system

Abstract

We provide the system. [Solution] A speech recognition means that collects voice data from elderly people and converts it into text data, A generative model means that analyzes the converted text data to generate a response, A speech synthesis means that converts the generated response into speech and outputs it, A health information extraction method that extracts health information from the daily conversations of elderly people, A profile update means for updating the user profile with extracted health information, A notification method that sends a notification to a designated contact when an anomaly is detected, A system that includes this.

Description

Technical Field

[0001] The technology of the present disclosure relates to a system.

Background Art

[0002] Patent Document 1 discloses a persona chatbot control method performed by at least one processor, the method including steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] An object of the present invention is to solve problems such as loneliness and decline in cognitive function experienced by elderly people living alone, and further, the difficulty for family members and caregivers to grasp the physical condition and emergencies of elderly people. Also, it aims to provide a safer and higher-quality living environment by easily extracting health information from the daily conversations of elderly people and appropriately managing and notifying it.

Means for Solving the Problems

[0005] The present invention solves the above problems with a system that includes the following means: a speech recognition means, a speech synthesis means, a generation model means, a health information extraction means, a profile update means, and a notification means. Specifically, the system includes a speech recognition means that collects voice data of an elderly person and converts it into text data, a generation model means that analyzes the converted text data and generates a response, a speech synthesis means that converts the generated response into speech and outputs it, a health information extraction means that extracts health information from the elderly person's daily conversation, a profile update means that updates the extracted health information in the user profile, and a notification means that sends a notification to a set contact when an abnormality is detected. With this system, it is possible to understand the health status of an elderly person while having a natural conversation and to take a quick response as needed.

[0006] "Voice recognition means" refers to a device or system that collects speech from elderly people and converts that speech data into text data.

[0007] "Speech synthesis means" refers to a device or system that generates speech data from generated text data and outputs the speech through a speaker or the like.

[0008] A "generative model means" is an AI-based model or algorithm that analyzes collected text data and generates an appropriate response.

[0009] A "health information extraction method" is a system that automatically extracts health-related information (e.g., symptoms and medication status) from the conversations of elderly people.

[0010] A "profile update method" is a system that reflects extracted health information in the profile of individual elderly people and records their latest health status and history.

[0011] A "notification method" is a system that sends notifications via email, messaging, or other means to designated contacts (e.g., family members or caregivers) when an abnormality is detected. [Brief explanation of the drawing]

[0012] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] This is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] This is a sequence diagram showing the processing flow of the data processing system in Example 2, which incorporates an emotion engine. [Figure 14] This is a sequence diagram showing the processing flow of the data processing system in Application Example 2, which combines an emotion engine. [Modes for carrying out the invention]

[0013] Hereinafter, an example of an embodiment of the system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0014] First, the terms used in the following description will be explained.

[0015] In the following embodiments, the numbered processor (hereinafter simply referred to as "processor") may be one arithmetic unit or a combination of a plurality of arithmetic units. Also, the processor may be one type of arithmetic unit or a combination of a plurality of types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), and the like.

[0016] In the following embodiments, the numbered RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0017] In the following embodiments, the numbered storage is one or more non-volatile storage devices that store various programs and various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes, and the like.

[0018] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0019] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0020] [First Embodiment]

[0021] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0022] As shown in Figure 1, the data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0023] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0024] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0025] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0026] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0027] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0028] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0029] As shown in Figure 2, in the data processing device 12, a specific processing is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" related to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0030] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0031] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0032] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0033] This invention provides a system that enables the understanding of the health status of elderly people living alone through natural conversation and allows for prompt action as needed. This system consists of speech recognition means, speech synthesis means, generative model means, health information extraction means, profile update means, and notification means. Details of each component and specific examples of systems combining them are described below.

[0034] Main processing flow

[0035] 1. Collection of audio data (device):

[0036] A voice recognition device, such as a smart speaker, is constantly on standby, waiting for the elderly person (user) to begin speaking. When the user says, "I have a headache today, so I took some medicine," the voice data is collected.

[0037] 2. Speech-to-text conversion (device):

[0038] The collected audio data is sent to a speech recognition system and converted into text data. For example, the text "I had a headache today, so I took some medicine" is generated.

[0039] 3. Response generation (server):

[0040] The converted text data is sent to the server, where a generative model analyzes it to generate an appropriate response. An example of a generated response is, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0041] 4. Audio output (device):

[0042] The generated response text is converted into speech data by a speech synthesis system, and the voice is output from the smart speaker. The user hears the response, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0043] 5. Extraction of health information (server):

[0044] The server extracts health-related information from the user's statements using a health information extraction method. In this example, the health information extracted includes "headache" and "took medicine."

[0045] 6. Profile update (server):

[0046] The extracted health information is updated in each user's profile. The server adds information to the profile such as "Had a headache on 2023-10-01 and took medication."

[0047] 7. Notifications and alerts (server):

[0048] If the server detects an anomaly in the profile's health information, it sends a notification to a pre-configured contact. For example, a message such as "Mr. / Ms. A complained of a headache and has taken medication" might be sent to a family member.

[0049] Specific system examples

[0050] Example: In the case of "Mr. A"

[0051] 1. Collection of audio data:

[0052] User A says, "I had a headache today, so I took some medicine."

[0053] Device: A smart speaker collects this statement.

[0054] 2. Text conversion:

[0055] Terminal: Converts the collected audio data into text data: "I had a headache today, so I took some medicine."

[0056] 3. Response generation:

[0057] Server: The generative model analyzes the text data and generates the response, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0058] 4. Audio output:

[0059] Terminal: The speech synthesis device converts the generated response into speech and outputs the following message from the speaker: "You have a headache, take care of yourself. Don't forget to take your medicine."

[0060] 5. Extraction of health information:

[0061] Server: The health information extraction method extracts information such as "headache" and "took medicine" from the text data.

[0062] 6. Profile Update:

[0063] Server: Records the extracted health information in Person A's profile and updates the data to "Had a headache on 2023-10-01 and took medication."

[0064] 7. Notifications and alerts:

[0065] Server: Detects an anomaly in the health profile (in this case, "headache") and sends a notification to the designated contact (A's family) stating, "A is complaining of a headache and has taken medication."

[0066] This specific workflow allows for real-time monitoring of the health status of elderly individuals and prompt provision of necessary information to family members and care staff. This system can improve the quality of life for elderly people living alone.

[0067] The following describes the processing flow.

[0068] Step 1:

[0069] The user speaks. For example, they might say, "I had a headache today, so I took some medicine."

[0070] Step 2:

[0071] The device collects the user's speech as audio data. This is done by smart speakers and voice recognition devices.

[0072] Step 3:

[0073] The device calls a speech recognition engine to convert the collected audio data into text data. For example, it might convert the audio data into the text "I had a headache today, so I took some medicine."

[0074] Step 4:

[0075] The terminal sends the converted text data to the server. This initiates the analysis of the text data.

[0076] Step 5:

[0077] The server uses a generative model to analyze the received text data and generate an appropriate response. For example, a response such as "You have a headache, take care of yourself. Don't forget to take your medicine." might be generated.

[0078] Step 6:

[0079] The server sends the generated response text to the terminal.

[0080] Step 7:

[0081] The terminal sends the received response text to a speech synthesis engine, which converts it into speech data. The audio is then output through the speaker. Specifically, the following message is played: "You have a headache, please take care of yourself. Don't forget to take your medicine."

[0082] Step 8:

[0083] The server extracts health information from text data. For example, it might extract health information such as "headache" and "took medication."

[0084] Step 9:

[0085] The server updates the user's profile with the extracted health information. For example, the information "Had a headache on 2023-10-01 and took medication" will be recorded in person A's profile.

[0086] Step 10:

[0087] If the server detects an anomaly, it will send a notification to the designated contact. For example, a notification stating, "Mr. / Ms. A complained of a headache and has taken medication," will be sent to the user's family.

[0088] This processing flow allows for real-time monitoring of the health status of elderly individuals and enables rapid response. The sequential execution of each step ensures the safety and security of the user.

[0089] (Example 1)

[0090] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0091] The aim is to solve the problem of difficulty in monitoring and responding promptly to the daily health conditions of elderly people living alone. In particular, there is a need to provide appropriate responses in real time when elderly people report any abnormalities.

[0092] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0093] In this invention, the server includes a generation model means, a health information extraction means, a profile update means, and a notification means. This makes it possible to analyze the content of elderly people's statements, extract and record health-related information, and send notifications quickly when an abnormality is detected.

[0094] "Speech recognition means" refers to a device and program that collects user speech as audio data and converts it into text data.

[0095] "Speech synthesis means" refers to a device and program for converting text data into speech data and playing it back.

[0096] A "generative model means" is a program that analyzes collected text data and generates an appropriate response.

[0097] "Health information extraction means" refers to a device and program that extracts health-related information from text data.

[0098] "Profile update means" refers to a device and program that records and updates extracted health information in each user's profile.

[0099] "Notification means" refers to a device and program that sends a notification to a pre-configured contact when an anomaly is detected.

[0100] This invention provides a system that enables the understanding of the health status of elderly people living alone through natural conversation and enables prompt action as needed. This system consists of a speech recognition means, a speech synthesis means, a generation model means, a health information extraction means, a profile update means, and a notification means.

[0101] Hardware and software to be used

[0102] 1. Speech recognition means

[0103] Hardware: Smart speaker

[0104] Software: Speech recognition API (e.g., Google Speech-to-Text API)

[0105] 2. Speech synthesis means

[0106] Software: Text-to-speech API (e.g., Google Text-to-Speech API)

[0107] 3. Generative Model Means

[0108] Software: Generative AI models (e.g., GPT-3)

[0109] 4. Health information extraction means

[0110] Software: Natural Language Processing Algorithms

[0111] 5. Profile update method

[0112] Software: Database management systems (e.g., MySQL)

[0113] 6. Means of notification

[0114] Software: Notification API (e.g., Twilio)

[0115] Specific example: In the case of "Mr. A"

[0116] Collection of audio data

[0117] User A says, "I had a headache today, so I took some medicine."

[0118] Device: A smart speaker collects this statement.

[0119] Speech-to-text conversion

[0120] Terminal: The collected audio data is converted into text data, "I had a headache today, so I took some medicine," using a speech recognition API.

[0121] Response generation

[0122] Server: The generative model analyzes the text data and generates the response, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0123] Audio output

[0124] Terminal: The response generated using a speech synthesis API is converted into speech and output from the speaker the following message: "You have a headache, take care of yourself. Don't forget to take your medicine."

[0125] Extraction of health information

[0126] Server: The health information extraction method extracts information such as "headache" and "took medicine" from the text data.

[0127] Profile Update

[0128] Server: Records the extracted health information in Person A's profile and updates the data to "Had a headache on 2023-10-01 and took medication."

[0129] Notifications and alerts

[0130] Server: Detects an anomaly in the health profile (in this case, "headache") and sends a notification to the configured contacts stating, "Person A complained of a headache and has taken medication."

[0131] Example of a prompt

[0132] An elderly person says, "I had a headache today, so I took some medicine."

[0133] The server uses a generated AI model to respond, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0134] This system allows for real-time monitoring of the health status of elderly people living alone and promptly provides necessary information to family members and care staff. This, in turn, can improve the quality of life for elderly people living alone.

[0135] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0136] Step 1:

[0137] Collection of audio data (device)

[0138] Input: The user speaks into the smart speaker.

[0139] Action: The user says, "I have a headache today, so I took some medicine." The smart speaker, acting as the device, uses its microphone to collect this statement.

[0140] Data processing: The microphone converts the audio signal into a digital signal and temporarily stores it in a buffer.

[0141] Output: Digital audio data.

[0142] Step 2:

[0143] Speech-to-text conversion (device)

[0144] Input: Digital audio data.

[0145] Operation: The device sends the collected audio data to a speech recognition API (e.g., Google Speech-to-Text API).

[0146] Data processing: The speech recognition API analyzes the audio data and identifies phonemes and context. This converts the audio data into text data: "I had a headache today, so I took some medicine."

[0147] Output: Text data "I had a headache today, so I took some medicine."

