Method for providing personalized conversation service for elderly user and system thereof

Abstract

A method for providing a conversation service tailored to an elderly user is disclosed. According to one aspect of the present invention, a method for providing a conversation service tailored to an elderly user is provided, comprising the steps of: a providing unit providing a conversation service including voice conversation and a visual interface to an elderly user through a terminal using conversational artificial intelligence; a collecting unit collecting user data during a conversation with the user; a determining unit determining a user's cognitive state based on the user data; and the providing unit changing a method of providing the conversation service according to the user's cognitive state.

Description

Technology Field

[0001] The present invention relates to a method and system for providing a conversation service tailored to elderly users. Background Technology

[0003] With the recent advancement of artificial intelligence technology, conversational AI capable of engaging in natural language-based dialogue with users is being utilized in various fields. Meanwhile, the elderly may experience emotional instability and social isolation due to aging, living alone, and reduced social activity, leading to an increasing demand for services to alleviate these issues. Accordingly, conversational AI-based services for the elderly (companionship services) can be proposed; however, since the cognitive state of the elderly varies by individual or situation, providing conversation in a uniform manner may result in reduced information comprehension or increased cognitive burden. Therefore, there is a need for technology capable of providing user-customized conversational services by adaptively changing the delivery method to reflect the user's state. Prior art literature

[0005] Republic of Korea Published Patent Application No. 10-2024-0078761 (Published June 4, 2024) The problem to be solved

[0006] The present invention provides a method or system for providing a conversational service tailored to elderly users that can reduce cognitive burden and improve the accuracy and comprehension of information transmission. means of solving the problem

[0008] According to one aspect of the present invention, a method for providing a conversation service tailored to an elderly user is provided, comprising the steps of: a providing unit providing a conversation service including voice conversation and a visual interface to an elderly user through a terminal using conversational artificial intelligence; a collecting unit collecting user data during a conversation with the user; a determining unit determining a user's cognitive state based on the user data; and a providing unit changing a method of providing the conversation service according to the user's cognitive state.

[0009] User data may include voice data obtained from the user's voice, gaze data obtained by tracking the user's gaze, and operation data obtained from the user's operation log.

[0010] The step of determining the user's perception state involves calculating a voice perception score based on response delay time and silence frequency measured in voice data, calculating an eye gaze perception score based on fixation time and gaze deviation frequency measured in eye gaze data, calculating an operation perception score based on selection reversal frequency measured in operation data, calculating a perception state score by weighting the voice perception score, the eye gaze perception score, and the operation perception score, and determining the user's perception state based on the perception state score.

[0011] The step of determining the user's cognitive state may determine the user's cognitive state as a normal state if the cognitive state score is greater than or equal to a first set score, determine the user's cognitive state as an attention state if the cognitive state score is less than the first set score but greater than or equal to a second set score, and determine the user's cognitive state as an overload state if the cognitive state score is less than the second set score.

[0012] The step of changing the method of providing conversational services may involve slowing down the response speed and replacing abstract words with concrete ones when the user's cognitive state is determined to be in an attention state, and providing conversational services primarily through a visual interface when the user's cognitive state is determined to be in an overload state.

[0013] According to another aspect of the present invention, a conversation service provision system tailored for elderly users is provided, comprising: a providing unit that provides a conversation service including voice conversation and a visual interface to an elderly user through a terminal using conversational artificial intelligence; a collecting unit that collects user data during a conversation with the user; and a determining unit that determines the user's cognitive state based on the user data, wherein the providing unit changes the method of providing the conversation service according to the user's cognitive state. Effects of the invention

[0015] According to the present invention, a method or system for providing a conversational service tailored to elderly users can be provided, which can reduce cognitive burden and improve the accuracy and comprehension of information transmission. Brief explanation of the drawing

[0017] FIG. 1 is a flowchart illustrating a method for providing a conversation service tailored to an elderly user according to an embodiment of the present invention. FIG. 2 is a schematic diagram showing a system for providing conversational services tailored to elderly users according to another embodiment of the present invention. Specific details for implementing the invention

[0018] The present invention is capable of various modifications and may have various embodiments; specific embodiments are illustrated in the drawings and described in detail in the detailed description. However, this is not intended to limit the present invention to specific embodiments, and it should be understood that it includes all modifications, equivalents, and substitutions that fall within the spirit and scope of the present invention. In describing the present invention, detailed descriptions of related prior art are omitted if it is determined that such detailed descriptions may obscure the essence of the present invention.

