health indicators are detected to be within an abnormal range, a safety alert is sent first regardless of the authentication process, a simple notification is sent to a pre-registered third-party terminal, and these notification records are stored in the safety record repository.

[0049] In subsequent steps, the server accumulates authentication data and transmission history in user profiles and meeting profiles, respectively, and uses them as the basis for subsequent boundary correction, prediction updates, and operational statistics.

[0050] In the center of the map in Fig. 3, the meeting location is marked as a reference marker, and a circular initial geofence is placed around it. This initial boundary is a basic protection range simply calculated using only location coordinates at the time of meeting creation, and it does not adequately consider realistic factors such as indoor / outdoor distinctions, building layouts, alley curvature, and zones where radio waves can reflect. Accordingly, the server sequentially collects contextual information such as the outlines of topographic features around the location, the outlines of adjacent buildings, inter-floor access routes, the distribution of underground entrances, the locations of obstructing structures, road flow and crossing points, the weather and visibility of the day, and nearby wireless signal quality and network congestion. When an indoor space is included, auxiliary indicators such as near-field wireless identifiers, Wi-Fi fingerprints, and indoor map information are used together to correct for external satellite coordinate errors. Based on this input, the server deconstructs the initial circle to modify the boundary so that it follows only the actual movement path and the space where dwelling is possible. Furthermore, when it encounters points where alleys bend or physical boundaries such as walls or fences, it bends the boundary line to prevent through-access from being misidentified as authentication. As a result, the irregular closed lines marked on the outer edge of the drawing flow along the contours of surrounding structures and movable areas, and are repositioned into a natural shape independent of the map grid.

[0051] Correction for individual users will proceed simultaneously on the same screen. The user indicator in the bottom-left corner represents an example of a user still located outside the boundary; the server adjusts the boundary application width on a per-user basis by considering factors such as the device's movement speed and directional stability, history of repeated tardiness or absences, use of mobility aids, age group and walking speed distribution, current traffic flow, and adverse weather conditions. Since users of mobility aids may have restricted access routes and slower transition speeds, the boundary line is gently widened outward at curves and stair access points to provide a generous allowance for dwell detection. However, in sections directly adjacent to vehicular traffic lanes, a policy is applied requiring the simultaneous satisfaction of dwell signals and directional stability to exclude through-passage access. Conversely, for users who walk quickly and frequently traverse urban alleys, corrections are applied so that the boundary folds inward at intersections and no-crossing zones to prevent the boundary from becoming unnecessarily wide, ensuring that authentication does not occur solely through passing nearby areas.

[0052] The decision-making process for boundary changes may be implemented in a dual structure where a learned model proposes the change and a policy engine verifies it. The model is applied by learning authentication logs from the same location in the past, congestion patterns at similar times, floor configurations and indoor movement patterns of surrounding buildings, and the average dwell points and avoidance paths of user groups. In one embodiment, for complex facilities where indoor and outdoor areas frequently intersect, a cooling period for boundary transitions is provided to prevent frequent blinking and excessive notifications. If signal quality is insufficient, the boundary may be temporarily returned to the initial circular boundary to provide a simple confirmation interface that allows the user to assist with authentication.

[0053] Additionally, when adverse weather conditions are detected, the boundary gradually expands in a direction that diverts movement paths to avoid areas prone to slippery surfaces or crowd density. If a temporary control zone, such as road construction or a procession, is detected, the boundary line is modified to curve around that area because actual access is impossible even if the zone is within the boundary. For gatherings with many users of mobility aids, weights are applied to make the boundary more sensitive toward entrances with elevator halls and gentle ramps, and for large events where various modes of transportation are mixed, conditions are set to automatically require stay-based verification at vehicle boarding / alighting areas and pedestrian transition points.

[0054] The result of individual user correction is applied by interpreting the boundary itself differently on a user-by-user basis. Even if the same outline is displayed, the server sets different allowed dwell ranges and judgment threshold criteria for each user; consequently, even if users access the same point simultaneously, one user may be classified as an authentication candidate while another is classified as a pass-through user. In this context, it will be made clear that the difference in judgment criteria is not unfair treatment, but a technical measure to preserve accuracy based on user characteristics.

