Information processing device, information processing method, program, and system

Abstract

This disclosure aims to facilitate smoother and more effective communication by accurately conveying the intended emotions of the message sender to the recipient while providing optimal notifications that take into account the recipient's situation. [Solution] This invention provides an information processing device, method, program, and system that integrates and analyzes the emotions contained in messages entered by parents and real-time contextual information (location, time, schedule deviation status, etc.) obtained from child monitoring devices, and dynamically optimizes the voice expression of messages (volume, tone, speed, etc.) and non-voice notification methods (screen display, vibration, etc.) based on the results to deliver them to children. In particular, parents can choose between fixed rules and AI-driven automatic learning for determining the notification style.

Description

Technical Field

[0001] The present disclosure relates to an information processing apparatus, an information processing method, a program, and a system.

Background Art

[0002] In a conventional child monitoring device, the message function mainly involves playing fixed phrases and mechanical voice readings with little emotion, and there is a problem that nuances of emotions such as the worry and joy of guardians are difficult to convey to children. For this reason, the subtlety of the emotions that the guardians want to convey is not fully conveyed, and there has been a problem that parent-child communication cannot be smoothly carried out.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The problem to be solved by the present disclosure is to provide an information processing apparatus improved from the conventional one.

Means for Solving the Problems

[0005] An information processing apparatus, comprising: a receiving unit that receives a message transmitted from a terminal device; an emotion estimation unit that estimates an emotion included in the message; a notification determination unit that determines a notification format from the emotion; and a transmission unit that transmits the message to a GPS terminal in the determined notification format.

Brief Description of the Drawings

[0006] [Figure 1] It is a schematic diagram showing a system configuration according to an embodiment of the present disclosure. [Figure 2]This figure shows the functional block configuration of an information processing device (server) according to one embodiment of the present disclosure. [Figure 3] This is a flowchart of a context-adaptive message notification process according to one embodiment of the present disclosure. [Figure 4] This figure shows an example of the structure of notification rule data according to one embodiment of this disclosure. [Figure 5] This figure shows the hardware configuration of a child monitoring device according to one embodiment of the present disclosure. [Figure 6A] This figure shows an example of an emotion icon display according to one embodiment of the present disclosure. [Figure 6B] This figure shows an example of an emergency icon display according to one embodiment of the present disclosure. [Figure 7] This figure shows an overview of an AI learning model according to one embodiment of this disclosure. [Figure 8] This figure shows an example of the screen configuration of a parental device according to one embodiment of the present disclosure. [Figure 9] This is a message notification processing flowchart for Example 1 (concern message at school) according to one embodiment of the present disclosure. [Figure 10] This is a flowchart of the message notification process for Example 2 (Emergency Message in a Hazardous Area) according to one embodiment of the present disclosure. [Figure 11] This is a flowchart for determining the notification format in Example 3 (Automated Learning by AI) according to one embodiment of the present disclosure. [Modes for carrying out the invention]

[0007] The embodiments of this disclosure will be described in detail below with reference to the drawings.

[0008] In addition to the above, the technology disclosed herein also has problems that undermine children's psychological safety, such as when a message is played at a uniform volume without considering the child's TPO (time, place, and situation), causing the child to disturb others or feel embarrassed if a message is played at a high volume during class. In particular, even when it is necessary to quickly and reliably transmit a message in an emergency, appropriate notifications are not made according to the surrounding circumstances, resulting in insufficient practicality and reliability as a monitoring function.

[0009] First, an overview of a context-adaptive message-to-voice conversion system 100 according to one embodiment of this disclosure will be described using Figure 1. This system 100 mainly consists of a parent terminal 10, a server 20, and a child monitoring device 30.

[0010] However, the functions performed by each of these components are not necessarily limited to the illustrated devices, and may be appropriately distributed or shared among multiple information processing devices constituting this system (for example, the parent terminal 10, the child monitoring device 30, or other terminal devices that cooperate with them, or a cloud-based distributed computing environment). For example, some or all of a series of processes such as message reception, emotion estimation, notification format determination, and message transmission may be performed directly between the parent terminal 10 and the child monitoring device 30 via P2P (Peer-to-Peer) communication without going through the server 20.

[0011] The child monitoring device 30 is assumed to be a device with GPS functionality, as described later, but the final recipient of message notifications in this system is not limited to this. For example, it may also include a configuration in which non-GPS devices such as smartwatches or smart speakers that do not have direct GPS functionality send message notifications via a hub device in the home. The parent terminal 10 is a device for parents to input and send messages to their children. The server 20 has a receiving unit that receives messages sent from the parent terminal 10 and provides a core function that estimates the emotions contained in the message and determines the optimal notification format based on context information (situation information) obtained from the child monitoring device 30. The child monitoring device 30 is equipped with GPS functionality and obtains context information indicating the child's current location and surrounding environment, and sends it to the server 20.

[0012] Furthermore, messages sent from server 20 are delivered to the child using the most appropriate voice or non-voice notification method based on a predetermined notification format. This system aims to promote smooth and heartwarming communication between parents and children and improve the child's psychological safety by comprehensively analyzing the emotions contained in the parent's message and the child's real-time situation, and intelligently mediating and optimizing communication between the two.

[0013] Note that although this system is called a "Context-Adaptive Message Voice Conversion System", it does not necessarily require voice notifications for message transmission. Depending on the situation and the purpose, configurations that use only non-voice notifications or mainly non-voice notifications are also included in the technical idea of this disclosure. The "terminal device" in this system is not necessarily limited to an input device directly operated by a human, but refers to any device that generates and transmits messages and data serving as information sources. This includes various sensor devices such as wearable sensors, environmental sensors, and IoT devices that automatically detect and generate information about a child's situation. Thus, even when a message is automatically generated and transmitted without direct human input, it can be broadly interpreted as a message reception from the "terminal device" of this application.

[0014] Next, the hardware configuration of the child monitoring device 30 will be described using FIG. 5. The child monitoring device 30 includes a GPS receiver 301, a communication unit 302, an audio output unit 303, a display unit 304, a vibration unit 305, a storage unit 306, and a control unit 307.

[0015] The GPS receiver 301 receives signals from GPS satellites and obtains the current position information of the child monitoring device 30. The obtained position information is transmitted to the server 20 as context information.

[0016] The communication unit 302 is a wireless communication module for transmitting and receiving messages, context information, notification formats, etc. with the server 20.