[0019] Terms such as "first," "second," etc., may be used to describe various components, but said components should not be limited by said terms. These terms are used solely for the purpose of distinguishing one component from another.

[0020] The terms used in this application are used merely to describe specific embodiments and are not intended to limit the invention. The singular expression includes the plural expression unless the context clearly indicates otherwise. In this application, terms such as "comprising" or "having" are intended to specify the presence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0021] Hereinafter, embodiments of the method and system (100) for providing a conversation service tailored to an elderly user according to the present invention will be described in detail with reference to the attached drawings. In describing with reference to the attached drawings, identical or corresponding components are given the same reference numerals, and redundant descriptions thereof will be omitted.

[0023] First, a method for providing a conversation service tailored to elderly users according to one embodiment of the present invention will be described.

[0024] According to the present embodiment, as illustrated in FIG. 1, a method for providing a conversation service tailored to an elderly user is presented, comprising the steps of: a providing unit (110) providing a conversation service including voice conversation and a visual interface to an elderly user through a terminal using conversational artificial intelligence (S110); a collecting unit (120) collecting user data during a conversation with the user (S120); a determining unit (130) determining the user's cognitive state based on the user data (S130); and the providing unit (110) changing the method of providing the conversation service according to the user's cognitive state (S140).

[0025] According to the present embodiment, while providing a conversational service using conversational artificial intelligence, the user's cognitive state is determined based on the user data of the elderly user, and the method of providing the conversational service is changed accordingly, thereby providing a customized conversational service to the elderly user, which can improve the accuracy and comprehension of information delivery.

[0026] Hereinafter, each step of the method for providing a conversation service tailored to an elderly user according to the present embodiment will be described with reference to FIGS. 1 and FIGS. 2.

[0028] The method according to the present embodiment can be performed by a computing device.

[0029] Step 110 allows the providing unit (110) to provide a conversational service including voice conversation and a visual interface to an elderly user through a terminal using conversational artificial intelligence.

[0030] By providing conversation services to elderly users, emotional stability can be provided and social isolation can be alleviated.

[0031] Conversational artificial intelligence can be natural language-based artificial intelligence such as ChatGPT, Gemini, Copilot, Claude, CLOVA X, etc., or variant models thereof.

[0032] The conversational service can be provided through a terminal, where the terminal may be a smartphone, tablet PC, conversational robot, etc.

[0033] The conversation service provided to the user may include not only auditory voice conversation but also visual interfaces such as selection buttons displayed on a touchscreen, but may be primarily voice-based.

[0034] Step 120 allows the collection unit (120) to collect user data during a conversation with the user.

[0035] User data can be data about the user's state, behavior, reactions, etc. that appear during a conversation.

[0036] Such user data can be collected using a camera, microphone, memory, etc. equipped in the terminal.

[0037] Step 130 allows the judgment unit (130) to determine the user's awareness state based on user data.

[0038] User cognitive state is a cognitive state of the user that may comprehensively represent information processing ability, attention span, comprehension, decision-making ability, etc.

[0039] Step 140 allows the providing unit (110) to change the method of providing conversation services according to the user's awareness state.

[0040] By changing the method of dynamically providing conversational services based on the user's cognitive state, the user's cognitive load can be reduced and the accuracy and comprehension of information delivery can be improved.

[0041] User data may include voice data obtained from the user's voice, gaze data obtained by tracking the user's gaze, and operation data obtained from the user's operation log.

[0042] Voice data can be collected by acquiring the user's voice through a microphone equipped in the terminal.

[0043] Eye gaze data can be collected by tracking the user's gaze through a camera equipped in the terminal.

[0044] Operation data can be collected by obtaining operation logs generated by user operation through a touchscreen provided on the terminal and stored in a storage medium such as memory provided on the terminal.