[0148] Step 3:

[0149] Response generation (server)

[0150] Input: Text data "I had a headache today, so I took some medicine."

[0151] Operation: The server inputs the received text data into a generating AI model (e.g., GPT-3) and generates an appropriate response.

[0152] Data processing: The generative AI model analyzes text data and generates a response based on the context. Specifically, it generates a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0153] Output: Response text "You have a headache, please take care of yourself. Don't forget to take your medicine."

[0154] Step 4:

[0155] Audio output (device)

[0156] Input: Response text: "You have a headache, please take care of yourself. Don't forget to take your medicine."

[0157] Operation: The device sends the response text to a text-to-speech API (e.g., Google Text-to-Speech API) and converts it into speech data.

[0158] Data processing: The speech synthesis API analyzes the text data and generates an audio waveform. This audio waveform is then prepared as audio data.

[0159] Output: Response voice data.

[0160] Step 5:

[0161] Extraction of health information (server)

[0162] Input: Text data "I had a headache today, so I took some medicine."

[0163] Operation: The server uses a health information extraction method to identify health-related information from text data.

[0164] Data processing: Analyze the text using natural language processing algorithms and extract health-related keywords such as "headache" and "took medicine."

[0165] Output: Health information "headache" and "took medication".

[0166] Step 6:

[0167] Profile update (server)

[0168] Input: Health information "headache" and "took medication".

[0169] Operation: The server records and updates the extracted health information in the user profile.

[0170] Data processing: Access the user profile database and add the information, "Had a headache on 2023-10-01 and took medication."

[0171] Output: Updated user profile.

[0172] Step 7:

[0173] Notifications and alerts (server)

[0174] Input: Updated user profile.

[0175] Operation: The server detects anomalies from the health profile and sends notifications to the configured contacts.

[0176] Data processing: Anomaly detection algorithms are used to extract anomalous data that matches the specified conditions. Specifically, a notification is generated stating, "Person A complained of a headache and took medication."

[0177] Output: A notification message sent to the contact.

[0178] This system's processing flow makes it possible to monitor the health status of elderly individuals in real time and quickly provide necessary information to their families and care staff.

[0179] (Application Example 1)

[0180] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0181] For elderly people living alone, monitoring their health and ensuring a rapid response in emergencies are major challenges. However, conventional systems have struggled to naturally monitor health conditions through voice conversations and to quickly detect and respond to abnormalities. In particular, the lack of adequate means for appropriate responses in emergencies and for promptly notifying family members and medical professionals has made ensuring the safety of the elderly a significant concern.

[0182] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0183] In this invention, the server includes speech recognition means, speech synthesis means, generation model means, health information extraction means, profile update means, notification means, speech response generation means, and health information transmission means. This makes it possible to analyze the voice spoken by elderly people on a daily basis in real time and constantly monitor their health status. Furthermore, when an abnormality is detected, it is possible to quickly generate an appropriate response, enabling emergency response and immediate notification to family members and medical professionals through remote transmission of health information.

[0184] "Speech recognition means" refers to a technological element that collects speech from elderly people as audio data and converts that audio data into text data.

[0185] "Speech synthesis means" refers to a technological element that converts generated text responses into speech data and transmits it to elderly people.

[0186] A "generative model means" is a technical element that analyzes text data of everyday conversations of elderly people and generates appropriate responses.

[0187] A "health information extraction method" is a technological element that extracts health-related information from everyday conversations.

[0188] A "profile update method" is a technological element that records and updates extracted health information in each user's profile.

[0189] A "notification method" is a technical element that sends a notification to a designated contact when an anomaly is detected.

[0190] A "voice response generation means" is a technological element that automatically generates an appropriate response in the event of an abnormality, thereby addressing the needs of the elderly.

[0191] A "health information transmission means" is a technical element that transmits extracted health information to a remote server.

[0192] The system of this invention provides support for elderly people to live safely and healthily in their homes or care facilities. This system is composed of the following hardware and software.

[0193] Hardware configuration

[0194] 1. Voice recognition device: Use a microphone built into the robot (e.g., USB microphone, built-in microphone of a smart speaker). This will pick up the speech of elderly people.

[0195] 2. Speech synthesis device: Uses the robot's built-in speaker to output the generated response as speech.

[0196] 3. Computing Unit: A server or high-performance computer that manages the entire system.

[0197] Software Configuration

[0198] 1. Speech recognition software: The speech_recognition library in Python is used to convert the speech of elderly people into text data.

[0199] 2. Speech synthesis software: Use the pyttsx3 library in Python to convert text responses into speech.

[0200] 3. Generative Models: AI models that use natural language processing (NLP) to analyze text data and generate appropriate responses. For example, OpenAI's GPT-3 is used.

[0201] 4. Communication method: Use the Python requests library to send health information to a remote server.

[0202] Explanation of the process

[0203] The server first receives audio data transmitted from the speech recognition device and converts it into text data. This process is performed using the speech_recognition library. Next, the generative model analyzes the text data and generates an appropriate response from what the elderly person is saying. The generated response is converted into audio data using the pyttsx3 library and output by the speech synthesis device.

[0204] Furthermore, health-related information is extracted from the generated text data using a health information extraction method. This extracted information is then updated in the user's profile by the server. If an abnormality is detected, a notification is sent to the contacts set by the notification method. In addition, by sending the extracted health information to a remote server using the health information transmission method, the health status of elderly individuals can be continuously monitored.

[0205] Explanation of specific examples

[0206] For example, if person A says, "I have a headache today, so I took some medicine," the system collects this utterance as audio data and converts it into text data. Next, a generative model analyzes this and generates a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine." This response is then spoken by a speech synthesis system and conveyed to person A. Furthermore, the information "headache" and "took medicine" is extracted by a health information extraction system and updated in person A's profile. If an abnormality is detected, person A's family is notified by email.

[0207] Example of a prompt

[0208] "In the following conversation, please provide an explanation automatically. If the user says, 'I have a headache today, so I took some medicine,' the robot should respond, 'You have a headache, please take care of yourself. Don't forget to take your medicine,' and send the health information to the server."

[0209] This will allow for continuous monitoring of the health status of elderly individuals and enable a swift response when abnormalities occur.

[0210] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0211] Step 1:

[0212] The device collects speech from elderly individuals as audio data through a speech recognition device (microphone). The input is the elderly person's voice, and the output is that audio data. This audio data is stored in internal memory for speech recognition processing.

[0213] Step 2:

[0214] The terminal uses the Python `speech_recognition` library to convert collected audio data into text data. The input is audio data, and the output is text data. This conversion process is performed using a speech recognition engine, and the text data will contain content such as "I had a headache today, so I took some medicine."

[0215] Step 3:

[0216] The server analyzes the transformed text data using a generative model (e.g., GPT-3) and generates an appropriate response. The input is text data, and the output is the response text. The generative model understands the context based on the input text data and outputs a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0217] Step 4:

[0218] The terminal uses a speech synthesis system (pyttsx3 library) to generate response text, which is then converted into audio data and output to the elderly person through a speaker. The input is response text, and the output is audio data. This audio data is played back through the speaker so that the elderly person can hear it.

[0219] Step 5:

[0220] The server uses a health information extraction method to extract health-related information from text data of elderly people's daily conversations. The input is text data of conversations, and the output is health information (e.g., "headache" and "took medicine"). This extraction process is performed using natural language processing techniques.

[0221] Step 6:

[0222] The server updates the user profile with extracted health information using a profile update mechanism. The input is health information, and the output is the updated profile data. This update process is carried out by adding the new health information to the user-specific profile database.

[0223] Step 7:

[0224] The server uses a notification system to send notifications to configured contacts when an anomaly is detected. Inputs are profile data and anomaly detection information, and output is a notification message. This notification is sent via methods such as email or SMS.

[0225] Step 8:

[0226] The server uses a health information transmission method to send extracted health information to a remote server. The input is health information, and the output is the transmitted data. This transmission process is carried out, for example, using an HTTP request, and the data is stored on a remote management server.

[0227] Through these steps, it becomes possible to monitor the health status of elderly individuals in real time and take necessary actions quickly.

[0228] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0229] This invention relates to a system that uses a speech recognition system combined with an emotion engine to grasp the emotions and health status of elderly people in real time and to take prompt action as needed. This system consists of a speech recognition means, a speech synthesis means, a generation model means, a health information extraction means, a profile update means, a notification means, and an emotion engine. The details of each component and a specific example of a system combining them are described below.

[0230] Main processing flow

[0231] 1. Collection of audio data (device):

[0232] A voice recognition device, such as a smart speaker, is constantly on standby, waiting for the elderly person (user) to begin speaking. When the user says, "I have a headache today, so I took some medicine," the voice data is collected.

[0233] 2. Speech-to-text conversion (device):

[0234] The collected audio data is sent to a speech recognition system and converted into text data. For example, the text "I had a headache today, so I took some medicine" is generated.

[0235] 3. Sentiment analysis (server):

[0236] Text and audio data are sent to the emotion engine, which analyzes the user's emotional state. For example, emotions such as "sadness" and "anxiety" may be detected.

[0237] 4. Response generation (server):

[0238] The converted text data and analyzed emotional state are sent to a generative model to generate an appropriate response. Based on the emotional state, a response such as "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone." is generated.

[0239] 5. Audio output (device):

[0240] The generated response text is converted into speech data by a speech synthesis system, and the voice is output from the smart speaker. The user hears the response, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[0241] 6. Extraction of health information (server):

[0242] The server extracts health-related information from the user's statements using a health information extraction method. In this example, the health information extracted includes "headache" and "took medicine."

[0243] 7. Profile update (server):

[0244] The extracted health information and emotional state are updated in each user's profile. The server adds the following information to the profile: "Had a headache on 2023-10-01 and took medication. Emotional state: Sad."

[0245] 8. Notifications and alerts (server):

[0246] If the server detects an anomaly in the profile's health information, it sends a notification to a pre-configured contact. For example, a message such as, "Person A complained of a headache and took medication. Their emotional state was analyzed as sadness," might be sent to family members.

[0247] Specific system examples

[0248] Example: In the case of "Mr. A"

[0249] 1. Collection of audio data:

[0250] User A says, "I had a headache today, so I took some medicine."

[0251] Device: A smart speaker collects this statement.

[0252] 2. Text conversion:

[0253] Terminal: Converts the collected audio data into text data: "I had a headache today, so I took some medicine."

[0254] 3. Emotion analysis:

[0255] Server: Text and audio data are sent to the emotion engine, and the emotion "sadness" is analyzed.

[0256] 4. Response generation:

[0257] Server: The generative model analyzes the text data and emotional state to generate a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[0258] 5. Audio output:

[0259] Terminal: The speech synthesis system converts the generated response into speech and outputs the following message from the speaker: "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[0260] 6. Extraction of health information:

[0261] Server: The health information extraction method extracts information such as "headache" and "took medicine" from the text data.

[0262] 7. Profile Update:

[0263] Server: Records the extracted health information and the emotional state of "sadness" in Person A's profile and updates the data to "Had a headache on 2023-10-01 and took medicine. Emotional state: sadness."

[0264] 8. Notifications and alerts:

[0265] Server: Detects anomalies in the health profile (in this case, the emotional states of "headache" and "sadness") and sends a notification to the designated contact (A's family) stating, "A complained of a headache and took medication. Their emotional state was also analyzed as sadness."

[0266] This specific workflow allows for real-time monitoring of the health and emotional state of elderly individuals, enabling prompt responses. By combining it with an emotional intelligence engine, more accurate monitoring and psychological care can be achieved.

[0267] The following describes the processing flow.

[0268] Step 1:

[0269] The user speaks. For example, they might say, "I had a headache today, so I took some medicine."

[0270] Step 2:

[0271] The device collects the user's speech as audio data. This collection is performed by smart speakers or voice recognition devices.

[0272] Step 3:

[0273] The terminal sends the collected voice data to the voice recognition engine and converts it into text data. Specifically, the voice data is converted into the character string "I have a headache today and took medicine."

[0274] Step 4:

[0275] The terminal sends the converted text data to the server. This data transmission initiates the analysis of the text data.