[0017] The audio output unit 303 is a speaker, an audio synthesis circuit, etc. that outputs the message received from the server 20 as audio based on the determined audio parameters.

[0018] The display unit 304 is a display that displays icons, text, etc. indicating the content and emotion of the message based on the notification format received from the server 20.

[0019] The vibration unit 305 generates different vibration patterns based on the notification format received from the server 20, depending on the urgency and type of emotion of the message, and notifies the child.

[0020] The memory unit 306 stores the OS, application programs, various setting data, and history of received messages necessary for the operation of the child monitoring device 30.

[0021] The control unit 307 is a processor or microcontroller that comprehensively controls the operation of the entire child monitoring device 30. Specifically, according to the program stored in the memory unit 306, it controls the components of the voice output unit 303, display unit 304, and vibration unit 305 based on the acquisition of location information from the GPS receiver unit 301, communication with the server 20 via the communication unit 302, and the notification format received from the server 20, thereby delivering messages to the child in the most optimal way. Although not shown, the child monitoring device 30 may also be equipped with various sensors such as an accelerometer, gyroscope, heart rate sensor, body temperature sensor, ambient volume / illumination sensor, camera, and microphone. These sensors are used to acquire more detailed contextual information, such as the child's activity level, biometric information, surrounding environment information, or the child's facial expressions and voice, and are transmitted to the contextual information acquisition unit 203 via the control unit 307.

[0022] The parent terminal 10 is an information processing device such as a typical smartphone or tablet, and includes an input unit (touch panel, microphone, etc.), a display unit, a communication unit, a storage unit, a control unit, etc.

[0023] Server 20 is a typical computer system, equipped with a processor, storage device, communication interface, input / output interface, etc. These hardware resources are implemented in software as functional blocks, as described later, and work together in coordination.

[0024] Next, the functional block configuration of server 20 will be explained using Figure 2. Server 20 comprises a receiving unit 201, an emotion estimation unit 202, a context information acquisition unit 203, a notification determination unit 204, and a transmission unit 205. These are functional blocks executed by the server 20's processor and are implemented by programs stored in the memory unit.

[0025] The receiving unit 201 receives text or voice messages sent from the parent terminal 10. Along with the message, it also receives identification information (such as a user ID) of the parent terminal 10 that sent the message. Furthermore, in addition to the text or voice messages described above, the receiving unit 201 may also receive mechanically generated information, such as structured event data, biometric data, or environmental data transmitted from various sensor devices, as "messages." Even if this data does not directly take the form of human language, it is treated as a "message" because it indicates the sender's intent or situation (for example, a situation where the parent is hastily sending a message based on GPS data from the parent terminal). In this case, the receiving unit 201 may include a function to convert this structured data into a format that the emotion estimation unit 202 can process.

[0026] The emotion estimation unit 202 estimates the emotions contained in the message received by the receiving unit 201. If the message is in text format, it uses natural language processing (NLP) techniques (e.g., a fine-tuned emotion vocabulary dictionary or a pre-trained language model such as BERT) to classify it into multiple emotion categories such as joy, worry, encouragement, anger, sadness, and urgency, and estimates the strength (score) of each emotion. If the message is in audio format, it estimates the emotions using a combination of speech recognition techniques to convert the audio to text and perform NLP processing, and speech emotion recognition (SER) techniques (directly estimating emotions from speech features such as pitch, volume, speech rate, and voice quality). SER is effective in capturing the nuances of emotion contained in the tone of voice. The main types of emotions to be estimated are assumed to be "joy," "worry," "encouragement," "anger," "sadness," and "urgency." In particular, "urgency" is designed to be assigned special processing different from other emotions (e.g., high volume, highest priority notification). Furthermore, the emotion estimation unit 202 can estimate the sender's intentions, emotions, or equivalent "importance" or "urgency" embedded in a message, not only from the content of the message (text, voice) but also from indirect information sources such as the sender's operation history (e.g., message input speed, emoji selection, presence or absence of specific emphasis or keywords, time interval until sending, modification history during input, etc.) and the sender's biometric information obtained from the parent terminal 10 (e.g., heart rate, skin potential, voice tone, finger pressure intensity, etc.). Even if this information is not directly classified as "emotion," it can be used as an indicator of notification priority or characteristics, such as "extremely important," "requires urgent action," or "warning," and is included in the concept of emotion estimation in this application. Furthermore, the "emotions" estimated by the emotion estimation unit 202 are not limited to being output as specific category names (e.g., joy, worry), but may also be numerical scores (e.g., confidence levels for each emotion category), multidimensional vectors (e.g., features representing the intensity or polarity of emotions), or any representation format that other machine learning models can use as input (e.g., emotion embeddings, two-dimensional emotion models in affective computing (such as Valence-Arousal))."Emotions" expressed in these formats are also treated as "emotions" in this application for determining the notification format. The emotion estimation unit 202 selects the optimal emotion estimation technique according to the message format (text / voice) and other related information, and appropriately estimates the emotion or equivalent intention or importance contained in the message (Claim 3). Furthermore, the emotion estimation unit 202 may estimate emotions by combining a wide range of information sources, such as image recognition technology (e.g., facial expression recognition), motion sensor data (e.g., gestures, body movements), and environmental sensor data (e.g., room temperature, humidity, illuminance, noise level). For example, by estimating joy from the guardian's facial expression, impatience from the pressure of the fingers when entering the message, or urgency of the message from the ambient noise level, and making a comprehensive judgment based on these, it becomes possible to estimate emotions from multiple perspectives and with higher accuracy.