[0055] Boundary changes can be continuous rather than a one-time occurrence. When a user changes their route or surrounding congestion subsides, the outline reverts to a shape close to its original, and if new obstacles are detected or map data updates are confirmed, the outline is repositioned to reflect that information. When signal quality temporarily deteriorates, it operates conservatively by giving greater weight to the consistency of the recent coordinate set and direction vectors to prevent the boundary from becoming excessively jagged.

[0056] Through this process, the range initially represented as a prototype evolves from a simple distance radius into an irregular closed area that reflects real-world mobility and user context, enabling more accurate estimation of arrival probability for example users located outside the drawing by incorporating real-time factors.

[0057] The verification of the position sensor, accelerometer, and gyroscope sensor (S201) in Fig. 4 illustrates a process of integrating multiple signals into a single state. Satellite-based coordinates, near-field wireless identification signals, and wireless network-based base station triangulation results provide candidate locations in space, while accelerometers and gyroscopes provide dynamic patterns such as transitions between movement and stop, and straight-line movement and rotation. Magnetic force and atmospheric pressure complement indoor / outdoor identification and indications of movement between floors, and if the user consents, heart rate and gait rhythms provided by the wearable can add evidence of movement continuity and stability of stay. In complex environments where indoor and outdoor spaces intersect, beacons and wireless fingerprints are interpreted first to reduce vertical errors, and in sections with low network quality, the sample collection cycle may be temporarily lowered to suppress battery consumption while allowing switching to detailed observation only at the moment when change detection is required.

[0058] Real-time entry determination (S202) is a procedure for determining boundary approach as an entry candidate. When the stability of movement speed and direction converges closely to a walking pattern, the likelihood of a pass-through approach decreases, and it may appear as attending a gathering. During the interpretation of vehicle movement, entry is not confirmed solely by boundary contact, taking into account the road flow and the shape of the access road. If congestion actually staying around the geofence is observed, it is promoted to an entry candidate state. In places where positional error is likely to increase, such as near indoor entrances or boarding / alighting points, the presence of a proximity signal and directional consistency may be required as additional evidence.

[0059] Subsequently, the process of verifying characteristics, reliability, mode of transportation, and traffic congestion (S203) and the process of the second artificial intelligence model (S204) will use the learned model to calculate the probability of attendance and determine whether to execute the authentication trigger. The input includes personal contexts such as user-specific movement characteristics, past tendencies toward tardiness and non-attendance, and recent trends in conscientiousness, as well as the traffic congestion and weather conditions of the day, arrival patterns based on the type of gathering and time of day, and similarity regarding how closely the current route matches the destination. The model estimates the probability of arrival by synthesizing the current speed and direction, the intensity of the lingering signal, and the variability of the signal quality. When the confidence level of the prediction is low, it may recommend a conservative policy such as extending the observation or mitigating the notification intensity on the user terminal. If this result is combined with the execution conditions of the authentication trigger, it leads to the generation of an authentication notification if the probability exceeds a certain level and evidence of lingering is secured. However, if spoofing circumstances are detected, such as traces of mock location usage, abnormal speed changes, or device time manipulation, authentication is suspended, and only a simple guide for fact verification is displayed on the terminal. If the battery level or network connection status is poor, the trigger switches to a safe save-and-retransmit mode instead of immediate transmission, ensuring that synchronization with the server is reliably completed while the user remains within the boundaries.

[0060] In one embodiment, a scene is captured where a user moving along a road reaches a boarding / alighting point near a boundary and switches to walking. As the periodicity of the walking cycle becomes evident in acceleration and gyroscopes, and the direction vector converges toward the meeting point, the model's arrival probability increases the moment a short stay is confirmed. At this time, if a proximity wireless signal is detected above a certain level, it is interpreted as a high probability of entering the indoor space; thus, authentication is confirmed after a brief observation, and if the user has enabled accessibility mode, a simplified procedure is applied to complete authentication without additional operation. Conversely, while vehicle movement continues, even if contact with the boundary line is repeated, it is classified as a pass-through approach, and entry authentication does not occur. For users with a history of accumulated past non-participation, the prediction is conservatively adjusted under the same conditions to require clearer evidence of stay. In locations where movement between floors occurs indoors, the instantaneous location reliability is recalibrated, and authentication proceeds when a clear stay is confirmed.