[0045] The step of determining the user's cognitive state (S130) can determine the user's cognitive state based on the cognitive state score calculated using user data.

[0046] Here, the cognitive state score can be calculated by weighting the voice recognition score calculated from voice data, the gaze recognition score calculated from gaze data, and the operation recognition score calculated from operation data.

[0047] At this time, the weights applied to each cognitive score can be set based on the importance of each cognitive score calculated according to the results of the artificial intelligence model learning the correlation between each cognitive score and the user's cognitive state.

[0048] Here, the artificial intelligence model may be a linear regression model, a logistic regression model, a decision tree model, a Random Forest model, an XGBoost model, or a combination thereof.

[0049] The voice recognition score can be calculated based on the response delay time and the frequency of silence occurrences measured in the voice data.

[0050] Response latency may be a feature value that measures the time elapsed from the moment a question is asked or information is provided to the user in voice data until the user begins to respond verbally.

[0051] The frequency of silence may be a feature value that measures the number of times a period during a user's utterance in voice data occurs in which no voice is detected for a certain period of time or longer.

[0052] The voice recognition score can be calculated higher as the response delay time is shorter and the frequency of silence occurrences is lower.

[0053] The gaze recognition score can be calculated based on the gaze fixation time and gaze deviation frequency measured from the gaze data.

[0054] Gaze fixation time can be a feature value in gaze data that measures the time a user's gaze continuously stays on a specific location or object.

[0055] The frequency of gaze deviation can be a feature value that measures the number of times the focus is disrupted in gaze data as the user's gaze fails to remain on the object being looked at and moves to another location.

[0056] The gaze recognition score can be calculated higher the longer the fixed gaze time and the lower the frequency of gaze deviation.

[0057] The manipulation recognition score can be calculated based on the frequency of choice reversal measured in the manipulation data.

[0058] The frequency of selection reversal may be a feature value measuring the number of times hesitation occurs in the manipulation data, such as canceling a user-selected item or changing it to another item.

[0059] The manipulation recognition score can be calculated higher as the frequency of choice reversal decreases.

[0060] Speech recognition scores, gaze recognition scores, and manipulation recognition scores can each be calculated using an artificial intelligence model, wherein the artificial intelligence model can receive feature values ​​corresponding to each recognition score as input to calculate the corresponding recognition score, and specifically, it may be a linear regression model, a decision tree model, a Random Forest model, an XGBoost model, an artificial neural network model, or a combination thereof.

[0061] The step of determining the user's cognitive state (S130) can determine the user's cognitive state as a normal state, that is, a state in which the user can normally perceive and understand the provided information and perform an appropriate response or decision-making, if the cognitive state score is greater than or equal to the first set score.

[0062] In addition, if the cognitive state score is less than the first set score but greater than or equal to the second set score (< first set score), the user's cognitive state can be determined as an attention state, that is, a state in which the user shows some difficulty in perceiving and understanding the provided information but can perform a response or decision if appropriate support is provided.

[0063] In addition, if the cognitive state score is less than the second set score, the user's cognitive state may be determined to be an overloaded state, that is, a state in which the user shows significant difficulty in perceiving or understanding the provided information or in making appropriate responses or decisions.

[0064] The first setting score and the second setting score may be set based on expert evaluation results, statistical analysis results, or the learning results of an artificial intelligence model, wherein the artificial intelligence model may specifically be a linear regression model, a logistic regression model, a decision tree model, a Random Forest model, an XGBoost model, or a combination thereof.

[0065] The step of changing the conversation service provision method (S140) can slow down the response speed and replace the words used from abstract to concrete when the user's cognitive state is determined to be an attention state.

[0066] For example, the sentence "Please proceed with the relevant procedure." can be changed to a more specific expression, such as "Please press the 'Apply' button at the bottom of the screen."

[0067] Additionally, the step of changing the method of providing the conversation service (S140) may provide the conversation service primarily through a visual interface when the user's cognitive state is determined to be overloaded.