[0027] The context information acquisition unit 203 acquires real-time context information (situation information) from the child monitoring device 30. The context information includes GPS location information, the type of location identified from the location information, the current time of day, schedule deviation status, geofencing status, etc. The child's current location is identified from the GPS location information, and the type of location is determined by comparing it with registered geofences (school, home, cram school, park, etc.). Note that the method of acquiring location information is not limited to GPS, and the location may be determined based on various wireless signals and sensor information that the child monitoring device 30 (or a device linked to it) can acquire, such as the SSID and signal strength of a Wi-Fi access point, the signal strength of a Bluetooth beacon, the radio wave strength (cell ID) from a base station, or other positioning technologies (e.g., indoor positioning system, dead reckoning using inertial sensors, LiDAR, RFID, etc.). Location information acquired by these methods is also included in "location information" in this application, which enables more accurate situational judgment both indoors and outdoors. The time of day determines whether the current time is daytime, nighttime, late night, or a specific activity time (e.g., lunch break, bedtime, lesson time). The schedule deviation status compares the child's schedule information, which has been set in advance by the guardian, with the current location information and time of day to determine whether or not the child is deviating from the schedule. The geofence status detects entry into or exit from a specific area (geofence), and the notification format is adjusted accordingly. In addition, the context information acquisition unit 203 may also acquire biometric information (activity level, heart rate, body temperature, sleep state, alertness level, etc.), environmental information (ambient volume, illuminance, temperature, humidity, air quality, etc.), and device usage status (screen on status, other application running status, battery level, network connection status, etc.) as context information, which are obtained from various sensors installed in or linked to the child monitoring device 30 (e.g., accelerometer, gyroscope, heart rate sensor, body temperature sensor, microphone, camera, etc.). For example, it becomes possible to make more detailed situational judgments, such as when a child's heart rate is abnormally high, when the surroundings are extremely noisy, or when the device has not been operated for a long time (is unresponsive).

[0028] The notification determination unit 204 comprehensively evaluates the emotion estimated by the emotion estimation unit 202 and the context information (situation information) acquired by the context information acquisition unit 203 to determine the optimal notification format, including the message's audio parameters (volume, speaking speed, tone of voice, intonation, etc.) and non-audio parameters (icon display, vibration pattern, number of replays, light, heat, biostimulation, etc.). The notification determination unit 204 determines the notification format based on both the emotion information received from the emotion estimation unit 202 and the situation information received from the context information acquisition unit 203 (Claim 2). If the emotion estimation unit 202 determines that the emotion contained in the message indicates urgency (Claim 4), the notification determination unit 204 prioritizes determining a specific notification format compared to other emotion cases. For example, it may maximize the volume, add a specific warning sound, make the display unit 304 flash a warning icon (Figure 6B), or generate strong, continuous vibrations in the vibration unit 305. Furthermore, as situational information, the system acquires at least one of the following from the GPS terminal 30: "location information," "type of location," and "current time of day." Based on this, it determines at least one of the following audio parameters of the message: "volume, speaking speed, tone of voice, intonation," and at least one of the following non-audio parameters: "icon display, vibration pattern, number of plays" (Claim 5). If the notification determination unit 204 determines that the type of location is "school or sleeping place," or that the current time of day is "during class or bedtime," it minimizes or silences the volume and prioritizes determining a notification based on the non-audio parameters (Claim 6). This allows the system to notify the child of an incoming message without disturbing their learning or rest. If the situational information includes the "schedule deviation status" of the GPS terminal 30 (Claim 7), the notification determination unit 204 determines the notification format based on the emotion estimated by the emotion estimation unit 202 and this schedule deviation status. For example, if a schedule deviation is detected, such as a child being at school past the scheduled dismissal time, and the emotion is "worried," adjustments will be made to increase the urgency of the notification (increase volume and notification frequency).Furthermore, the notification determination unit 204 may further fine-tune the notification format by considering the recipient's past response history and characteristic information (e.g., hearing impairment, visual impairment, age, sensitivity to specific notifications, response history to specific sounds or lights, preferences, etc.) to ensure that the determined notification format is optimal for the recipient. For example, for a child who is sensitive to a particular vibration pattern, adjustments may be made to emphasize or suppress that vibration pattern. Also, if there is a history of the recipient noticing a message with a particular notification format in the past, adaptive adjustments may be made to avoid that format and select a different format.

[0029] The logic for determining the notification style can be selected between a "fixed rule-based logic" and an "artificial intelligence (AI)-based automated learning logic," depending on the settings made by the parent (Claim 9). In fixed rule mode, the notification style is determined based on a "notification rule set" (see Figure 4) that has been set in advance by the parent or customized by the parent. This option is provided on the settings screen of the parent terminal 10, as shown in Figure 8, and can be set by the parent via the GUI. It is defined by mapping a specific notification style to a specific combination of emotion and specific situational information. In AI-based automated learning mode, past notification history, feedback from the parent (e.g., evaluations such as "was this notification appropriate / inappropriate"), and the child's response (e.g., message opening time, whether a reply was made, changes in behavior after notification (e.g., whether the child moved away from a dangerous area after receiving the notification, or returned to their schedule)) are continuously collected and analyzed, and the logic for determining the notification style is optimized using a machine learning model (e.g., reinforcement learning, supervised learning). As shown in Figure 7, the AI ​​learning model consists of a neural network with an input layer (emotion, context information), an intermediate layer, and an output layer (notification style parameters), and learns the optimal mapping based on the training data. The AI ​​learns which notification style was most effective for a particular combination of emotion and situation, and dynamically determines a new notification style based on the learning results. If automatic learning by this AI is selected, the AI ​​learning model uses data such as past notification history, feedback from parents, and information on the child's response as training data, and continuously learns to derive the optimal notification style from emotion and situation information (Claim 10). Here, "determining the notification style" includes not only selecting from existing notification styles, but also the AI ​​learning model "generating" or "optimally adjusting" speech synthesis parameters (e.g., pitch, tone quality, subtle intonation patterns for emotional expression), animation of displayed icons, vibration patterns, light flashing patterns, etc., in real time based on estimated emotion and context information.For example, the concept of "determining the notification style" in this application also includes cases where the AI ​​creates more nuanced and personalized notification styles that do not exist in the existing training data, by directly inputting emotion scores or emotion vectors. Furthermore, the AI ​​learning model can also exploratoryally find the optimal notification style by using reinforcement learning techniques. In this case, the logic for determining the notification style is optimized by using not only feedback from parents and the child's reactions, but also the child's behavioral changes after notification (e.g., whether they moved away from the danger zone after receiving the notification, returned to their schedule, or continued to ignore the notification) as rewards. This leads to the derivation of more practical and effective notification styles, contributing to the achievement of the system's objectives. In addition, by introducing the concept of transfer learning and using pre-trained general-purpose emotion recognition models or speech synthesis models, and then fine-tuning them with data specific to this system, efficient and highly accurate learning can also be achieved.

[0030] The transmitting unit 205 transmits a message to the child monitoring device 30 along with the determined notification format. The child monitoring device 30 plays and notifies the message based on the received notification format. At this time, the display unit 304 and vibration unit 305 of the child monitoring device 30 perform a notification based on non-voice parameters based on the transmitted notification format. Depending on the type of emotion and urgency contained in the message, at least one of the visual characteristics of the icon display (such as the type of icon, color, flashing speed, and animation as shown in Figures 6A and 6B) and the type or intensity of the vibration pattern (e.g., weak, medium, strong, continuous, intermittent, etc.) is changed (Claim 8).