[0061] This procedure is designed not to rely on a single sensor or rule, taking into account the limitations of positioning and the complexity of real-world movement. Noise suppression and map-based path correction in the sensor fusion stage, the requirement for evidence of presence and pass-through exclusion in the candidate determination stage, the simultaneous reflection of personal and environmental contexts in the model stage, and spoofing detection and energy policy integration in the trigger stage will work in tandem to achieve a balance that prevents accidental contact or arbitrary manipulation near boundaries from being misidentified as authentication without excessively delaying actual arrival.

[0062] In the map on the left side of Fig. 5, the meeting place is marked as a central marker, and the surrounding closed boundary appears as a curved shape following the actual terrain. On the left side of the boundary, markers of two adjacent users are arranged to appear vertically overlapping, symbolizing instances where they passed through the same section within a short time interval or stayed together just inside the boundary line. The server integrates location samples transmitted from the terminal and various signals to verify whether the movement trajectories of the two users flowed in similar directions and speeds for a certain period, whether their stays inside the boundary overlap, and whether location reliability is sufficiently maintained even during background operations with the screen off. At points where location is prone to instability, such as entrances where indoor and outdoor areas meet, or areas where one transitions to walking after disembarking from a vehicle, weight is given to the stay signal and directional consistency, and pass-through approaches involving only repeated boundary contact will be excluded from the accompaniment determination.

[0063] The determination of cohabitation is not based solely on simple temporal or coordinate proximity, but also considers cross-referenced evidence between the users. If the sensor patterns of the two terminals exhibit similar cycles, the movement state transitions to stop together followed by the capture of the same proximity signal, and there is a continuous sequence of events on the map indicating entry into the interior via the same entrance or passageway, it is interpreted as having a high probability of cohabitation. The server groups the two users into a single group only when these consecutive alignments are satisfied, and provides summarized arrival notifications based on that group. On the host screen, management indicators such as group size, the group's initial point of stay, and group dispersion are briefly displayed along with group identification information; conversely, on the general participant screen, only group-level information may be provided to prevent the exposure of excessive details regarding individual users.

[0064] Once the accompaniment determination is complete, reliability correction follows. Combinations that frequently arrived together in past records are given increased weight for repeatability, and even if they traveled different routes, the effect of mutual correction is maintained if it is confirmed that they stayed side-by-side for a certain period of time or longer in the final segment. Within a group, the user leading the arrival process and the user following them are not treated with equal weight; if the leader secures sufficient grounds for staying first and the follower follows to enter stably, the reliability of both users increases slightly. However, if the follower's stay is insufficient or abnormal speed changes occur, the correction effect will be applied only to a limited extent. To prevent fraudulent activities such as proxy attendance, the unique signature and one-time token of each device are included in the authentication message. If traces suggesting device duplication or time manipulation are detected, the accompaniment determination is suspended or invalidated, and only a brief notice for fact verification is displayed on the user's device.

[0065] In one embodiment, a scenario can be envisioned where two users who disembarked together on the outer road of the performance venue move toward the entrance and enter the interior after a brief stay relative to the boundary line. If similar walking cycles are detected in acceleration and gyroscopes, proximity wireless signals strengthen equally, and a pattern of entering the front entrance via the same rotation angle is maintained in map matching, the server determines the two users as a companion group. The arrival notification generated at this time is summarized and transmitted to other participant terminals on a group basis, and the reliability calculation unit determines the correction range by reflecting the recent conscientiousness flow and non-participation tendencies of both users. If repeated entry and exit are detected while the user stays for a long time directly above the boundary line, it is considered a waiting situation due to structural congestion at the entrance to avoid misjudgment; furthermore, for a group including a user with a mobility aid, low dispersion of movement speed is recognized as grounds for companionship, but the stay requirement is not relaxed.