[0068] By changing the method of providing conversational services to match the user's cognitive state in this way, the user's cognitive load can be reduced, and the accuracy and comprehension of information delivery can be improved.

[0069] Specifically, the step of changing the conversation service provision method (S140) can change the syntactic structure of the response to be mainly short sentences when the user's cognitive state is determined to be an overloaded state.

[0070] In addition, the response to the question can be requested as a yes or no choice.

[0071] In addition, background music and sound effects can be removed, and only voice in a specific frequency band with high clarity can be output.

[0072] For example, the sentence "Would you please confirm whether you took your morning and lunch medications for today's medication management?" can be changed to a short question "Did you take your medication?", and the user can respond by selecting "Yes / No" on the touchscreen.

[0073] After the step of changing the method of providing the conversation service (S140), the step of sending an emergency notification (S150) may be further included.

[0074] The step of transmitting an emergency notification (S150) is that if the user's response time after the conversation service provision method is changed is longer than the set time, the judgment unit (130) determines that it is an emergency situation and transmits an emergency notification to the user's guardian.

[0075] In addition, if the user's response time is less than the set time after the conversation service provision method is changed, it is determined that the user's cognitive state has been recovered, and the conversation service provision method can be restored to the original method.

[0076] Here, the setting time can be dynamically set per user based on the average response time calculated from the user's past response history.

[0077] Meanwhile, after the step of determining the user's cognitive state (S130), a step of recommending welfare services (S160) may be further included.

[0078] In the step of recommending welfare services (S160), the recommendation unit (160) can recommend welfare services for the user based on the user's cognitive state score, and the guardian's asset information and dependent family information.

[0079] The guardian's asset information and dependent family information can be entered in advance by requesting them via the guardian's device.

[0080] Specifically, the step of recommending welfare services (S160) may include a step of calculating a self-determination index based on the user's cognitive state score (S161), a step of calculating a support capacity index based on the guardian's asset information and dependent family information (S162), and a step of providing a list of recommended welfare services according to a recommendation index calculated by weighting the self-determination index and the support capacity index (S163).

[0081] The self-determination index is a characteristic value that quantifies a user's cognitive ability to select welfare services independently, and it can be calculated as higher as the cognitive state score increases.

[0082] The support capacity index is a characteristic value that quantifies a caregiver's economic ability to accept the costs of welfare services; it can be calculated to be higher when the caregiver has more assets and fewer dependents.

[0083] The self-determination index and the support capacity index can each be calculated using an artificial intelligence model, wherein the artificial intelligence model can calculate the corresponding index by receiving feature values ​​corresponding to each index as input, and specifically, it may be a linear regression model, a decision tree model, a Random Forest model, an XGBoost model, an artificial neural network model, or a combination thereof.

[0084] The step of providing a list of welfare services (S163) may calculate a recommendation index by weighting the self-determination index and the support capacity index, and provide a list of recommended welfare services to the user and / or guardian according to the recommendation index.

[0085] At this time, the weights applied to each index may be set based on the importance of each index calculated according to the results of the artificial intelligence model learning the correlation between each index and the welfare service recommendation results (whether the recommended welfare service is used, whether it is selected, satisfaction, suitability, etc.), and the artificial intelligence model may be a linear regression model, a logistic regression model, a decision tree model, a Random Forest model, an XGBoost model, or a combination thereof.

[0086] A set index is pre-labeled for each welfare service, and welfare services with a set index greater than or equal to the recommendation index may be included in the recommended welfare service list. In this case, the set index may be set based on expert evaluation results, statistical analysis results, or the learning results of an artificial intelligence model. Specifically, the artificial intelligence model may be a linear regression model, a logistic regression model, a decision tree model, a Random Forest model, an XGBoost model, or a combination thereof.

[0087] The step of providing a list of recommended welfare services (S163) increases the weight assigned to the self-determination index when the self-determination index is greater than or equal to the first set value, and recommends welfare services according to the recommendation index calculated accordingly, while considering the user's preferences, it may recommend leisure or cultural services relatively first.