[0031] Next, the structure of the notification rule data 400 will be explained using Figure 4. The notification rule data 400 is an example of a data structure that the notification determination unit 204 refers to when it operates in fixed rule mode. This data includes fields for emotion category 401, context condition 402, voice parameter 403, and non-voice parameter 404.

[0032] Emotion category 401 stores the type of emotion inferred from the message (e.g., joy, worry, encouragement, urgency, etc.).

[0033] Context condition 402 stores status conditions (e.g., type of location (school, home, park), time of day (daytime, nighttime, during class), schedule deviation status, geofence status, etc.) to determine the notification format, and multiple conditions can be combined using AND / OR.

[0034] The voice parameter 403 stores parameters related to voice notifications that are applied when the emotion category 401 and contextual conditions 402 are matched (e.g., volume (minimum, normal, loud), speech rate (slow, normal, fast), tone (gentle, normal, harsh), intonation (flat, normal, rich), etc.).

[0035] Non-audio parameter 404 stores parameters related to non-audio notifications that are applied when the emotion category 401 and contextual conditions 402 are matched (e.g., icon display (icon ID, animation, color), vibration pattern (pattern ID, intensity, duration), number of plays, etc.). For icon displays in non-audio notifications, as shown in Figures 6A and 6B, multiple icons are provided to visually represent the estimated type of emotion (joy, worry, encouragement, urgency, etc.) and its intensity. In particular, for "urgent" emotions, a red warning mark or a flashing icon is displayed. The animation (flashing speed, scaling) and color of the icon are changed according to the intensity of the emotion. For vibration patterns in non-audio notifications, different patterns are assigned depending on the type of emotion and urgency. For example, for "urgent" messages, a strong, continuous vibration or a pattern with a specific rhythm is generated to immediately draw attention and convey the importance of the message. The intensity and duration of the vibration are also changed according to the intensity of the emotion and the urgency of the notification. Furthermore, non-audio parameters 404 may also include parameters related to light notifications (e.g., LED flashing patterns, color, brightness, pulse width), thermal notifications (e.g., temperature changes in specific areas, cooling / heating patterns), and biostimulation notifications (e.g., weak electrical stimulation, ultrasonic stimulation, temperature changes due to Peltier elements). This enables more reliable and flexible notifications by appealing to a variety of sensory organs. For example, for recipients with hearing impairments, audio parameters can be disabled, and a notification style combining light, vibration, heat, and biostimulation can be applied. It is also possible to integrate with modules (not shown) that provide odor (olfactory stimulation) or subtle gustatory stimulation, and include these as non-audio parameters.

[0036] Next, Figure 3 will be used to explain the flow of the context-adaptive message notification process. This process is executed in the following steps. First, in step S101, the receiving unit 201 of the server 20 receives a message in text or voice format from the parent terminal 10. At this time, the receiving unit 201 also receives metadata regarding the message content (text data, voice data), the sender's identification information, the date and time the message was sent, and the format of the message (text, voice).

[0037] Next, in step S102, the emotion estimation unit 202 of the server 20 estimates the emotion from the received message. For text messages, natural language processing is applied, and for voice messages, voice emotion recognition is applied to estimate emotion categories such as joy, worry, encouragement, anger, sadness, and urgency, and their respective strengths. Here, the emotion estimation unit 202 comprehensively analyzes not only the content of the message but also the sender's operation history and biometric information (not shown) to estimate more detailed emotions (e.g., strong worry, mild joy) or importance (e.g., high importance, medium importance).

[0038] Next, in step S103, the context information acquisition unit 203 of the server 20 acquires real-time context information (GPS location, type of location, time of day, schedule deviation status, geofencing status, etc.) from the child monitoring device 30. Furthermore, the context information acquisition unit 203 also acquires biometric information such as the child's activity level, ambient noise level, brightness, the child's heart rate, body temperature, and sleep state from various sensors installed in the child monitoring device 30 (e.g., accelerometer, microphone, illuminance sensor, etc.), and integrates this as current situation information.

[0039] Subsequently, in step S104, the notification determination unit 204 of the server 20 selects either "fixed rule" or "AI-based automatic learning" as the logic for determining the notification format, based on the settings made by the parent. This setting is provided on the settings screen of the parent terminal 10, as shown in Figure 8, and can be flexibly switched according to the parent's communication policy and the child's characteristics.

[0040] Once a notification logic is selected, the notification format is determined in step S105. Specifically, if "fixed rule" is selected (step S105A), the notification determination unit 204, based on the estimated emotion and acquired contextual information, refers to predefined notification rule data 400 (Figure 4) to determine the notification format that includes the optimal voice and non-voice parameters. In this case, the notification determination unit 204 preferentially applies the following decision logic. (1) Urgency determination: If the emotion estimation unit 202 determines that the message indicates urgency (Claim 4), a specific notification method (e.g., maximum volume, addition of warning sound, flashing icon, strong continuous vibration) is preferentially determined compared to other emotions. This determination prioritizes ensuring that highly urgent messages are reliably delivered to the recipient. (2) Volume control and non-audio priority based on location and time: If the location is determined to be a "school" or "bedroom," or if the current time is determined to be "during class" or "bedtime," the volume will be minimized or muted, and notifications based on non-audio parameters (e.g., icon display, vibration, light) will be prioritized. This ensures consideration for those around the child and the child's psychological safety while still conveying the arrival of a message. (3) Coordination with schedule deviations: If a schedule deviation is detected (Claim 7), the urgency and frequency of the notification are adjusted by combining the emotion (especially "worry" or "urgency") with the deviation. For example, if a schedule deviation is detected, such as a child being at school past the dismissal time, and the emotion is "worry," the urgency of the notification is increased (volume and notification frequency are increased) to more strongly convey the parent's intentions. (4) Consideration of recipient characteristics: The notification determination unit 204 may further refine the notification style by referring to the recipient's characteristic information (e.g., presence or absence of hearing impairment, sensitivity to specific notifications, history of reactions to specific sounds or lights, preferences, etc.) held in the memory unit 306 of the child monitoring device 30. For example, for a child with a hearing impairment, the notification style will be determined by completely disabling the voice parameter and combining non-voice notifications such as icon display, vibration, light, or biostimulation. Adaptive adjustments will also be made, such as avoiding a notification style and selecting a different style if there is a history of not noticing a message with a particular notification style in the past. If "automatic learning by AI" is selected (step S105B), the AI ​​model in the notification determination unit 204 uses past notification history, feedback from guardians, and information on the child's reactions as learning data to derive the optimal notification style from estimated emotions and contextual information. The AI ​​model is, for example, a neural network based on deep learning. It receives a vector in its input layer that combines sentiment scores, contextual information (such as quantified location, time, and environmental information), and receiver characteristic information. The output layer outputs the optimal combination of speech parameters (numerical values ​​representing volume, speech rate, tone, and intonation) and non-speech parameters (such as icon ID, vibration pattern ID, light pattern ID, thermal change, and biostimulus intensity). During the learning process, the system collects feedback on the child's behavior after a specific notification style is applied (such as opening the message, replying, or moving to a different location) and evaluations from parents. This feedback is used to perform reinforcement learning, updating the parameters of the AI ​​model. As a result, the AI ​​automatically generates more effective and personalized notification styles over time, continuously optimizing the overall communication efficiency of the system.