[0066] To protect personal information and ensure safety, the detailed basis for companionship assessments and reliability adjustments is not exposed externally; only assessment results and summary metrics are accumulated in user and group profiles. The server stores a key summary of the assessment basis in logs to allow for the re-evaluation of companionship group creation and dissolution at any time, and maintains a level of visibility without excessive exposure even when the group expands as new participants continuously enter via the same path.

[0067] Near the left boundary of Fig. 6, an arriving user who has crossed the boundary line is displayed, and the moment the authentication of this user is confirmed, the server generates an event to select a transmission target. The transmission targets may include terminals participating in the same meeting, the organizer's managed terminal, and third-party terminals designated in advance by the user, and different display levels and scopes of disclosure are applied to each target group. If the network condition is good, transmission is performed immediately via a push path, and if the connection quality is low, a reliable path is prioritized to send a concise summary first, after which detailed information can be added when the connection is restored.

[0068] Except for the area near the left boundary, subsequent users who are still in the pre-confirmation stage are deployed, and the red trajectory on the map indicates that notifications generated by preceding arriving users are delivered to these terminals. As soon as the subsequent user's terminal receives the notification, it re-evaluates the current location, direction of movement, and stay signal, temporarily increases the observation cycle if necessary, and updates the probability of entering the boundary and the arrival prediction.

[0069] Notification transmission may be implemented not through simple broadcasting, but through role-based differential provision. The organizer's management screen displays aggregates such as the current attendance count, the time of the most recent authentication, and whether an accompaniment judgment has been made, while the general participant screen will summarize only the fact of arrival to avoid excessive exposure of other people's personal information. Third-party devices will concisely display only the user's arrival status or any abnormalities during movement, and pressing the confirmation button sends a response to the server, which is recorded in the safety log.

[0070] In one embodiment, when an arriving user who entered first from the northwest side of the boundary completes authentication, the server sends an arrival notification to the terminals of the participants in the same gathering. Immediately upon receiving this, the terminals re-evaluate their movement status to verify pedestrian transition and stay signals, and authentication is confirmed immediately. Based on the notification, user terminals on the right side may re-search for valid entry / exit routes within the boundary and correct their movement toward the destination. For example, a terminal moving along the southwest outer perimeter, which was determined to be passing through, can be guided to a nearby entrance after receiving the prior arrival notification and enter the authentication stage after a short wait. This sequence transparently shares the progress of the gathering in real time while reducing the operational burden on individual users and lowering the possibility of misrecognition due to environmental variables such as entrance congestion.

[0071] When the entry of a user is confirmed near the boundary on the left side of the map in Fig. 7, the server determines the transmission target and the level of disclosure. Since other users in the center of the map are still in the pre-definitive authentication stage, a lightweight notification arrives first to them, along with the fact of arrival, prompting them to re-evaluate their current location and direction of movement.

[0072] The right side of the diagram represents notification to a third-party terminal, namely a trusted contact such as family, acquaintances, or guardians. This channel is activated only when explicit consent from the party is secured at the time of service registration or during the process of participating in a gathering. The notification includes only minimal information directly related to safety and security, such as whether the destination has been reached or if there were any abnormalities during transit. A confirmation button is provided on the guardian's terminal to send a response to the server confirming receipt, and this response information is stored in the safety record database to serve as a basis for responding to disputes and improving quality after the fact.

[0073] In situations where abnormal information regarding user characteristics is detected, the scope of disclosure will be adjusted more finely. When movement discontinuities or abnormal biometric responses are detected by the device's sensors or wearable indicators, the server first displays a status check notice on the individual's device and sends a brief warning to the guardian's device only if immediate danger is presumed. To prevent excessive exposure of personal information, the meeting participant's device may display only a condensed summary indicating a possible schedule delay or the need for assistance, instead of a detailed reason. The trigger for anomaly notifications is based on continuous observation rather than a single measurement, and to prevent misjudgment, certainty may be increased through short-term verification observations.

[0074] In one embodiment, when arrival is confirmed at the northwest boundary, the server sends an arrival summary to the meeting participants and simultaneously sends a safe arrival notification to pre-registered family members and guardians. The user at the center of the map re-evaluates their movement status immediately after receiving this notification, and the authentication process is reliably completed as authentication proceeds upon confirmation of walking transition and stay signals. If an abnormal heart rate pattern is continuously detected by the central user's wearable, the server first displays a rest recommendation to the user's terminal and suggests switching to a mode with reduced future notification frequency, along with a simple warning to the guardian's terminal.