[0088] In addition, if the self-determination index is less than the second set value, the weight assigned to the support capacity index is increased, and welfare services are recommended according to the recommendation index calculated accordingly, but medical or care services may be recommended relatively preferentially in consideration of the user's health.

[0089] In addition, if the support capacity index is above the third setting value, welfare services requiring relatively high costs can be included in the recommended welfare service list.

[0090] Meanwhile, if an emergency notification is transmitted, the weight assigned to the self-determination index is reduced to 0, and welfare services are recommended according to the recommendation index calculated accordingly, and in this case, medical or care services optimized for emergency situations may be recommended first.

[0091] The first, second, and third settings may be set based on expert evaluation results, statistical analysis results, or the learning results of an artificial intelligence model, wherein the artificial intelligence model may specifically be a linear regression model, a logistic regression model, a decision tree model, a Random Forest model, an XGBoost model, or a combination thereof.

[0093] Next, a system (100) providing a conversation service tailored to an elderly user according to one embodiment of the present invention will be described.

[0094] According to the present embodiment, as illustrated in FIG. 2, a conversation service provision system (100) customized for elderly users is presented, comprising a providing unit (110) that provides a conversation service including voice conversation and a visual interface to an elderly user through a terminal using conversational artificial intelligence, a collecting unit (120) that collects user data during a conversation with the user, and a determining unit (130) that determines the user's cognitive state based on the user data, wherein the providing unit (110) changes the method of providing the conversation service according to the user's cognitive state.

[0095] According to the present embodiment, while providing a conversational service using conversational artificial intelligence, the user's cognitive state is determined based on the user data of the elderly user, and the method of providing the conversational service is changed accordingly, thereby providing a customized conversational service to the elderly user, which can improve the accuracy and comprehension of information delivery.

[0096] With reference to FIGS. 1 and FIGS. 2, each component of the elderly user customized conversation service provision system (100) according to the present embodiment will be described below, and the specific details thereof may follow the content of the elderly user customized conversation service provision method described above.

[0098] The providing unit (110) can provide a conversational service including voice conversation and a visual interface to an elderly user through a terminal using conversational artificial intelligence.

[0099] The collection unit (120) can collect user data during a conversation with the user.

[0100] The judgment unit (130) can determine the user's awareness status based on user data.

[0101] The providing unit (110) can change the method of providing conversation services according to the user's awareness state.

[0102] User data may include voice data obtained from the user's voice, gaze data obtained by tracking the user's gaze, and operation data obtained from the user's operation log.

[0103] The judgment unit (130) can calculate a voice recognition score based on the response delay time and silence occurrence frequency measured in voice data, calculate an eye gaze recognition score based on the gaze fixation time and gaze departure frequency measured in eye gaze data, calculate an operation recognition score based on the selection reversal occurrence frequency measured in operation data, calculate a recognition state score by weighting the voice recognition score, the eye gaze recognition score, and the operation recognition score, and determine the user recognition state according to the recognition state score.

[0104] The judgment unit (130) can determine the user's perception state to be normal if the perception state score is greater than or equal to the first set score, determine the user's perception state to be at attention if the perception state score is less than the first set score but greater than or equal to the second set score, and determine the user's perception state to be overloaded if the perception state score is less than the second set score.

[0105] The providing unit (110) can slow down the response speed and replace the words used from abstract to concrete when the user's cognitive state is determined to be in an attention state, and can provide the conversation service mainly through a visual interface when the user's cognitive state is determined to be in an overload state.

[0106] When the user's perception state is determined to be overloaded, the providing unit (110) can change the syntactic structure of the answer to be mainly short sentences, request a response to the question in a yes or no selection manner, remove background sounds and sound effects, and output only voice in a specific frequency band with high clarity.

[0107] The judgment unit (130) can determine that if the user's response time is less than the set time after the conversation service provision method is changed, the user's cognitive state has been restored and the conversation service provision method is restored to the original method, and if the user's response time is greater than or equal to the set time after the conversation service provision method is changed, it can determine that it is an emergency situation and send an emergency notification to the user's guardian.

[0108] A recommendation unit (160) that recommends welfare services for the user based on the cognitive state score, and the guardian's asset information and dependent family information received from the user's guardian may be further included.