[0041] Next, in step S106, the transmission unit 205 of the server 20 sends the message to the child monitoring device 30 along with the determined notification format. The transmission unit 205 also compresses or encrypts the message data as needed to ensure secure and efficient data transmission.

[0042] Then, in step S107, the child monitoring device 30 transmits a message to the child using the most appropriate voice expression (voice output unit 303) or non-voice notification (display unit 304, vibration unit 305) based on the received notification format. In this case, notifications based on non-voice parameters change the visual characteristics of the icon display (Figures 6A, 6B) or the type or intensity of the vibration pattern according to the type of emotion and urgency (Claim 8). In addition, the control unit 307 drives various non-voice notification means according to the notification format, such as light notifications (not shown, e.g., flashing patterns or color changes of LED indicators), thermal notifications (not shown, e.g., temperature changes on the device surface), and biostimulation notifications (not shown, e.g., skin stimulation by weak electric current or vibration), to convey the intent and importance of the message to the child. As a result, the child can reliably recognize the message and respond appropriately, regardless of the surrounding circumstances or their own state.

[0043] In this embodiment, "GPS terminal" refers to a terminal device having GPS functionality and capable of acquiring location information, and in this embodiment corresponds to the child monitoring device 30. However, the "destination" of messages in this disclosure is not necessarily directly limited to this GPS terminal. For example, a configuration in which a message and notification format are sent to another notification device (e.g., a smartwatch, smart speaker, tablet, smartphone, VR / AR device, etc. that does not have GPS functionality) based on context information (situation information) acquired by the GPS terminal (child monitoring device 30), and the notification device ultimately notifies the message, is also within the scope of the technical idea of ​​this application. In this case, the GPS terminal functions as a source of context information and / or a relay point for message notification. "Notification format" is a general term for the method of expression when conveying a message to a child, and refers to a combination of voice parameters (volume, speaking speed, tone of voice, intonation, etc.) and non-voice parameters (icon display, vibration pattern, number of plays, light, heat, biostimulation, etc.). "Contextual information" refers to information indicating the child's real-time situation and includes GPS location information, type of location, current time of day, schedule deviation status, geofencing status, biometric information (heart rate, activity level, etc.), environmental information (noise level, illuminance, etc.), and device usage status. This notification format is not limited to a form that includes all voice and non-voice parameters, and may consist of only one or a combination of both. In particular, notification formats that do not include voice parameters and convey the emotion or importance of the message only through non-voice parameters (e.g., icon display, vibration, light, heat, biostimulation, etc.) are also included in the scope of the notification format of this application. Furthermore, "message" may be text, voice, image, video, or a combination thereof, or structured data such as sensor data or biometric information. This data is treated as a "message" of this application, and may be associated with emotions and importance automatically estimated by the system, as well as emotions explicitly entered by the sender.

[0044] This embodiment applies to message notifications from parents to children, but the scope of this disclosure is not limited thereto. For example, it can be applied to various communication scenarios, such as a monitoring system in which a caregiver sends messages to a person being cared for, or an educational system in which a teacher notifies students of important information. Furthermore, the term "message" in this application is not necessarily limited to human-readable linguistic information (text, voice), but refers to all information that a sender has sent with a specific intention or emotion. For example, even mechanically generated or acquired information, such as sensor data, biometric information, or specific behavioral patterns, can be included in the "message" of this application if it is associated with emotion or a specific intention. With this modification, emotion estimation technology is not limited to text-based emotion analysis or voice emotion recognition, but may be combined with technologies that estimate emotions from facial expression recognition or biometric information (heart rate, skin potential, etc.). Specifically, the emotion estimation unit 202's functionality also includes detecting the physical state of an object (e.g., vibration patterns, temperature, position changes) using sensors (sound wave sensors, thermal sensors, acceleration sensors, etc.) and indirectly estimating the sender's intentions and situation (e.g., emotions such as anxiety, relief, or danger detection, or similar states) from the change patterns. Contextual information can also be configured to utilize a wider range of information, such as device sensor information (temperature, humidity, acceleration), ambient sound environment, and information about people in the surroundings, to perform more detailed situational judgments. Regarding notification methods, a combination of diverse methods, such as display on VR / AR devices, olfactory and gustatory stimulation, and haptic feedback, can be considered to create more immersive or reliable means of information transmission. In addition, notification methods can include configurations that indirectly convey the "importance" or "emotion" of a message by introducing subtle changes in biometric information (body temperature, heart rate, skin conductivity, etc.). This includes, for example, technologies that induce subtle changes in sensation in the recipient (child) that are either unconscious or only noticeable as a form of information transmission, by applying very weak electrical stimuli from a wearable device to a biosensor or by applying temperature changes using a tiny Peltier element.This makes it possible to effectively convey urgency or specific emotions by using highly confidential notifications that are less likely to be noticed by those around, or by appealing to more direct physical sensations. Furthermore, the term "information processing device" in this application is not necessarily limited to a single physical device, but can refer to an entire system in which multiple information processing devices cooperate to perform each of the functions described in the claims. For example, the functional blocks of Claim 1, such as the "receiving unit," "emotion estimation unit," "notification determination unit," and "transmission unit," can be distributed among parent terminals, child monitoring devices, servers, etc., and operate in coordination via a network. This can provide advantages such as load balancing for processing, improved system availability, or operation in an offline environment. In addition, the term "terminal device" in this application broadly refers to any information processing device that is the source of a message, and includes not only human-operated devices such as parent terminals, but also sensor devices and IoT devices that detect specific situations or events and automatically generate and transmit them as messages, or software modules that transmit information (treated as messages) based on specific triggers defined within the system. This clarifies that even when the source of the "message" received by the "receiving unit" is an automatically generated information source without human intervention, it falls within the scope of the technical concept of this application. In particular, depending on the content and context of the message, or the characteristics of the recipient (e.g., hearing impairment), it is possible to configure the system to convey the emotion, importance, and content of the message solely through icon display, text display, flashing of specific light patterns, or vibration patterns from a vibration unit, without providing any voice notification. This allows for a wide range of use cases, such as accommodating the hearing impaired, showing consideration for others, or ensuring a reliable communication route other than voice in emergencies, further improving the versatility and practicality of the system. The elements of each of the above embodiments can be combined as appropriate without departing from the spirit of this disclosure. The technical concept of this disclosure can also be developed, for example, in conjunction with blockchain technology, to create a system that ensures the transparency and reliability of notification history and AI learning data.By recording the validity of notifications and the basis for learning on the blockchain, the system's reliability can be improved, and it can demonstrate an advantage from a privacy protection standpoint. Furthermore, by strengthening integration with IoT platforms, it is possible to collect richer contextual information from various sensors, further enhancing the accuracy and diversity of notifications.