[0276] Step 5:

[0277] The server sends the received text data and voice data to the emotion engine to analyze the user's emotional state. For example, an emotion such as "sadness" is detected from the content of the speech and the voice tone.

[0278] Step 6:

[0279] The server sends the output result of the emotion engine and the text data to the generation model means to generate an appropriate response. Based on the emotional state, the generated response is "You have a headache. Take care of yourself. If you have any concerns, please consult someone."

[0280] Step 7:

[0281] The server sends the generated response text to the terminal. Thereby, the response content is processed on the terminal side.

[0282] Step 8:

[0283] The terminal sends the received response text to the speech synthesis engine and converts it into voice data. Then, the voice "You have a headache. Take care of yourself. If you have any concerns, please consult someone." is output from the smart speaker.[[ID=,41]]

[0284] Step 9:

[0285] The server extracts health information from text data and emotional states. Specifically, in addition to the information of "headache" and "took medicine", it extracts the emotional state of "sadness".

[0286] Step 10:

[0287] The server updates the extracted health information and emotional state to the user's profile. Data such as "had a headache on October 1, 2023 and took medicine. Emotional state: sadness" is added to user A's profile.

[0288] Step 11:

[0289] When the server detects an abnormality, it sends a notification to the set contact. Since the emotional state of "sadness" is detected, the server sends a notification to user A's family saying "User A has complained of a headache and has taken medicine. Also, the emotional state is analyzed as sadness."

[0290] In this way, by performing processing one by one, the health and emotional states of the elderly can be grasped in real time, enabling prompt response. It is expected that this system will further enrich the care for the physical and mental health of the elderly and improve the quality of life.

[0291] (Example 2)

[0292] Next, Example 2 will be described. In the following description, the data processing device 12 is referred to as the "server", and the smart device 14 is referred to as the "terminal".

[0293] Grasping the health and emotional states of the elderly in real time and responding promptly is a difficult task. Especially with the increasing number of elderly people living alone, it is required to accurately monitor the daily health status and emotional changes and provide appropriate support when necessary. Also, in the current situation where the means for medical institutions and families to grasp the status of the elderly are limited, a mechanism that allows the health status of the elderly to be confirmed even from a remote location is needed.

[0294] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0295] In this invention, the server includes speech recognition means, speech synthesis means, generation model means, emotion analysis means, health information extraction means, profile update means, and notification means. This allows for real-time monitoring of the health and emotional state of elderly individuals, even from a remote location, by analyzing their speech content to determine their health and emotional state and generating and outputting appropriate responses. Furthermore, it enables rapid notification in the event of an abnormality, prompting appropriate action.

[0296] "Speech recognition means" refers to a device or software for converting speech data into text data.

[0297] "Speech synthesis means" refers to a device or software for converting text data into speech data.

[0298] "Generative model means" refers to an AI model that generates an appropriate response based on input data.

[0299] "Emotional analysis means" refers to technology for analyzing and identifying a user's emotional state from voice data and text data.

[0300] A "health information extraction means" is a program or device for extracting health-related information from a user's statements or actions.

[0301] A "profile update method" is a system that records and updates extracted health information and emotional states in the user profile.

[0302] A "notification means" is a device or software that sends a notification to a designated contact when an anomaly is detected.

[0303] The present invention is a system that uses a voice recognition system combined with sentiment analysis to monitor the health and emotional states of the elderly in real time and enables prompt responses as needed. This system is composed of voice recognition means, voice synthesis means, generation model means, sentiment analysis means, health information extraction means, profile update means, and notification means.

[0304] Specific examples of hardware and software to be used

[0305] Hardware

[0306] Smart speaker: Used as a terminal to collect the daily conversations of the elderly. Specific examples include Amazon Echo and Google Nest speakers.

[0307] Server: A computer server for data analysis and processing. Cloud-based servers (e.g., AWS, Google Cloud) can also be used.

[0308] Software

[0309] Voice recognition software: Used to convert voice data into text data. Specific examples include Google Speech-to-Text and Microsoft Azure Speech.

[0310] Voice synthesis software: Used to convert text data into voice data. Specific examples include Amazon Polly and Google Text-to-Speech.

[0311] Sentiment analysis engine: Used to analyze sentiment from text data and voice data. Specific examples include IBM Watson Tone Analyzer and Amazon Comprehend.

[0312] Generative AI models: Used to generate appropriate responses based on user input. A specific example is OpenAI GPT-4.

[0313] Database system: Used to record and update health information and emotional status in user profiles. Specific examples include MySQL and MongoDB.

[0314] Notification system: Used to send notifications when an anomaly is detected. Specific examples include Twilio and SMTP servers.

[0315] Specific details of the actions

[0316] Collection of audio data

[0317] The user speaks to the smart speaker, saying, "I have a headache today, so I took some medicine." The smart speaker collects this voice data.

[0318] Speech-to-text conversion

[0319] The device uses voice recognition to convert the voice data into text data that reads, "I had a headache today, so I took some medicine."

[0320] Emotion analysis

[0321] The server sends text and audio data to an emotion analysis engine to analyze emotions such as "sadness" and "anxiety."

[0322] Response generation

[0323] The server generates an appropriate response using a generative model based on the analyzed emotional state. The generated response text will read something like, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[0324] Audio output

[0325] The terminal converts the generated response into voice data using speech synthesis technology and outputs it to the user via the smart speaker.

[0326] Extraction of health information

[0327] The server uses a health information extraction method to extract health information such as "headache" and "took medicine" from the user's speech.

[0328] Profile update

[0329] The server records the extracted health information and emotional state in the user profile. For example, it adds data such as, "Had a headache on 2023-10-01 and took medication. Emotional state: Sad."

[0330] Notifications and alerts

[0331] If an anomaly is detected, the server will send a notification to the designated contact. The message will include something like, "Person A complained of a headache and took medication. Their emotional state was analyzed as sadness."

[0332] Examples of prompt statements

[0333] If someone says, "I have a headache today, so I took some medicine," a specific example of a response like, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone," is generated and output by the speech synthesis system.

[0334] By combining the above elements, the present invention makes it possible to monitor the health and emotional state of elderly people in real time and to take prompt and appropriate action when necessary.

[0335] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0336] Step 1:

[0337] Collection of audio data

[0338] User: An elderly person says, "I have a headache today, so I took some medicine."

[0339] Specific action: An elderly person speaks into a smart speaker.

[0340] Input: User's voice.

[0341] Output: Audio data captured by the smart speaker.

[0342] Step 2:

[0343] Speech-to-text conversion

[0344] Terminal: The smart speaker collects voice data and converts it into text data using speech recognition technology.

[0345] Specific operation: Speech recognition software (e.g., Google Speech-to-Text API) analyzes the audio data and generates a corresponding text string.

[0346] Input: Audio data.

[0347] Output: Text data "I had a headache today, so I took some medicine."

[0348] Step 3:

[0349] Emotion analysis

[0350] Server: The converted text and audio data are sent to the sentiment analysis engine.

[0351] Specific operation: A sentiment analysis engine on the server (e.g., IBM Watson Tone Analyzer) analyzes text data and voice tone to identify the user's emotional state.

[0352] Input: Text data and audio data.

[0353] Output: Emotional analysis results such as "sadness" and "anxiety".

[0354] Step 4:

[0355] Response generation

[0356] Server: Generates an appropriate response using a generative model based on the analyzed emotional state.

[0357] Specific operation: Based on the text input as a prompt and sentiment data, a generative AI model (e.g., OpenAI GPT-4) generates an appropriate response.

[0358] Input: Sentiment analysis results and text data.

[0359] Output: A text response saying, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[0360] Step 5:

[0361] Audio output

[0362] Terminal: The generated response text is converted into speech data using speech synthesis means.

[0363] Specific operation: Text-to-speech software (e.g., Amazon Polly) converts the response text into speech data.

[0364] Input: Text response data.

[0365] Output: Audio data of the response.

[0366] Device: Outputs voice messages to the user from the smart speaker.

[0367] Specific operation: Synthesized audio is played from the speaker and heard by the user.

[0368] Input: Audio data of the response.

[0369] Output: "You have a headache, please take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone." (Voice)

[0370] Step 6:

[0371] Extraction of health information

[0372] Server: Using a health information extraction method, it extracts health information such as "headache" or "took medicine" from the user's statements.

[0373] Specific operation: A text analysis engine (e.g., SpaCy or NLTK) identifies and extracts medical-related keywords from the text.

[0374] Input: Text data.

[0375] Output: Health information such as "headache" or "took medicine".

[0376] Step 7:

[0377] Profile update

[0378] Server: Records extracted health information and emotional state in each user's profile.

[0379] Specific action: The database system (e.g., MySQL or MongoDB) adds the information "Had a headache on 2023-10-01 and took medication. Emotional state: Sad" to the user profile.

[0380] Input: Health information and emotional state.

[0381] Output: Updated user profile.

[0382] Step 8:

[0383] Notifications and alerts

[0384] Server: If an anomaly is detected, it will send a notification to the configured contact.

[0385] Specific action: The notification system (e.g., Twilio or an SMTP server) sends a message to the configured contact stating, "Person A complained of a headache and took medication. Their emotional state was analyzed as sad."

[0386] Input: Health information and emotional state in which abnormalities were detected.

[0387] Output: Sent notification message.

[0388] (Application Example 2)

[0389] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0390] It is extremely important to understand the health and emotional state of the elderly in real time and to respond quickly and appropriately. However, conventional systems lacked the ability to integrate emotion analysis, autonomous response generation, and real-time health monitoring and notification, leaving them insufficient. In particular, the risk of delayed response in emergencies is high for elderly people living alone, so a new system is needed to solve this problem.

[0391] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes a speech recognition means, a speech synthesis means, a generation model means, an emotion analysis means, a health information extraction means, a profile update means, a real-time monitoring means, a push notification system, and a cloud database. This makes it possible to collect voice data from elderly people, perform emotion analysis and extract health information in real time, generate appropriate responses based on the content of speech, and send notifications quickly in emergencies.

[0392] "Speech recognition means" refers to technology that collects speech data and converts that speech into text.

[0393] "Speech synthesis means" refers to a technology that converts text data into speech data and outputs it as speech.

[0394] "Generative model means" refers to AI model technology that generates appropriate responses based on input data.

[0395] A "health information extraction method" is a technology that extracts health-related information from everyday conversations.

[0396] "Profile update means" refers to a technology that adds or updates extracted health information to the user's individual profile.

[0397] "Notification means" refers to a technology that sends notifications to designated contacts when an anomaly is detected.

[0398] "Emotional analysis methods" refer to technologies that analyze a user's emotional state from voice and text data.

[0399] "Real-time monitoring means" refers to technology that collects, analyzes, and processes the speech of elderly people in real time.

[0400] An "autonomous response generation means" is a technology that generates an appropriate response based on the results of an analysis of emotional states and other factors.

[0401] A "cloud database" is an online database that stores and manages data via the internet.

[0402] A "push notification system" is a technology that automatically sends important information and warnings to recipients in real time.

[0403] To implement this invention, a system is needed to collect and analyze voice data of elderly people in real time. This system consists of voice recognition means, voice synthesis means, generative model means, emotion analysis means, health information extraction means, profile update means, real-time monitoring means, push notification system, and cloud database.

[0404] First, the user speaks through a device such as a smartphone or a pet-like robot. The device's speech recognition system collects this speech data and converts it into text data using a speech recognition engine such as Google Cloud Speech-to-Text. For example, if the user says, "I have a headache today, so I took some medicine," that statement is converted into the text data "I have a headache today, so I took some medicine."

[0405] Next, the converted text data is analyzed using sentiment analysis tools such as IBM Watson Tone Analyzer to clarify the user's emotional state. As a result of the analysis, emotions such as "sadness" or "anxiety" may be detected.

[0406] The converted text data and analyzed emotional state are sent to a generative AI model such as OpenAI's GPT-4 to generate an appropriate response. The generated response consists of kind words directed at the user, such as, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[0407] The generated responses are converted into audio data via a speech synthesis engine such as Amazon Polly and played back through a smartphone or robot's speaker. Users can gain a sense of reassurance by listening to the generated audio.