[0109] The recommendation unit (160) can calculate a self-determination index that indexes the cognitive ability of the user to choose welfare services based on a cognitive state score, calculate a support capacity index that indexes the economic ability of the guardian to accept the cost of welfare services based on the guardian's asset information and dependent family information received in advance from the user's guardian, and provide a list of recommended welfare services according to the recommendation index calculated by weighting the self-determination index and the support capacity index.

[0110] The recommendation unit (160) can increase the weight assigned to the self-determination index when the self-determination index is greater than or equal to the first set value, and recommend leisure or cultural services first; when the self-determination index is less than the second set value, it can increase the weight assigned to the support capacity index, and recommend medical or care services first; when the support capacity index is greater than or equal to the third set value, it can include welfare services that incur relatively high costs in the recommended welfare service list; and when there is a transmission of an emergency notification, it can decrease the weight assigned to the self-determination index to 0, and recommend medical or care services optimized for emergency situations first.

[0112] Meanwhile, the components of the aforementioned embodiments can be easily identified from a process perspective. That is, each component can be identified as a separate process. Furthermore, the processes of the aforementioned embodiments can be easily identified from the perspective of the device components.

[0113] In addition, the technical details described above may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either individually or in combination. The program instructions recorded on the medium may be those specifically designed and configured for the embodiments, or they may be those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. Hardware devices may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

[0115] Although an embodiment of the present invention has been described above, those skilled in the art may modify and change the present invention in various ways by adding, changing, deleting, or adding components, etc., without departing from the spirit of the present invention as described in the claims, and such modifications and changes are also to be included within the scope of the rights of the present invention. Explanation of the symbols

[0117] 100: System providing customized conversation services for elderly users 110: Provider 120: Collection Department 130: Judgment Department 160: Recommendation Section

Claims

Claim 1 A providing unit provides a conversational service including voice conversation and a visual interface to an elderly user through a terminal using conversational artificial intelligence; a collecting unit collects user data during a conversation with the user; a determining unit determines the user's cognitive state based on the user data; The providing unit includes a step of changing the conversation service provision method according to the user perception state, wherein the user perception state is a state indicating the user's understanding of information, attention concentration, and decision burden, and the user data includes voice data obtained from the user's voice, gaze data obtained by tracking the user's gaze, and operation data obtained from the user's operation log, and the step of determining the user perception state calculates a voice perception score based on the response delay time and silence frequency measured in the voice data, calculates a gaze perception score based on the gaze fixation time and gaze deviation frequency measured in the gaze data, calculates an operation perception score based on the selection reversal frequency measured in the operation data, calculates a perception state score by weighting the voice perception score, the gaze perception score, and the operation perception score, and determines the user perception state according to the perception state score, wherein the selection reversal frequency is the number of times an act of canceling an item selected by the user or changing it to another item occurs, and the step of determining the user perception state determines the user perception state as a normal state when the perception state score is greater than or equal to a first set score, and the perception state If the score is less than the first set score but greater than or equal to the second set score, the user perception state is determined to be an attention state; if the perception state score is less than the second set score, the user perception state is determined to be an overload state; and the step of changing the conversation service provision method is, when the user perception state is determined to be an attention state,A method for providing a conversation service tailored to an elderly user, comprising: slowing down the response speed and replacing words used from abstract to concrete; and, when the user's cognitive state is determined to be the overloaded state, providing the conversation service primarily through the visual interface, wherein providing the conversation service primarily through the visual interface includes changing the syntactic structure of the response to be mainly short sentences and requesting a response to a question in a yes or no selection manner; and, after the step of changing the conversation service provision method, the judgment unit further includes the step of determining that the user's cognitive state has recovered and returning the conversation service provision method to the original method if the user's response time is less than a set time after the conversation service provision method is changed, and determining that it is an emergency situation and sending an emergency notification to the user's guardian if the response time is greater than or equal to the set time, wherein the set time is dynamically set for each user based on the average response time calculated from the user's past response history. Claim 2 delete Claim 3 delete Claim 4 delete Claim 5 delete

Images