[0045] [Examples]

[0046] This embodiment describes a specific scenario in which a parent sends a message to their child, and a child monitoring device adapts the message to the context and notifies the child. The specific technical problem that this embodiment aims to solve is "the decline in the quality of parent-child communication and the psychological burden on the child, caused by the parent's emotions being conveyed without adequate consideration of the child's situation."

[0047] (Example 1: Sending a "worry" message during a school lesson - including variations in AI learning mode)

[0048] Figure 9 is a flowchart of the message notification process according to this embodiment. The parent sends a text message from the parent terminal 10 saying, "Come straight home after school today. I'm a little worried." This process corresponds to step S101. The emotion estimation unit 202 of the server 20 estimates from this message that it strongly contains the emotion of "worry." This process corresponds to step S102. At the same time, the context information acquisition unit 203 of the server 20 identifies that the child is "on school" based on GPS location information acquired from the child monitoring device 30 and confirms that the current time is "during class." It also compares this with the schedule information and confirms that there is no schedule deviation. This process corresponds to step S103. The notification determination unit 204 determines the notification format based on the estimated emotion of "worry," the contextual information that the situation is "on school and during class," and the notification logic settings selected by the parent. This process corresponds to steps S104 and S105.

[0049] If the parent has selected the "Fixed Rules" mode (Yes branch in S104, S105A in Figure 9), the notification rule data 400 is referenced. The notification rule data 400 has settings registered to "minimize or mute" the volume and "prioritize" non-audio notifications when "on school and during class". In addition, a specific "icon display (e.g., the blue flashing cloud mark shown in Figure 6A)" and a "specific weak vibration pattern" are set for the emotion of "worry". Based on this, the notification determination unit 204 determines a notification style in which the sound is turned off, an icon indicating "worry" is displayed on the display unit 304, and a weak vibration is generated on the vibration unit 305.

[0050] On the other hand, if the parent selects the "AI-driven automatic learning" mode (No branch in S104 and S105B in Figure 9), the AI ​​model in the notification determination unit 204 uses past notification history, feedback from the parent, and information on the child's reaction as learning data to derive the optimal notification format from the estimated emotion of "worry" and the contextual information of "being at school and during class." For example, if past learning shows that when an audio notification was sent during class, the child did not check the message and the feedback from the parent was "inappropriate," the AI ​​adjusts its learning to suppress audio notifications during class and recommend non-audio notifications (icon display or vibration), and then determines the notification format based on the results.

[0051] The transmission unit 205 of the server 20 sends the message to the child monitoring device 30 along with the determined notification format. This process corresponds to step S106. The child monitoring device 30 notifies the child of the incoming message using the determined notification format (for example, displaying an icon on the display unit 304 and generating a weak vibration, or using an optimal non-verbal expression generated by AI) without interrupting the lesson. This process corresponds to step S107.

[0052] (Example 2: Sending an "Emergency" message when a child enters a dangerous area - including variations of AI learning mode)

[0053] Figure 10 is a flowchart of the message notification process according to this embodiment. The parent sends a voice message from the parent terminal 10 saying, "Get away from that place immediately! It's dangerous!" This process corresponds to step S201. The emotion estimation unit 202 of the server 20 estimates that the voice message contains a very strong emotion of "urgency." This process corresponds to step S202. At the same time, the context information acquisition unit 203 of the server 20 detects that the child has "entered" a pre-set "danger zone" (geofence) based on GPS location information acquired from the child monitoring device 30. This process corresponds to step S203.

[0054] The notification determination unit 204 determines the notification format based on the estimated "urgent" emotion, the contextual information of "intrusion into a dangerous area," and the notification logic settings selected by the guardian. This process corresponds to steps S204 and S205.

[0055] If the parent has selected the "fixed rule" mode (No branch in S204, S205B in Figure 10), the notification determination unit 204 determines that the emotion is "urgent" and therefore prioritizes a specific notification style compared to other emotions. Specifically, it determines voice parameters that set the volume to "maximum," the speaking speed to "faster," and the tone of voice to "harsh," as well as non-voice parameters that display a "flashing red warning icon" (Figure 6B) on the display unit 304 and generate a "strong, continuous vibration pattern" on the vibration unit 305.

[0056] On the other hand, if the parent selects the "AI-driven automatic learning" mode (Yes branch in S204 and S205A in Figure 10), the AI ​​model in the notification determination unit 204 uses past notification history, feedback from the parent, and information on the child's reaction as learning data to derive the optimal notification format from the estimated emotion of "urgency" and contextual information of "entry into a dangerous area." For example, if there is a history of audio notifications being ignored and vibration notifications being effective when entering a dangerous area in the past, the AI ​​will determine a notification format that further increases the intensity of vibration notifications in that situation, or enhances visual notifications.

[0057] The child monitoring device 30, based on its predetermined notification format, plays a message at a loud volume, flashes a red warning icon, and generates strong, continuous vibrations to immediately convey the importance of the message to the child and encourage them to avoid dangerous situations. This process corresponds to steps S206 and S207.