[0408] Furthermore, the server extracts health-related information from text data using a health information extraction mechanism. For example, specific health information such as "headache" and "took medicine" is extracted. The extracted health information and emotional state are then updated in the user profile stored in a cloud database (e.g., Firebase).

[0409] The real-time monitoring system constantly monitors user speech and, if an anomaly is detected, immediately sends a notification to a designated contact (family or medical staff) using a push notification system (e.g., Firebase Cloud Messaging). This notification might contain a message such as, "Person A complained of a headache and took medication. Their emotional state was analyzed as sadness."

[0410] As a concrete example, here is an example of a prompt message:

[0411] User comment: "I had a headache today, so I took some medicine."

[0412] Detected emotion: "sadness"

[0413] Generate an appropriate response.

[0414] As described above, the system for implementing this invention can grasp the health and emotional state of elderly people in real time and respond quickly. This makes it possible to ensure the safety and security of elderly people living alone.

[0415] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0416] Step 1:

[0417] The user speaks. The user uses a device such as a smartphone or a pet-like robot to say something like, "I have a headache today, so I took some medicine." This voice data becomes the system's input.

[0418] Step 2:

[0419] The device collects audio data. Using speech recognition technology (e.g., Google Cloud Speech-to-Text), the collected audio data is converted into text data in real time. Here, the input is audio data, and the output is text data.

[0420] Step 3:

[0421] The server sends text and audio data to an emotion analysis device (e.g., IBM Watson Tone Analyzer) to analyze the emotional state. The input is text and audio data, and the output is the user's emotional state (e.g., "sadness").

[0422] Step 4:

[0423] The server sends text data and analyzed emotional states to a generative model (e.g., OpenAI GPT-4) to generate an appropriate response. The input is text data and emotional states, and the output is the response text (e.g., "You have a headache, take care of yourself.").

[0424] Step 5:

[0425] The server sends the generated response text to a speech synthesis system (e.g., Amazon Polly) to convert it into speech data. The input is the response text, and the output is the response speech data.

[0426] Step 6:

[0427] The device plays a voice response and responds to the user. The user can hear a voice message that says, "You have a headache, please take care of yourself."

[0428] Step 7:

[0429] The server uses a health information extraction method to extract health information from text data. The input is text data, and the output is health information (e.g., "headache," "took medicine").

[0430] Step 8:

[0431] The server accesses a cloud database (e.g., Firebase) and updates the user profile with extracted health and emotional status information. The input is health and emotional status, and the output is the updated profile information.

[0432] Step 9:

[0433] The server continuously monitors user speech using real-time monitoring methods. If an anomaly is detected (e.g., "sadness" or "headache"), a push notification system (e.g., Firebase Cloud Messaging) is used to send a notification to the configured contacts. Inputs are profile information and anomaly detection results, and output is a notification message.

[0434] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0435] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0436] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0437] [Second Embodiment]

[0438] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0439] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0440] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0441] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0442] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0443] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0444] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0445] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0446] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0447] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0448] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0449] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0450] This invention provides a system that enables the understanding of the health status of elderly people living alone through natural conversation and allows for prompt action as needed. This system consists of speech recognition means, speech synthesis means, generative model means, health information extraction means, profile update means, and notification means. Details of each component and specific examples of systems combining them are described below.

[0451] Main processing flow

[0452] 1. Collection of audio data (device):

[0453] A voice recognition device, such as a smart speaker, is constantly on standby, waiting for the elderly person (user) to begin speaking. When the user says, "I have a headache today, so I took some medicine," the voice data is collected.

[0454] 2. Speech-to-text conversion (device):

[0455] The collected audio data is sent to a speech recognition system and converted into text data. For example, the text "I had a headache today, so I took some medicine" is generated.

[0456] 3. Response generation (server):

[0457] The converted text data is sent to the server, where a generative model analyzes it to generate an appropriate response. An example of a generated response is, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0458] 4. Audio output (device):

[0459] The generated response text is converted into speech data by a speech synthesis system, and the voice is output from the smart speaker. The user hears the response, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0460] 5. Extraction of health information (server):

[0461] The server extracts health-related information from the user's statements using a health information extraction method. In this example, the health information extracted includes "headache" and "took medicine."

[0462] 6. Profile update (server):

[0463] The extracted health information is updated in each user's profile. The server adds information to the profile such as "Had a headache on 2023-10-01 and took medication."

[0464] 7. Notifications and alerts (server):

[0465] If the server detects an anomaly in the profile's health information, it sends a notification to a pre-configured contact. For example, a message such as "Mr. / Ms. A complained of a headache and has taken medication" might be sent to a family member.

[0466] Specific system examples

[0467] Example: In the case of "Mr. A"

[0468] 1. Collection of audio data:

[0469] User A says, "I had a headache today, so I took some medicine."

[0470] Device: A smart speaker collects this statement.

[0471] 2. Text conversion:

[0472] Terminal: Converts the collected audio data into text data: "I had a headache today, so I took some medicine."

[0473] 3. Response generation:

[0474] Server: The generative model analyzes the text data and generates the response, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0475] 4. Audio output:

[0476] Terminal: The speech synthesis device converts the generated response into speech and outputs the following message from the speaker: "You have a headache, take care of yourself. Don't forget to take your medicine."

[0477] 5. Extraction of health information:

[0478] Server: The health information extraction method extracts information such as "headache" and "took medicine" from the text data.

[0479] 6. Profile Update:

[0480] Server: Records the extracted health information in Person A's profile and updates the data to "Had a headache on 2023-10-01 and took medication."

[0481] 7. Notifications and alerts:

[0482] Server: Detects an anomaly in the health profile (in this case, "headache") and sends a notification to the designated contact (A's family) stating, "A is complaining of a headache and has taken medication."

[0483] This specific workflow allows for real-time monitoring of the health status of elderly individuals and prompt provision of necessary information to family members and care staff. This system can improve the quality of life for elderly people living alone.

[0484] The following describes the processing flow.

[0485] Step 1:

[0486] The user speaks. For example, they might say, "I had a headache today, so I took some medicine."

[0487] Step 2:

[0488] The device collects the user's speech as audio data. This is done by smart speakers and voice recognition devices.

[0489] Step 3:

[0490] The device calls a speech recognition engine to convert the collected audio data into text data. For example, it might convert the audio data into the text "I had a headache today, so I took some medicine."

[0491] Step 4:

[0492] The terminal sends the converted text data to the server. This initiates the analysis of the text data.

[0493] Step 5:

[0494] The server uses a generative model to analyze the received text data and generate an appropriate response. For example, a response such as "You have a headache, take care of yourself. Don't forget to take your medicine." might be generated.

[0495] Step 6:

[0496] The server sends the generated response text to the terminal.

[0497] Step 7:

[0498] The terminal sends the received response text to a speech synthesis engine, which converts it into speech data. The audio is then output through the speaker. Specifically, the following message is played: "You have a headache, please take care of yourself. Don't forget to take your medicine."

[0499] Step 8:

[0500] The server extracts health information from text data. For example, it might extract health information such as "headache" and "took medication."

[0501] Step 9:

[0502] The server updates the user's profile with the extracted health information. For example, the information "Had a headache on 2023-10-01 and took medication" will be recorded in person A's profile.

[0503] Step 10:

[0504] If the server detects an anomaly, it will send a notification to the designated contact. For example, a notification stating, "Mr. / Ms. A complained of a headache and has taken medication," will be sent to the user's family.

[0505] This processing flow allows for real-time monitoring of the health status of elderly individuals and enables rapid response. The sequential execution of each step ensures the safety and security of the user.

[0506] (Example 1)

[0507] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0508] The aim is to solve the problem of difficulty in monitoring and responding promptly to the daily health conditions of elderly people living alone. In particular, there is a need to provide appropriate responses in real time when elderly people report any abnormalities.

[0509] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0510] In this invention, the server includes a generation model means, a health information extraction means, a profile update means, and a notification means. This makes it possible to analyze the content of elderly people's statements, extract and record health-related information, and send notifications quickly when an abnormality is detected.

[0511] "Speech recognition means" refers to a device and program that collects user speech as audio data and converts it into text data.

[0512] "Speech synthesis means" refers to a device and program for converting text data into speech data and playing it back.

[0513] A "generative model means" is a program that analyzes collected text data and generates an appropriate response.

[0514] "Health information extraction means" refers to a device and program that extracts health-related information from text data.

[0515] "Profile update means" refers to a device and program that records and updates extracted health information in each user's profile.

[0516] "Notification means" refers to a device and program that sends a notification to a pre-configured contact when an anomaly is detected.

[0517] This invention provides a system that enables the understanding of the health status of elderly people living alone through natural conversation and enables prompt action as needed. This system consists of a speech recognition means, a speech synthesis means, a generation model means, a health information extraction means, a profile update means, and a notification means.

[0518] Hardware and software to be used

[0519] 1. Speech recognition means

[0520] Hardware: Smart speaker

[0521] Software: Speech recognition API (e.g., Google Speech-to-Text API)

[0522] 2. Speech synthesis means

[0523] Software: Text-to-speech API (e.g., Google Text-to-Speech API)

[0524] 3. Generative Model Means

[0525] Software: Generative AI models (e.g., GPT-3)

[0526] 4. Health information extraction means

[0527] Software: Natural Language Processing Algorithms

[0528] 5. Profile update method

[0529] Software: Database management systems (e.g., MySQL)

[0530] 6. Means of notification

[0531] Software: Notification API (e.g., Twilio)

[0532] Specific example: In the case of "Mr. A"

[0533] Collection of audio data

[0534] User A says, "I had a headache today, so I took some medicine."

[0535] Device: A smart speaker collects this statement.

[0536] Speech-to-text conversion

[0537] Terminal: The collected audio data is converted into text data, "I had a headache today, so I took some medicine," using a speech recognition API.

[0538] Response generation

[0539] Server: The generative model analyzes the text data and generates the response, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0540] Audio output

[0541] Terminal: The response generated using a speech synthesis API is converted into speech and output from the speaker the following message: "You have a headache, take care of yourself. Don't forget to take your medicine."

[0542] Extraction of health information

[0543] Server: The health information extraction method extracts information such as "headache" and "took medicine" from the text data.

[0544] Profile Update

[0545] Server: Records the extracted health information in Person A's profile and updates the data to "Had a headache on 2023-10-01 and took medication."

[0546] Notifications and alerts

[0547] Server: Detects an anomaly in the health profile (in this case, "headache") and sends a notification to the configured contacts stating, "Person A complained of a headache and has taken medication."

[0548] Example of a prompt

[0549] An elderly person says, "I had a headache today, so I took some medicine."

[0550] The server uses a generated AI model to respond, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0551] This system allows for real-time monitoring of the health status of elderly people living alone and promptly provides necessary information to family members and care staff. This, in turn, can improve the quality of life for elderly people living alone.

[0552] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0553] Step 1:

[0554] Collection of audio data (device)

[0555] Input: The user speaks into the smart speaker.

[0556] Action: The user says, "I have a headache today, so I took some medicine." The smart speaker, acting as the device, uses its microphone to collect this statement.

[0557] Data processing: The microphone converts the audio signal into a digital signal and temporarily stores it in a buffer.

[0558] Output: Digital audio data.

[0559] Step 2:

[0560] Speech-to-text conversion (device)

[0561] Input: Digital audio data.

[0562] Operation: The device sends the collected audio data to a speech recognition API (e.g., Google Speech-to-Text API).

[0563] Data processing: The speech recognition API analyzes the audio data and identifies phonemes and context. This converts the audio data into text data: "I had a headache today, so I took some medicine."

[0564] Output: Text data "I had a headache today, so I took some medicine."

[0565] Step 3:

[0566] Response generation (server)

[0567] Input: Text data "I had a headache today, so I took some medicine."

[0568] Operation: The server inputs the received text data into a generating AI model (e.g., GPT-3) and generates an appropriate response.