[0075] Although specific embodiments of the method for authenticating meeting attendance and the system for such authentication according to the present invention have been described above, the present invention is not limited to the above embodiments. Within the scope of the technical spirit and essence of the present invention, various modifications, changes, additions, or substitutions may be made by those skilled in the art, and all such modified embodiments should also be interpreted as being included within the scope of the rights of the present invention.

Claims

Claim 1 A method for authenticating meeting attendance performed by a server including a processor and memory, wherein the server sets an initial geofence area based on the coordinates of the meeting location, and a user-specific reliability calculated based on at least one of the following: a user's arrival time history, meeting attendance history, meeting tardiness history, meeting cancellation history, and meeting non-attendance history; (b) a step of dynamically changing the initial geofence area on an individual user basis by reflecting user characteristics including at least one of the user's physical condition, use of mobility assistive devices, age, movement speed, and means of transportation; (b) a step in which the server determines in real time whether to enter the dynamically changed geofence area using at least one sensor data among a location sensor, an accelerometer, and a gyroscope received from the user terminal, and calculates the probability of the user attending the meeting in real time using an artificial intelligence model that takes at least one of the user-specific reliability, user characteristics, means of transportation, and traffic congestion as input, and determines whether to authenticate attendance at the meeting based on the entry determination result and the attendance probability; (c) a step in which the server generates a notification authenticating attendance at the meeting to the user according to the determination, wherein if multiple user terminals belonging to the same meeting generate the notification within a preset time interval, the multiple users are determined to be a group of companions, and the user-specific reliability is corrected based on the mutual behavior of the group of companions; (d) a step in which the server transmits the generated attendance authentication to other user terminals of the meeting; comprising a meeting attendance authentication method. method Claim 2 A method for authenticating attendance at a meeting, wherein, in claim 1, the geofence area dynamically changed in step (a) is an irregularly formed closed area that additionally reflects at least one of the topographical characteristics of the meeting location, weather conditions, the number of floors of a building, and geographical obstacles. Claim 3 A method for authenticating attendance at a meeting according to claim 1, wherein step (b) confirms entry only when the stability of the user's movement speed and direction of movement converges to a walking pattern and a stay signal within the geofence area is confirmed for a preset period or longer, and does not confirm entry based solely on contact with the geofence area while it is determined to be vehicle movement. Claim 4 A meeting attendance authentication method according to claim 1, wherein step (d) is characterized by transmitting a meeting attendance authentication including abnormal details to a third-party terminal when a physical condition or movement abnormality of the user is detected. Claim 5 In a meeting attendance authentication system comprising a central processing unit and a memory, the central processing unit executes instructions for executing a meeting attendance authentication method stored in the memory, wherein the method comprises: (a) a server setting an initial geofence area based on the coordinates of a meeting location, wherein the server has a user-specific reliability calculated based on at least one of the following: a user's arrival time history, meeting attendance history, meeting tardiness history, meeting cancellation history, and meeting non-attendance history; (b) a step of dynamically changing the initial geofence area on an individual user basis by reflecting user characteristics including at least one of the user's physical condition, use of mobility assistive devices, age, movement speed, and means of transportation; (b) a step in which the server determines in real time whether to enter the dynamically changed geofence area using at least one sensor data among a location sensor, an accelerometer, and a gyroscope received from the user terminal, and calculates the probability of the user attending the meeting in real time using an artificial intelligence model that takes at least one of the user-specific reliability, user characteristics, means of transportation, and traffic congestion as input, and determines whether to authenticate attendance at the meeting based on the entry determination result and the attendance probability; (c) a step in which the server generates a notification authenticating attendance at the meeting to the user according to the determination, wherein if multiple user terminals belonging to the same meeting generate the notification within a preset time interval, the multiple users are determined to be a group of companions, and the user-specific reliability is corrected based on the mutual behavior of the group of companions; (d) a step in which the server transmits the generated attendance authentication to other user terminals of the meeting; comprising a meeting attendance authentication method. System

Images