[0058] (Example 3: Utilization of AI-based automatic learning mode)

[0059] Figure 11 is a flowchart for determining the notification format based on the AI ​​learning model according to this embodiment. When a parent selects the "AI-based automatic learning" mode as the logic for the notification judgment unit 204, the system performs the following learning. This process corresponds to step S304. The server 20 continuously collects past notification history (emotion, context, notification format), feedback from parents (e.g., evaluations such as "the notification was appropriate" or "it was inappropriate"), and information on the child's response (e.g., message opening time, whether or not a reply was sent, changes in behavior in response to the notification). As shown in Figure 7, the AI ​​learning model uses this learning data as input to learn which notification format was most effective for a particular combination of emotion and situation. For example, if a child receives a "joy" message while playing in the park, and there is a history of the child immediately checking the voice notification and replying, the AI ​​will learn to prioritize voice notifications in that situation. Conversely, if a child does not check the message when a voice notification is sent during class, and the feedback from the parent is "inappropriate," the AI ​​will adjust its learning to suppress voice notifications during class and recommend non-voice notifications. Based on these learning results, the AI ​​dynamically determines a new notification format, providing more personalized and optimal notifications that adapt to the child's growth, changes in their living environment, and changes in the parent's communication style. This process corresponds to step S305B. As a result, the system will be able to support parent-child communication with greater accuracy over time.

[0060] <Summary>

[0061] [General tasks] One of the purposes of this disclosure is to facilitate smoother and more effective communication by accurately conveying the intended emotions of the message sender to the recipient while providing optimal notifications that take into account the recipient's situation. Issues corresponding to [Appendix 1] One of the purposes of this disclosure is to appropriately convey the emotions contained in parents' messages to their children and to promote smooth and heartwarming communication between parents and children. [Note 1] An information processing device comprising: a receiving unit that receives a message transmitted from a terminal device; an emotion estimation unit that estimates the emotion contained in the message; a notification determination unit that determines a notification format from the emotion; and a transmitting unit that transmits the message to a GPS terminal in the determined notification format. According to the above information processing device, it becomes possible to appropriately convey the nuances of emotions that parents want to communicate to their children, thereby promoting smooth and heartwarming communication between parents and children. More specifically, by appropriately conveying the nuances of emotions that parents wish to communicate to their children, and by providing optimal notifications that meticulously consider the child's real-time situation, it becomes possible to promote smooth, heartwarming, and deeper understanding-based communication between parents and children. This solves the conventional problems of insufficient emotional communication and the disruption of children's psychological safety due to notifications that do not consider the time, place, and occasion, and dramatically improves the practicality and reliability of the monitoring function by providing optimal communication according to the situation, such as in emergencies. Furthermore, this system enhances the certainty of information transmission and reduces unnecessary psychological burden by providing a more personalized and situation-appropriate flexible means of communication for both message senders and receivers. In particular, through continuous learning by AI, it can automatically adapt to the changing parent-child relationship and living environment, as well as the growth and characteristics of individual children, and always maintain and develop the optimal notification style. This ensures that messages are "understood" and provides a truly valuable communication experience that encourages "behavioral change" and "security" in those who "understood". Furthermore, by combining multiple notification methods, we can ensure reliable message delivery and eliminate information disparities, even for recipients with impairments in specific sensory organs.

[0062] Issues corresponding to [Appendix 2] One of the purposes of this disclosure is to improve children's psychological safety by considering the child's time, place, and occasion (TPO) and providing appropriate notifications that are suitable for the situation. [Note 2] The information processing apparatus according to claim 1, wherein the notification determination unit determines a notification format from the emotion and the situation information based on the situation information obtained from the GPS terminal. This allows for notifications that take the child's situation into consideration, thereby improving the child's psychological safety.

[0063] Issues corresponding to [Appendix 3] One of the purposes of this disclosure is to accurately estimate emotions from both text and audio messages and use that information to provide optimal notifications. [Note 3] The information processing apparatus according to claim 1, wherein the emotion estimation unit estimates emotions using natural language processing technology when the message is in text format, and estimates emotions using speech emotion recognition technology when the message is in voice format. This allows for accurate capture of emotions from messages in diverse input formats, enabling the determination of more appropriate notification methods.

[0064] Issues corresponding to [Appendix 4] One of the purposes of this disclosure is to ensure that urgent messages are transmitted reliably and quickly, and to improve the practicality and reliability of the monitoring function. [Note 4] The information processing apparatus according to claim 1 or 2, wherein the emotion estimation unit has a function to determine whether or not the emotion contained in the message indicates urgency, and the notification determination unit, when it is determined that the emotion indicates urgency, preferentially determines a specific notification format compared to the case of other emotions. This ensures that messages are reliably transmitted in emergencies, improving the practicality and reliability of the monitoring function.

[0065] Issues corresponding to [Appendix 5] One of the purposes of this disclosure is to determine a detailed notification format that is tailored to the child's circumstances (location, place, and time of day), thereby balancing the protection of the child's privacy with consideration for those around them. [Note 5] The information processing apparatus according to claim 2, wherein the situation information includes at least one of the location information of the GPS terminal, the type of place identified from the location information, and the current time period, and the notification determination unit determines at least one of the voice parameters of the message, such as volume, speaking speed, tone of voice, and intonation, and at least one of the non-voice parameters of the icon display, vibration pattern, and playback count, based on the situation information. This allows for the setting of optimal voice and non-voice notification parameters based on the child's current situation, enabling more considerate message delivery.

[0066] Issues corresponding to [Appendix 6] One of the purposes of this disclosure is to convey messages without disturbing others in specific locations and times, thereby reducing the psychological burden on children. [Note 6] The information processing apparatus according to claim 5, wherein the notification determination unit determines that the type of place is a school or a sleeping place, or determines that the current time is during class or bedtime, and prioritizes determining notifications based on non-speech parameters. This ensures that messages are reliably delivered to children even in situations where voice notifications are inappropriate, such as during class or while sleeping, while also being considerate of those around them and ensuring the child's psychological safety.