[0569] Data processing: The generative AI model analyzes text data and generates a response based on the context. Specifically, it generates a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0570] Output: Response text "You have a headache, please take care of yourself. Don't forget to take your medicine."

[0571] Step 4:

[0572] Audio output (device)

[0573] Input: Response text: "You have a headache, please take care of yourself. Don't forget to take your medicine."

[0574] Operation: The device sends the response text to a text-to-speech API (e.g., Google Text-to-Speech API) and converts it into speech data.

[0575] Data processing: The speech synthesis API analyzes the text data and generates an audio waveform. This audio waveform is then prepared as audio data.

[0576] Output: Response voice data.

[0577] Step 5:

[0578] Extraction of health information (server)

[0579] Input: Text data "I had a headache today, so I took some medicine."

[0580] Operation: The server uses a health information extraction method to identify health-related information from text data.

[0581] Data processing: Analyze the text using natural language processing algorithms and extract health-related keywords such as "headache" and "took medicine."

[0582] Output: Health information "headache" and "took medication".

[0583] Step 6:

[0584] Profile update (server)

[0585] Input: Health information "headache" and "took medication".

[0586] Operation: The server records and updates the extracted health information in the user profile.

[0587] Data processing: Access the user profile database and add the information, "Had a headache on 2023-10-01 and took medication."

[0588] Output: Updated user profile.

[0589] Step 7:

[0590] Notifications and alerts (server)

[0591] Input: Updated user profile.

[0592] Operation: The server detects anomalies from the health profile and sends notifications to the configured contacts.

[0593] Data processing: Anomaly detection algorithms are used to extract anomalous data that matches the specified conditions. Specifically, a notification is generated stating, "Person A complained of a headache and took medication."

[0594] Output: A notification message sent to the contact.

[0595] This system's processing flow makes it possible to monitor the health status of elderly individuals in real time and quickly provide necessary information to their families and care staff.

[0596] (Application Example 1)

[0597] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0598] For elderly people living alone, monitoring their health and ensuring a rapid response in emergencies are major challenges. However, conventional systems have struggled to naturally monitor health conditions through voice conversations and to quickly detect and respond to abnormalities. In particular, the lack of adequate means for appropriate responses in emergencies and for promptly notifying family members and medical professionals has made ensuring the safety of the elderly a significant concern.

[0599] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0600] In this invention, the server includes speech recognition means, speech synthesis means, generation model means, health information extraction means, profile update means, notification means, speech response generation means, and health information transmission means. This makes it possible to analyze the voice spoken by elderly people on a daily basis in real time and constantly monitor their health status. Furthermore, when an abnormality is detected, it is possible to quickly generate an appropriate response, enabling emergency response and immediate notification to family members and medical professionals through remote transmission of health information.

[0601] "Speech recognition means" refers to a technological element that collects speech from elderly people as audio data and converts that audio data into text data.

[0602] "Speech synthesis means" refers to a technological element that converts generated text responses into speech data and transmits it to elderly people.

[0603] A "generative model means" is a technical element that analyzes text data of everyday conversations of elderly people and generates appropriate responses.

[0604] A "health information extraction method" is a technological element that extracts health-related information from everyday conversations.

[0605] A "profile update method" is a technological element that records and updates extracted health information in each user's profile.

[0606] A "notification method" is a technical element that sends a notification to a designated contact when an anomaly is detected.

[0607] A "voice response generation means" is a technological element that automatically generates an appropriate response in the event of an abnormality, thereby addressing the needs of the elderly.

[0608] A "health information transmission means" is a technical element that transmits extracted health information to a remote server.

[0609] The system of this invention provides support for elderly people to live safely and healthily in their homes or care facilities. This system is composed of the following hardware and software.

[0610] Hardware configuration

[0611] 1. Voice recognition device: Use a microphone built into the robot (e.g., USB microphone, built-in microphone of a smart speaker). This will pick up the speech of elderly people.

[0612] 2. Speech synthesis device: Uses the robot's built-in speaker to output the generated response as speech.

[0613] 3. Computing Unit: A server or high-performance computer that manages the entire system.

[0614] Software Configuration

[0615] 1. Speech recognition software: The speech_recognition library in Python is used to convert the speech of elderly people into text data.

[0616] 2. Speech synthesis software: Use the pyttsx3 library in Python to convert text responses into speech.

[0617] 3. Generative Models: AI models that use natural language processing (NLP) to analyze text data and generate appropriate responses. For example, OpenAI's GPT-3 is used.

[0618] 4. Communication method: Use the Python requests library to send health information to a remote server.

[0619] Explanation of the process

[0620] The server first receives audio data transmitted from the speech recognition device and converts it into text data. This process is performed using the speech_recognition library. Next, the generative model analyzes the text data and generates an appropriate response from what the elderly person is saying. The generated response is converted into audio data using the pyttsx3 library and output by the speech synthesis device.

[0621] Furthermore, health-related information is extracted from the generated text data using a health information extraction method. This extracted information is then updated in the user's profile by the server. If an abnormality is detected, a notification is sent to the contacts set by the notification method. In addition, by sending the extracted health information to a remote server using the health information transmission method, the health status of elderly individuals can be continuously monitored.

[0622] Explanation of specific examples

[0623] For example, if person A says, "I have a headache today, so I took some medicine," the system collects this utterance as audio data and converts it into text data. Next, a generative model analyzes this and generates a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine." This response is then spoken by a speech synthesis system and conveyed to person A. Furthermore, the information "headache" and "took medicine" is extracted by a health information extraction system and updated in person A's profile. If an abnormality is detected, person A's family is notified by email.

[0624] Example of a prompt

[0625] "In the following conversation, please provide an explanation automatically. If the user says, 'I have a headache today, so I took some medicine,' the robot should respond, 'You have a headache, please take care of yourself. Don't forget to take your medicine,' and send the health information to the server."

[0626] This will allow for continuous monitoring of the health status of elderly individuals and enable a swift response when abnormalities occur.

[0627] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0628] Step 1:

[0629] The device collects speech from elderly individuals as audio data through a speech recognition device (microphone). The input is the elderly person's voice, and the output is that audio data. This audio data is stored in internal memory for speech recognition processing.

[0630] Step 2:

[0631] The terminal uses the Python `speech_recognition` library to convert collected audio data into text data. The input is audio data, and the output is text data. This conversion process is performed using a speech recognition engine, and the text data will contain content such as "I had a headache today, so I took some medicine."

[0632] Step 3:

[0633] The server analyzes the transformed text data using a generative model (e.g., GPT-3) and generates an appropriate response. The input is text data, and the output is the response text. The generative model understands the context based on the input text data and outputs a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0634] Step 4:

[0635] The terminal uses a speech synthesis system (pyttsx3 library) to generate response text, which is then converted into audio data and output to the elderly person through a speaker. The input is response text, and the output is audio data. This audio data is played back through the speaker so that the elderly person can hear it.

[0636] Step 5:

[0637] The server uses a health information extraction method to extract health-related information from text data of elderly people's daily conversations. The input is text data of conversations, and the output is health information (e.g., "headache" and "took medicine"). This extraction process is performed using natural language processing techniques.

[0638] Step 6:

[0639] The server updates the user profile with extracted health information using a profile update mechanism. The input is health information, and the output is the updated profile data. This update process is carried out by adding the new health information to the user-specific profile database.

[0640] Step 7:

[0641] The server uses a notification system to send notifications to configured contacts when an anomaly is detected. Inputs are profile data and anomaly detection information, and output is a notification message. This notification is sent via methods such as email or SMS.

[0642] Step 8:

[0643] The server uses a health information transmission method to send extracted health information to a remote server. The input is health information, and the output is the transmitted data. This transmission process is carried out, for example, using an HTTP request, and the data is stored on a remote management server.

[0644] Through these steps, it becomes possible to monitor the health status of elderly individuals in real time and take necessary actions quickly.

[0645] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0646] This invention relates to a system that uses a speech recognition system combined with an emotion engine to grasp the emotions and health status of elderly people in real time and to take prompt action as needed. This system consists of a speech recognition means, a speech synthesis means, a generation model means, a health information extraction means, a profile update means, a notification means, and an emotion engine. The details of each component and a specific example of a system combining them are described below.

[0647] Main processing flow

[0648] 1. Collection of audio data (device):

[0649] A voice recognition device, such as a smart speaker, is constantly on standby, waiting for the elderly person (user) to begin speaking. When the user says, "I have a headache today, so I took some medicine," the voice data is collected.

[0650] 2. Speech-to-text conversion (device):

[0651] The collected audio data is sent to a speech recognition system and converted into text data. For example, the text "I had a headache today, so I took some medicine" is generated.

[0652] 3. Sentiment analysis (server):

[0653] Text and audio data are sent to the emotion engine, which analyzes the user's emotional state. For example, emotions such as "sadness" and "anxiety" may be detected.

[0654] 4. Response generation (server):

[0655] The converted text data and analyzed emotional state are sent to a generative model to generate an appropriate response. Based on the emotional state, a response such as "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone." is generated.

[0656] 5. Audio output (device):

[0657] The generated response text is converted into speech data by a speech synthesis system, and the voice is output from the smart speaker. The user hears the response, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[0658] 6. Extraction of health information (server):

[0659] The server extracts health-related information from the user's statements using a health information extraction method. In this example, the health information extracted includes "headache" and "took medicine."

[0660] 7. Profile update (server):

[0661] The extracted health information and emotional state are updated in each user's profile. The server adds the following information to the profile: "Had a headache on 2023-10-01 and took medication. Emotional state: Sad."

[0662] 8. Notifications and alerts (server):

[0663] If the server detects an anomaly in the profile's health information, it sends a notification to a pre-configured contact. For example, a message such as, "Person A complained of a headache and took medication. Their emotional state was analyzed as sadness," might be sent to family members.

[0664] Specific system examples

[0665] Example: In the case of "Mr. A"

[0666] 1. Collection of audio data:

[0667] User A says, "I had a headache today, so I took some medicine."

[0668] Device: A smart speaker collects this statement.

[0669] 2. Text conversion:

[0670] Terminal: Converts the collected audio data into text data: "I had a headache today, so I took some medicine."

[0671] 3. Emotion analysis:

[0672] Server: Text and audio data are sent to the emotion engine, and the emotion "sadness" is analyzed.

[0673] 4. Response generation:

[0674] Server: The generative model analyzes the text data and emotional state to generate a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[0675] 5. Audio output:

[0676] Terminal: The speech synthesis system converts the generated response into speech and outputs the following message from the speaker: "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[0677] 6. Extraction of health information:

[0678] Server: The health information extraction method extracts information such as "headache" and "took medicine" from the text data.

[0679] 7. Profile Update:

[0680] Server: Records the extracted health information and the emotional state of "sadness" in Person A's profile and updates the data to "Had a headache on 2023-10-01 and took medicine. Emotional state: sadness."

[0681] 8. Notifications and alerts:

[0682] Server: Detects anomalies in the health profile (in this case, the emotional states of "headache" and "sadness") and sends a notification to the designated contact (A's family) stating, "A complained of a headache and took medication. Their emotional state was also analyzed as sadness."

[0683] This specific workflow allows for real-time monitoring of the health and emotional state of elderly individuals, enabling prompt responses. By combining it with an emotional intelligence engine, more accurate monitoring and psychological care can be achieved.

[0684] The following describes the processing flow.

[0685] Step 1:

[0686] The user speaks. For example, they might say, "I had a headache today, so I took some medicine."

[0687] Step 2:

[0688] The device collects the user's speech as audio data. This collection is performed by smart speakers or voice recognition devices.

[0689] Step 3:

[0690] The device sends the collected voice data to a speech recognition engine, which converts it into text data. Specifically, the voice data is converted into the string "I had a headache today, so I took some medicine."

[0691] Step 4:

[0692] The terminal sends the converted text data to the server. This data transmission initiates the analysis of the text data.

[0693] Step 5:

[0694] The server sends the received text and audio data to the emotion engine, which then analyzes the user's emotional state. For example, emotions such as "sadness" can be detected from the content of speech and tone of voice.

[0695] Step 6:

[0696] The server sends the output of the emotion engine and text data to the generation model to generate an appropriate response. Based on the emotional state, the generated response will be, "You have a headache, take care of yourself. If you have any concerns, please talk to someone."

[0697] Step 7:

[0698] The server sends the generated response text to the terminal. The terminal then processes the response.

[0699] Step 8:

[0700] The device sends the received response text to a speech synthesis engine, which converts it into speech data. Then, the smart speaker outputs the following voice message: "You have a headache, please take care of yourself. If you have any concerns, please talk to someone."

[0701] Step 9:

[0702] The server extracts health information from text data and emotional states. Specifically, it extracts information such as "headache" and "took medicine," as well as the emotional state of "sadness."

[0703] Step 10:

[0704] The server updates the user's profile with the extracted health information and emotional state. The following data is added to Person A's profile: "Had a headache on 2023-10-01 and took medication. Emotional state: Sad."

[0705] Step 11:

[0706] If the server detects an anomaly, it will send a notification to the designated contacts. Because the emotional state "sadness" was detected, the server will send a notification to Person A's family stating, "Person A complained of a headache and took medication. Also, their emotional state was analyzed as sadness."

[0707] By processing the data step by step in this way, it becomes possible to grasp the health and emotional state of elderly individuals in real time and respond quickly. This system is expected to further enhance the physical and mental care of the elderly and improve their quality of life.

[0708] (Example 2)

[0709] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0710] Monitoring the health and emotional state of elderly individuals in real time and responding promptly is a challenging task. Especially with the increasing number of elderly people living alone, there is a need to accurately monitor their daily health and emotional changes and provide appropriate support when needed. Furthermore, given the limited means available to medical institutions and families to monitor the condition of elderly individuals, a system is needed that allows for monitoring their health remotely.

[0711] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0712] In this invention, the server includes speech recognition means, speech synthesis means, generation model means, emotion analysis means, health information extraction means, profile update means, and notification means. This allows for real-time monitoring of the health and emotional state of elderly individuals, even from a remote location, by analyzing their speech content to determine their health and emotional state and generating and outputting appropriate responses. Furthermore, it enables rapid notification in the event of an abnormality, prompting appropriate action.

[0713] "Speech recognition means" refers to a device or software for converting speech data into text data.

[0714] "Speech synthesis means" refers to a device or software for converting text data into speech data.

[0715] "Generative model means" refers to an AI model that generates an appropriate response based on input data.

[0716] "Emotional analysis means" refers to technology for analyzing and identifying a user's emotional state from voice data and text data.

[0717] A "health information extraction means" is a program or device for extracting health-related information from a user's statements or actions.

[0718] A "profile update method" is a system that records and updates extracted health information and emotional states in the user profile.

[0719] A "notification means" is a device or software that sends a notification to a designated contact when an anomaly is detected.

[0720] This invention provides a system that uses a speech recognition system combined with emotion analysis to monitor the health and emotional state of elderly individuals in real time and enable prompt responses as needed. This system consists of speech recognition means, speech synthesis means, generative model means, emotion analysis means, health information extraction means, profile update means, and notification means.

[0721] Specific examples of hardware and software to be used

[0722] hardware

[0723] Smart speakers: Used as devices to collect everyday conversations of the elderly. Specific examples include Amazon Echo and Google Nest speakers.

[0724] Server: A computer server used for data analysis and processing. Cloud-based servers (e.g., AWS, Google Cloud) can also be used.

[0725] software

[0726] Speech recognition software: Used to convert speech data into text data. Specific examples include Google Speech-to-Text and Microsoft Azure Speech.

[0727] Text-to-speech software: Used to convert text data into speech data. Specific examples include Amazon Polly and Google Text-to-Speech.

[0728] Sentiment analysis engine: Used to analyze emotions from text and audio data. Specific examples include IBM Watson Tone Analyzer and Amazon Comprehend.

[0729] Generative AI models: Used to generate appropriate responses based on user input. A specific example is OpenAI GPT-4.

[0730] Database system: Used to record and update health information and emotional status in user profiles. Specific examples include MySQL and MongoDB.

[0731] Notification system: Used to send notifications when an anomaly is detected. Specific examples include Twilio and SMTP servers.

[0732] Specific details of the actions

[0733] Collection of audio data

[0734] The user speaks to the smart speaker, saying, "I have a headache today, so I took some medicine." The smart speaker collects this voice data.

[0735] Speech-to-text conversion

[0736] The device uses voice recognition to convert the voice data into text data that reads, "I had a headache today, so I took some medicine."

[0737] Emotion analysis

[0738] The server sends text and audio data to an emotion analysis engine to analyze emotions such as "sadness" and "anxiety."

[0739] Response generation

[0740] The server generates an appropriate response using a generative model based on the analyzed emotional state. The generated response text will read something like, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[0741] Audio output

[0742] The terminal converts the generated response into voice data using speech synthesis technology and outputs it to the user via the smart speaker.

[0743] Extraction of health information

[0744] The server uses a health information extraction method to extract health information such as "headache" and "took medicine" from the user's speech.

[0745] Profile update

[0746] The server records the extracted health information and emotional state in the user profile. For example, it adds data such as, "Had a headache on 2023-10-01 and took medication. Emotional state: Sad."

[0747] Notifications and alerts

[0748] If an anomaly is detected, the server will send a notification to the designated contact. The message will include something like, "Person A complained of a headache and took medication. Their emotional state was analyzed as sadness."

[0749] Examples of prompt statements

[0750] If someone says, "I have a headache today, so I took some medicine," a specific example of a response like, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone," is generated and output by the speech synthesis system.

[0751] By combining the above elements, the present invention makes it possible to monitor the health and emotional state of elderly people in real time and to take prompt and appropriate action when necessary.

[0752] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0753] Step 1:

[0754] Collection of audio data

[0755] User: An elderly person says, "I have a headache today, so I took some medicine."

[0756] Specific action: An elderly person speaks into a smart speaker.

[0757] Input: User's voice.

[0758] Output: Audio data captured by the smart speaker.

[0759] Step 2:

[0760] Speech-to-text conversion

[0761] Terminal: The smart speaker collects voice data and converts it into text data using speech recognition technology.

[0762] Specific operation: Speech recognition software (e.g., Google Speech-to-Text API) analyzes the audio data and generates a corresponding text string.

[0763] Input: Audio data.

[0764] Output: Text data "I had a headache today, so I took some medicine."

[0765] Step 3:

[0766] Emotion analysis

[0767] Server: The converted text and audio data are sent to the sentiment analysis engine.

[0768] Specific operation: A sentiment analysis engine on the server (e.g., IBM Watson Tone Analyzer) analyzes text data and voice tone to identify the user's emotional state.

[0769] Input: Text data and audio data.

[0770] Output: Emotional analysis results such as "sadness" and "anxiety".

[0771] Step 4:

[0772] Response generation

[0773] Server: Generates an appropriate response using a generative model based on the analyzed emotional state.

[0774] Specific operation: Based on the text input as a prompt and sentiment data, a generative AI model (e.g., OpenAI GPT-4) generates an appropriate response.

[0775] Input: Sentiment analysis results and text data.

[0776] Output: A text response saying, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[0777] Step 5:

[0778] Audio output

[0779] Terminal: The generated response text is converted into speech data using speech synthesis means.

[0780] Specific operation: Text-to-speech software (e.g., Amazon Polly) converts the response text into speech data.

[0781] Input: Text response data.

[0782] Output: Audio data of the response.

[0783] Device: Outputs voice messages to the user from the smart speaker.

[0784] Specific operation: Synthesized audio is played from the speaker and heard by the user.

[0785] Input: Audio data of the response.

[0786] Output: "You have a headache, please take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone." (Voice)

[0787] Step 6:

[0788] Extraction of health information

[0789] Server: Using a health information extraction method, it extracts health information such as "headache" or "took medicine" from the user's statements.

[0790] Specific operation: A text analysis engine (e.g., SpaCy or NLTK) identifies and extracts medical-related keywords from the text.

[0791] Input: Text data.

[0792] Output: Health information such as "headache" or "took medicine".

[0793] Step 7:

[0794] Profile update

[0795] Server: Records extracted health information and emotional state in each user's profile.

[0796] Specific action: The database system (e.g., MySQL or MongoDB) adds the information "Had a headache on 2023-10-01 and took medication. Emotional state: Sad" to the user profile.

[0797] Input: Health information and emotional state.

[0798] Output: Updated user profile.

[0799] Step 8:

[0800] Notifications and alerts

[0801] Server: If an anomaly is detected, it will send a notification to the configured contact.

[0802] Specific action: The notification system (e.g., Twilio or an SMTP server) sends a message to the configured contact stating, "Person A complained of a headache and took medication. Their emotional state was analyzed as sad."

[0803] Input: Health information and emotional state in which abnormalities were detected.

[0804] Output: Sent notification message.

[0805] (Application Example 2)

[0806] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0807] It is extremely important to understand the health and emotional state of the elderly in real time and to respond quickly and appropriately. However, conventional systems lacked the ability to integrate emotion analysis, autonomous response generation, and real-time health monitoring and notification, leaving them insufficient. In particular, the risk of delayed response in emergencies is high for elderly people living alone, so a new system is needed to solve this problem.

[0808] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes a speech recognition means, a speech synthesis means, a generation model means, an emotion analysis means, a health information extraction means, a profile update means, a real-time monitoring means, a push notification system, and a cloud database. This makes it possible to collect voice data from elderly people, perform emotion analysis and extract health information in real time, generate appropriate responses based on the content of speech, and send notifications quickly in emergencies.

[0809] "Speech recognition means" refers to technology that collects speech data and converts that speech into text.

[0810] "Speech synthesis means" refers to a technology that converts text data into speech data and outputs it as speech.

[0811] "Generative model means" refers to AI model technology that generates appropriate responses based on input data.

[0812] A "health information extraction method" is a technology that extracts health-related information from everyday conversations.

[0813] "Profile update means" refers to a technology that adds or updates extracted health information to the user's individual profile.

[0814] "Notification means" refers to a technology that sends notifications to designated contacts when an anomaly is detected.

[0815] "Emotional analysis methods" refer to technologies that analyze a user's emotional state from voice and text data.

[0816] "Real-time monitoring means" refers to technology that collects, analyzes, and processes the speech of elderly people in real time.

[0817] An "autonomous response generation means" is a technology that generates an appropriate response based on the results of an analysis of emotional states and other factors.

[0818] A "cloud database" is an online database that stores and manages data via the internet.

[0819] A "push notification system" is a technology that automatically sends important information and warnings to recipients in real time.

[0820] To implement this invention, a system is needed to collect and analyze voice data of elderly people in real time. This system consists of voice recognition means, voice synthesis means, generative model means, emotion analysis means, health information extraction means, profile update means, real-time monitoring means, push notification system, and cloud database.

[0821] First, the user speaks through a device such as a smartphone or a pet-like robot. The device's speech recognition system collects this speech data and converts it into text data using a speech recognition engine such as Google Cloud Speech-to-Text. For example, if the user says, "I have a headache today, so I took some medicine," that statement is converted into the text data "I have a headache today, so I took some medicine."

[0822] Next, the converted text data is analyzed using sentiment analysis tools such as IBM Watson Tone Analyzer to clarify the user's emotional state. As a result of the analysis, emotions such as "sadness" or "anxiety" may be detected.

[0823] The converted text data and analyzed emotional state are sent to a generative AI model such as OpenAI's GPT-4 to generate an appropriate response. The generated response consists of kind words directed at the user, such as, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[0824] The generated responses are converted into audio data via a speech synthesis engine such as Amazon Polly and played back through a smartphone or robot's speaker. Users can gain a sense of reassurance by listening to the generated audio.

[0825] Furthermore, the server extracts health-related information from text data using a health information extraction mechanism. For example, specific health information such as "headache" and "took medicine" is extracted. The extracted health information and emotional state are then updated in the user profile stored in a cloud database (e.g., Firebase).