[0067] Issues corresponding to [Appendix 7] One of the purposes of this disclosure is to take into account children's deviations from their schedules and to more effectively communicate messages that convey concerns to parents in particular. [Note 7] The information processing apparatus according to claim 2, wherein the status information includes the schedule deviation status of the GPS terminal, and the notification determination unit determines the notification format based on the emotion estimated by the emotion estimation unit and the schedule deviation status. This ensures that messages expressing concern from parents are more reliably conveyed in urgent situations, such as when a child deviates from their schedule, thereby increasing the effectiveness of the monitoring function.

[0068] Issues corresponding to [Appendix 8] One of the purposes of this disclosure is to accurately express the type of emotion and urgency even in non-audio notifications, and to convey the nuances of the message visually and tactilely. [Note 8] The information processing apparatus according to claim 5, wherein the notification based on the non-voice parameter changes at least one of the visual characteristics of the icon display and the type or intensity of the vibration pattern depending on the type of emotion or urgency contained in the message. This allows for the effective transmission of emotional and urgent messages through visual changes in icons and vibration patterns, especially in situations where voice communication is unavailable or where a more intuitive way of conveying emotions is desired.

[0069] Issues corresponding to [Appendix 9] One of the purposes of this disclosure is to enhance trust and peace of mind by providing parents with options to control the system's notification logic. [Note 9] The information processing apparatus according to claim 1, wherein the notification determination unit selects the logic for determining the notification format from between a logic based on fixed rules and a logic based on automatic learning by artificial intelligence (AI), according to the settings made by the guardian. This improves the reliability and acceptability of the system, as parents can choose between fixed rules if they want complete control over the system's behavior, or AI learning if they desire more flexible optimization.

[0070] Issues corresponding to [Appendix 10] One of the purposes of this disclosure is to continuously optimize notification formats by adapting to children's growth and environmental changes through AI-driven automated learning. [Note 10] If the notification determination unit selects a logic based on automatic learning by artificial intelligence (AI), the AI ​​uses at least one of past notification history, feedback from parents, and information on the child's reaction as learning data, and learns to derive the optimal notification format from the emotion and situation information, as described in claim 9. This allows the AI ​​to automatically optimize notification formats through learning, adapting to changes in children's growth, living environments, and parents' communication styles, thus providing optimal notifications in the long term. [Explanation of symbols]

[0071] 10: Parental control device 20: Server 30: Child monitoring devices 100: Context-Adaptive Message Speech Conversion System 201: Receiving Unit 202: Emotion estimation part 203: Context Information Acquisition Unit 204: Notification judgment section 205: Transmitter 301: GPS receiver 302: Communications Department 303: Audio output section 304: Display section 305: Vibration section 306: Storage section 307: Control Unit 400: Notification rule data 401: Emotion Category 402: Context Condition 403: Audio Parameters 404: Non-voice parameter S101: Message receiving step S102: Emotion estimation step S103: Step to obtain context information S104: Notification logic selection step S105: Notification format determination step S105A: Decision step based on fixed rules S105B: AI-driven learning and decision-making steps S106: Message sending step S107: Message notification step S201: Message receiving step S202: Emotion Estimation Step S203: Step to obtain context information S204: Notification logic selection step S205A: Decision step based on fixed rules S206: Message sending step S207: Message notification step S301: Message reception step S302: Emotion estimation step S303: Step to obtain context information S304: Notification logic selection step S305B: AI-driven learning and decision-making steps S306: Message sending step S307: Message notification step

Claims

1. An information processing device, A receiving unit that receives messages sent from a terminal device, An emotion estimation unit that estimates the emotions contained in the message, A notification determination unit that determines the notification format based on the emotion, It comprises a transmitting unit that sends a message to a GPS terminal in a predetermined notification format, Information processing device.

2. The notification determination unit determines the notification format based on the emotion and the situation information obtained from the GPS terminal. The information processing apparatus according to claim 1.

3. The emotion estimation unit, when the message is in text format, estimates the emotion using natural language processing techniques. If the aforementioned message is in audio format, the emotion is estimated using speech emotion recognition technology. The information processing apparatus according to claim 1.

4. The emotion estimation unit has a function to determine whether or not the emotion contained in the message indicates urgency. The notification determination unit, when it determines that the emotion indicates urgency, prioritizes a specific notification format compared to other emotions. The information processing apparatus according to claim 1 or 2.

5. The aforementioned situational information includes at least one of the location information of the GPS terminal, the type of location identified from the location information, and the current time period. The notification determination unit determines, based on the situation information, at least one of the audio parameters of the message, such as volume, speaking speed, tone of voice, and intonation, and at least one of the non-audio parameters of the message, such as icon display, vibration pattern, and number of plays. The information processing apparatus according to claim 2.

6. The notification determination unit, when it determines that the type of location is a school or a sleeping place, or when it determines that the current time is during class or bedtime, minimizes or silences the volume and prioritizes determining notifications based on the non-sound parameters. The information processing apparatus according to claim 5.

7. The aforementioned status information includes the schedule deviation status of the GPS terminal, The notification determination unit determines the notification format based on the emotion estimated by the emotion estimation unit and the schedule deviation status. The information processing apparatus according to claim 2.

8. The notification based on the aforementioned non-voice parameters changes at least one of the visual characteristics of the icon display and the type or intensity of the vibration pattern depending on the type of emotion or urgency contained in the message. The information processing apparatus according to claim 5.

9. The notification determination unit selects, according to the settings made by the parent, the logic for determining the notification format from between a logic based on fixed rules and a logic based on automatic learning by artificial intelligence (AI). The information processing apparatus according to claim 1.

10. If the notification determination unit selects a logic based on automatic learning by artificial intelligence (AI), the AI ​​uses at least one of the past notification history, feedback from parents, and information on the child's response as learning data, and learns to derive the optimal notification format from the emotion and situation information. The information processing apparatus according to claim 9.

11. The processor receives a message sent from the terminal device. The processor estimates the emotion contained in the message, The processor determines the notification format based on the emotion, The processor sends a message to the GPS terminal in the determined notification format. method.

12. The processor receives the message sent from the terminal device. The processor is instructed to estimate the emotions contained in the message. The processor is instructed to determine the notification format based on the emotion in question. The processor is instructed to send a message to the GPS terminal in a predetermined notification format. program.

13. The system comprises a terminal device, a GPS terminal device, and a server. The aforementioned server, A receiving unit that receives messages sent from a terminal device, An emotion estimation unit that estimates the emotions contained in the message, A notification determination unit that determines the notification format based on the emotion, It comprises a transmitting unit that sends a message to a GPS terminal in a predetermined notification format, system.

Images