[0826] The real-time monitoring system constantly monitors user speech and, if an anomaly is detected, immediately sends a notification to a designated contact (family or medical staff) using a push notification system (e.g., Firebase Cloud Messaging). This notification might contain a message such as, "Person A complained of a headache and took medication. Their emotional state was analyzed as sadness."

[0827] As a concrete example, here is an example of a prompt message:

[0828] User comment: "I had a headache today, so I took some medicine."

[0829] Detected emotion: "sadness"

[0830] Generate an appropriate response.

[0831] As described above, the system for implementing this invention can grasp the health and emotional state of elderly people in real time and respond quickly. This makes it possible to ensure the safety and security of elderly people living alone.

[0832] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0833] Step 1:

[0834] The user speaks. The user uses a device such as a smartphone or a pet-like robot to say something like, "I have a headache today, so I took some medicine." This voice data becomes the system's input.

[0835] Step 2:

[0836] The device collects audio data. Using speech recognition technology (e.g., Google Cloud Speech-to-Text), the collected audio data is converted into text data in real time. Here, the input is audio data, and the output is text data.

[0837] Step 3:

[0838] The server sends text and audio data to an emotion analysis device (e.g., IBM Watson Tone Analyzer) to analyze the emotional state. The input is text and audio data, and the output is the user's emotional state (e.g., "sadness").

[0839] Step 4:

[0840] The server sends text data and analyzed emotional states to a generative model (e.g., OpenAI GPT-4) to generate an appropriate response. The input is text data and emotional states, and the output is the response text (e.g., "You have a headache, take care of yourself.").

[0841] Step 5:

[0842] The server sends the generated response text to a speech synthesis system (e.g., Amazon Polly) to convert it into speech data. The input is the response text, and the output is the response speech data.

[0843] Step 6:

[0844] The device plays a voice response and responds to the user. The user can hear a voice message that says, "You have a headache, please take care of yourself."

[0845] Step 7:

[0846] The server uses a health information extraction method to extract health information from text data. The input is text data, and the output is health information (e.g., "headache," "took medicine").

[0847] Step 8:

[0848] The server accesses a cloud database (e.g., Firebase) and updates the user profile with extracted health and emotional status information. The input is health and emotional status, and the output is the updated profile information.

[0849] Step 9:

[0850] The server continuously monitors user speech using real-time monitoring methods. If an anomaly is detected (e.g., "sadness" or "headache"), a push notification system (e.g., Firebase Cloud Messaging) is used to send a notification to the configured contacts. Inputs are profile information and anomaly detection results, and output is a notification message.

[0851] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0852] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0853] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0854] [Third Embodiment]

[0855] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0856] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0857] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0858] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0859] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0860] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0861] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0862] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0863] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0864] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0865] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0866] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0867] This invention provides a system that enables the understanding of the health status of elderly people living alone through natural conversation and allows for prompt action as needed. This system consists of speech recognition means, speech synthesis means, generative model means, health information extraction means, profile update means, and notification means. Details of each component and specific examples of systems combining them are described below.

[0868] Main processing flow

[0869] 1. Collection of audio data (device):

[0870] A voice recognition device, such as a smart speaker, is constantly on standby, waiting for the elderly person (user) to begin speaking. When the user says, "I have a headache today, so I took some medicine," the voice data is collected.

[0871] 2. Speech-to-text conversion (device):

[0872] The collected audio data is sent to a speech recognition system and converted into text data. For example, the text "I had a headache today, so I took some medicine" is generated.

[0873] 3. Response generation (server):

[0874] The converted text data is sent to the server, where a generative model analyzes it to generate an appropriate response. An example of a generated response is, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0875] 4. Audio output (device):

[0876] The generated response text is converted into speech data by a speech synthesis system, and the voice is output from the smart speaker. The user hears the response, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0877] 5. Extraction of health information (server):

[0878] The server extracts health-related information from the user's statements using a health information extraction method. In this example, the health information extracted includes "headache" and "took medicine."

[0879] 6. Profile update (server):

[0880] The extracted health information is updated in each user's profile. The server adds information to the profile such as "Had a headache on 2023-10-01 and took medication."

[0881] 7. Notifications and alerts (server):

[0882] If the server detects an anomaly in the profile's health information, it sends a notification to a pre-configured contact. For example, a message such as "Mr. / Ms. A complained of a headache and has taken medication" might be sent to a family member.

[0883] Specific system examples

[0884] Example: In the case of "Mr. A"

[0885] 1. Collection of audio data:

[0886] User A says, "I had a headache today, so I took some medicine."

[0887] Device: A smart speaker collects this statement.

[0888] 2. Text conversion:

[0889] Terminal: Converts the collected audio data into text data: "I had a headache today, so I took some medicine."

[0890] 3. Response generation:

[0891] Server: The generative model analyzes the text data and generates the response, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0892] 4. Audio output:

[0893] Terminal: The speech synthesis device converts the generated response into speech and outputs the following message from the speaker: "You have a headache, take care of yourself. Don't forget to take your medicine."

[0894] 5. Extraction of health information:

[0895] Server: The health information extraction method extracts information such as "headache" and "took medicine" from the text data.

[0896] 6. Profile Update:

[0897] Server: Records the extracted health information in Person A's profile and updates the data to "Had a headache on 2023-10-01 and took medication."

[0898] 7. Notifications and alerts:

[0899] Server: Detects an anomaly in the health profile (in this case, "headache") and sends a notification to the designated contact (A's family) stating, "A is complaining of a headache and has taken medication."

[0900] This specific workflow allows for real-time monitoring of the health status of elderly individuals and prompt provision of necessary information to family members and care staff. This system can improve the quality of life for elderly people living alone.

[0901] The following describes the processing flow.

[0902] Step 1:

[0903] The user speaks. For example, they might say, "I had a headache today, so I took some medicine."

[0904] Step 2:

[0905] The device collects the user's speech as audio data. This is done by smart speakers and voice recognition devices.

[0906] Step 3:

[0907] The device calls a speech recognition engine to convert the collected audio data into text data. For example, it might convert the audio data into the text "I had a headache today, so I took some medicine."

[0908] Step 4:

[0909] The terminal sends the converted text data to the server. This initiates the analysis of the text data.

[0910] Step 5:

[0911] The server uses a generative model to analyze the received text data and generate an appropriate response. For example, a response such as "You have a headache, take care of yourself. Don't forget to take your medicine." might be generated.

[0912] Step 6:

[0913] The server sends the generated response text to the terminal.

[0914] Step 7:

[0915] The terminal sends the received response text to a speech synthesis engine, which converts it into speech data. The audio is then output through the speaker. Specifically, the following message is played: "You have a headache, please take care of yourself. Don't forget to take your medicine."

[0916] Step 8:

[0917] The server extracts health information from text data. For example, it might extract health information such as "headache" and "took medication."

[0918] Step 9:

[0919] The server updates the user's profile with the extracted health information. For example, the information "Had a headache on 2023-10-01 and took medication" will be recorded in person A's profile.

[0920] Step 10:

[0921] If the server detects an anomaly, it will send a notification to the designated contact. For example, a notification stating, "Mr. / Ms. A complained of a headache and has taken medication," will be sent to the user's family.

[0922] This processing flow allows for real-time monitoring of the health status of elderly individuals and enables rapid response. The sequential execution of each step ensures the safety and security of the user.

[0923] (Example 1)

[0924] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0925] The aim is to solve the problem of difficulty in monitoring and responding promptly to the daily health conditions of elderly people living alone. In particular, there is a need to provide appropriate responses in real time when elderly people report any abnormalities.

[0926] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0927] In this invention, the server includes a generation model means, a health information extraction means, a profile update means, and a notification means. This makes it possible to analyze the content of elderly people's statements, extract and record health-related information, and send notifications quickly when an abnormality is detected.

[0928] "Speech recognition means" refers to a device and program that collects user speech as audio data and converts it into text data.

[0929] "Speech synthesis means" refers to a device and program for converting text data into speech data and playing it back.

[0930] A "generative model means" is a program that analyzes collected text data and generates an appropriate response.

[0931] "Health information extraction means" refers to a device and program that extracts health-related information from text data.

[0932] "Profile update means" refers to a device and program that records and updates extracted health information in each user's profile.

[0933] "Notification means" refers to a device and program that sends a notification to a pre-configured contact when an anomaly is detected.

[0934] This invention provides a system that enables the understanding of the health status of elderly people living alone through natural conversation and enables prompt action as needed. This system consists of a speech recognition means, a speech synthesis means, a generation model means, a health information extraction means, a profile update means, and a notification means.

[0935] Hardware and software to be used

[0936] 1. Speech recognition means

[0937] Hardware: Smart speaker

[0938] Software: Speech recognition API (e.g., Google Speech-to-Text API)

[0939] 2. Speech synthesis means

[0940] Software: Text-to-speech API (e.g., Google Text-to-Speech API)

[0941] 3. Generative Model Means

[0942] Software: Generative AI models (e.g., GPT-3)

[0943] 4. Health information extraction means

[0944] Software: Natural Language Processing Algorithms

[0945] 5. Profile update method

[0946] Software: Database management systems (e.g., MySQL)

[0947] 6. Means of notification

[0948] Software: Notification API (e.g., Twilio)

[0949] Specific example: In the case of "Mr. A"

[0950] Collection of audio data

[0951] User A says, "I had a headache today, so I took some medicine."

[0952] Device: A smart speaker collects this statement.

[0953] Speech-to-text conversion

[0954] Terminal: The collected audio data is converted into text data, "I had a headache today, so I took some medicine," using a speech recognition API.

[0955] Response generation

[0956] Server: The generative model analyzes the text data and generates the response, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0957] Audio output

[0958] Terminal: The response generated using a speech synthesis API is converted into speech and output from the speaker the following message: "You have a headache, take care of yourself. Don't forget to take your medicine."

[0959] Extraction of health information

[0960] Server: The health information extraction method extracts information such as "headache" and "took medicine" from the text data.

[0961] Profile Update

[0962] Server: Records the extracted health information in Person A's profile and updates the data to "Had a headache on 2023-10-01 and took medication."

[0963] Notifications and alerts

[0964] Server: Detects an anomaly in the health profile (in this case, "headache") and sends a notification to the configured contacts stating, "Person A complained of a headache and has taken medication."

[0965] Example of a prompt

[0966] An elderly person says, "I had a headache today, so I took some medicine."

[0967] The server uses a generated AI model to respond, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0968] This system allows for real-time monitoring of the health status of elderly people living alone and promptly provides necessary information to family members and care staff. This, in turn, can improve the quality of life for elderly people living alone.

[0969] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0970] Step 1:

[0971] Collection of audio data (device)

[0972] Input: The user speaks into the smart speaker.

[0973] Action: The user says, "I have a headache today, so I took some medicine." The smart speaker, acting as the device, uses its microphone to collect this statement.

[0974] Data processing: The microphone converts the audio signal into a digital signal and temporarily stores it in a buffer.

[0975] Output: Digital audio data.

[0976] Step 2:

[0977] Speech-to-text conversion (device)

[0978] Input: Digital audio data.

[0979] Operation: The device sends the collected audio data to a speech recognition API (e.g., Google Speech-to-Text API).

[0980] Data processing: The speech recognition API analyzes the audio data and identifies phonemes and context. This converts the audio data into text data: "I had a headache today, so I took some medicine."

[0981] Output: Text data "I had a headache today, so I took some medicine."

[0982] Step 3:

[0983] Response generation (server)

[0984] Input: Text data "I had a headache today, so I took some medicine."

[0985] Operation: The server inputs the received text data into a generating AI model (e.g., GPT-3) and generates an appropriate response.

[0986] Data processing: The generative AI model analyzes text data and generates a response based on the context. Specifically, it generates a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine."

[0987] Output: Response text "You have a headache, please take care of yourself. Don't forget to take your medicine."

[0988] Step 4:

[0989] Audio output (device)

[0990] Input: Response text: "You have a headache, please take care of yourself. Don't forget to take your medicine."

[0991] Operation: The device sends the response text to a text-to-speech API (e.g., Google Text-to-Speech API) and converts it into speech data.

[0992] Data processing: The speech synthesis API analyzes the text data and generates an audio waveform. This audio waveform is then prepared as audio data.

[0993] Output: Response voice data.

[0994] Step 5:

[0995] Extraction of health information (server)

[0996] Input: Text data "I had a headache today, so I took some medicine."

[0997] Operation: The server uses a health information extraction method to identify health-related information from text data.

[0998] Data processing: Analyze the text using natural language processing algorithms and extract health-related keywords such as "headache" and "took medicine."

[0999] Output: Health information "headache" and "took medication".

[1000] Step 6:

[1001] Profile update (server)

[1002] Input: Health information "headache" and "took medication".

[1003] Operation: The server records and updates the extracted health information in the user profile.

[1004] Data processing: Access the user profile database and add the information, "Had a headache on 2023-10-01 and took medication."

[1005] Output: Updated user profile.

[1006] Step 7:

[1007] Notifications and alerts (server)

[1008] Input: Updated user profile.

[1009] Operation: The server detects anomalies from the health profile and sends notifications to the configured contacts.

[1010] Data processing: Anomaly detection algorithms are used to extract anomalous data that matches the specified conditions. Specifically, a notification is generated stating, "Person A complained of a headache and took medication."

[1011] Output: A notification message sent to the contact.

[1012] This system's processing flow makes it possible to monitor the health status of elderly individuals in real time and quickly provide necessary information to their families and care staff.

[1013] (Application Example 1)

[1014] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1015] For elderly people living alone, monitoring their health and ensuring a rapid response in emergencies are major challenges. However, conventional systems have struggled to naturally monitor health conditions through voice conversations and to quickly detect and respond to abnormalities. In particular, the lack of adequate means for appropriate responses in emergencies and for promptly notifying family members and medical professionals has made ensuring the safety of the elderly a significant concern.

[1016] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1017] In this invention, the server includes speech recognition means, speech synthesis means, generation model means, health information extraction means, profile update means, notification means, speech response generation means, and health information transmission means. This makes it possible to analyze the voice spoken by elderly people on a daily basis in real time and constantly monitor their health status. Furthermore, when an abnormality is detected, it is possible to quickly generate an appropriate response, enabling emergency response and immediate notification to family members and medical professionals through remote transmission of health information.

[1018] "Speech recognition means" refers to a technological element that collects speech from elderly people as audio data and converts that audio data into text data.

[1019] "Speech synthesis means" refers to a technological element that converts generated text responses into speech data and transmits it to elderly people.

[1020] A "generative model means" is a technical element that analyzes text data of everyday conversations of elderly people and generates appropriate responses.

[1021] A "health information extraction method" is a technological element that extracts health-related information from everyday conversations.

[1022] A "profile update method" is a technological element that records and updates extracted health information in each user's profile.

[1023] A "notification method" is a technical element that sends a notification to a designated contact when an anomaly is detected.

[1024] A "voice response generation means" is a technological element that automatically generates an appropriate response in the event of an abnormality, thereby addressing the needs of the elderly.

[1025] A "health information transmission means" is a technical element that transmits extracted health information to a remote server.

[1026] The system of this invention provides support for elderly people to live safely and healthily in their homes or care facilities. This system is composed of the following hardware and software.

[1027] Hardware configuration

[1028] 1. Voice recognition device: Use a microphone built into the robot (e.g., USB microphone, built-in microphone of a smart speaker). This will pick up the speech of elderly people.

[1029] 2. Speech synthesis device: Uses the robot's built-in speaker to output the generated response as speech.

[1030] 3. Computing Unit: A server or high-performance computer that manages the entire system.

[1031] Software Configuration

[1032] 1. Speech recognition software: The speech_recognition library in Python is used to convert the speech of elderly people into text data.

[1033] 2. Speech synthesis software: Use the pyttsx3 library in Python to convert text responses into speech.

[1034] 3. Generative Models: AI models that use natural language processing (NLP) to analyze text data and generate appropriate responses. For example, OpenAI's GPT-3 is used.

[1035] 4. Communication method: Use the Python requests library to send health information to a remote server.

[1036] Explanation of the process

[1037] The server first receives audio data transmitted from the speech recognition device and converts it into text data. This process is performed using the speech_recognition library. Next, the generative model analyzes the text data and generates an appropriate response from what the elderly person is saying. The generated response is converted into audio data using the pyttsx3 library and output by the speech synthesis device.

[1038] Furthermore, health-related information is extracted from the generated text data using a health information extraction method. This extracted information is then updated in the user's profile by the server. If an abnormality is detected, a notification is sent to the contacts set by the notification method. In addition, by sending the extracted health information to a remote server using the health information transmission method, the health status of elderly individuals can be continuously monitored.

[1039] Explanation of specific examples

[1040] For example, if person A says, "I have a headache today, so I took some medicine," the system collects this utterance as audio data and converts it into text data. Next, a generative model analyzes this and generates a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine." This response is then spoken by a speech synthesis system and conveyed to person A. Furthermore, the information "headache" and "took medicine" is extracted by a health information extraction system and updated in person A's profile. If an abnormality is detected, person A's family is notified by email.

[1041] Example of a prompt

[1042] "In the following conversation, please provide an explanation automatically. If the user says, 'I have a headache today, so I took some medicine,' the robot should respond, 'You have a headache, please take care of yourself. Don't forget to take your medicine,' and send the health information to the server."

[1043] This will allow for continuous monitoring of the health status of elderly individuals and enable a swift response when abnormalities occur.

[1044] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1045] Step 1:

[1046] The device collects speech from elderly individuals as audio data through a speech recognition device (microphone). The input is the elderly person's voice, and the output is that audio data. This audio data is stored in internal memory for speech recognition processing.

[1047] Step 2:

[1048] The terminal uses the Python `speech_recognition` library to convert collected audio data into text data. The input is audio data, and the output is text data. This conversion process is performed using a speech recognition engine, and the text data will contain content such as "I had a headache today, so I took some medicine."

[1049] Step 3:

[1050] The server analyzes the transformed text data using a generative model (e.g., GPT-3) and generates an appropriate response. The input is text data, and the output is the response text. The generative model understands the context based on the input text data and outputs a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine."

[1051] Step 4:

[1052] The terminal uses a speech synthesis system (pyttsx3 library) to generate response text, which is then converted into audio data and output to the elderly person through a speaker. The input is response text, and the output is audio data. This audio data is played back through the speaker so that the elderly person can hear it.

[1053] Step 5:

[1054] The server uses a health information extraction method to extract health-related information from text data of elderly people's daily conversations. The input is text data of conversations, and the output is health information (e.g., "headache" and "took medicine"). This extraction process is performed using natural language processing techniques.

[1055] Step 6:

[1056] The server updates the user profile with extracted health information using a profile update mechanism. The input is health information, and the output is the updated profile data. This update process is carried out by adding the new health information to the user-specific profile database.

[1057] Step 7:

[1058] The server uses a notification system to send notifications to configured contacts when an anomaly is detected. Inputs are profile data and anomaly detection information, and output is a notification message. This notification is sent via methods such as email or SMS.

[1059] Step 8:

[1060] The server uses a health information transmission method to send extracted health information to a remote server. The input is health information, and the output is the transmitted data. This transmission process is carried out, for example, using an HTTP request, and the data is stored on a remote management server.

[1061] Through these steps, it becomes possible to monitor the health status of elderly individuals in real time and take necessary actions quickly.

[1062] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1063] This invention relates to a system that uses a speech recognition system combined with an emotion engine to grasp the emotions and health status of elderly people in real time and to take prompt action as needed. This system consists of a speech recognition means, a speech synthesis means, a generation model means, a health information extraction means, a profile update means, a notification means, and an emotion engine. The details of each component and a specific example of a system combining them are described below.

[1064] Main processing flow

[1065] 1. Collection of audio data (device):

[1066] A voice recognition device, such as a smart speaker, is constantly on standby, waiting for the elderly person (user) to begin speaking. When the user says, "I have a headache today, so I took some medicine," the voice data is collected.

[1067] 2. Speech-to-text conversion (device):

[1068] The collected audio data is sent to a speech recognition system and converted into text data. For example, the text "I had a headache today, so I took some medicine" is generated.

[1069] 3. Sentiment analysis (server):

[1070] Text and audio data are sent to the emotion engine, which analyzes the user's emotional state. For example, emotions such as "sadness" and "anxiety" may be detected.

[1071] 4. Response generation (server):

[1072] The converted text data and analyzed emotional state are sent to a generative model to generate an appropriate response. Based on the emotional state, a response such as "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone." is generated.

[1073] 5. Audio output (device):

[1074] The generated response text is converted into speech data by a speech synthesis system, and the voice is output from the smart speaker. The user hears the response, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[1075] 6. Extraction of health information (server):

[1076] The server extracts health-related information from the user's statements using a health information extraction method. In this example, the health information extracted includes "headache" and "took medicine."

[1077] 7. Profile update (server):

[1078] The extracted health information and emotional state are updated in each user's profile. The server adds the following information to the profile: "Had a headache on 2023-10-01 and took medication. Emotional state: Sad."

[1079] 8. Notifications and alerts (server):

[1080] If the server detects an anomaly in the profile's health information, it sends a notification to a pre-configured contact. For example, a message such as, "Person A complained of a headache and took medication. Their emotional state was analyzed as sadness," might be sent to family members.

[1081] Specific system examples

[1082] Example: In the case of "Mr. A"

[1083] 1. Collection of audio data:

[1084] User A says, "I had a headache today, so I took some medicine."

[1085] Device: A smart speaker collects this statement.

[1086] 2. Text conversion:

[1087] Terminal: Converts the collected audio data into text data: "I had a headache today, so I took some medicine."

[1088] 3. Emotion analysis:

[1089] Server: Text and audio data are sent to the emotion engine, and the emotion "sadness" is analyzed.

[1090] 4. Response generation:

[1091] Server: The generative model analyzes the text data and emotional state to generate a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[1092] 5. Audio output:

[1093] Terminal: The speech synthesis system converts the generated response into speech and outputs the following message from the speaker: "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[1094] 6. Extraction of health information:

[1095] Server: The health information extraction method extracts information such as "headache" and "took medicine" from the text data.

[1096] 7. Profile Update:

[1097] Server: Records the extracted health information and the emotional state of "sadness" in Person A's profile and updates the data to "Had a headache on 2023-10-01 and took medicine. Emotional state: sadness."

[1098] 8. Notifications and alerts:

[1099] Server: Detects anomalies in the health profile (in this case, the emotional states of "headache" and "sadness") and sends a notification to the designated contact (A's family) stating, "A complained of a headache and took medication. Their emotional state was also analyzed as sadness."

[1100] This specific workflow allows for real-time monitoring of the health and emotional state of elderly individuals, enabling prompt responses. By combining it with an emotional intelligence engine, more accurate monitoring and psychological care can be achieved.

[1101] The following describes the processing flow.

[1102] Step 1:

[1103] The user speaks. For example, they might say, "I had a headache today, so I took some medicine."

[1104] Step 2:

[1105] The device collects the user's speech as audio data. This collection is performed by smart speakers or voice recognition devices.

[1106] Step 3:

[1107] The device sends the collected voice data to a speech recognition engine, which converts it into text data. Specifically, the voice data is converted into the string "I had a headache today, so I took some medicine."

[1108] Step 4:

[1109] The terminal sends the converted text data to the server. This data transmission initiates the analysis of the text data.

[1110] Step 5:

[1111] The server sends the received text and audio data to the emotion engine, which then analyzes the user's emotional state. For example, emotions such as "sadness" can be detected from the content of speech and tone of voice.

[1112] Step 6:

[1113] The server sends the output of the emotion engine and text data to the generation model to generate an appropriate response. Based on the emotional state, the generated response will be, "You have a headache, take care of yourself. If you have any concerns, please talk to someone."

[1114] Step 7:

[1115] The server sends the generated response text to the terminal. The terminal then processes the response.

[1116] Step 8:

[1117] The device sends the received response text to a speech synthesis engine, which converts it into speech data. Then, the smart speaker outputs the following voice message: "You have a headache, please take care of yourself. If you have any concerns, please talk to someone."

[1118] Step 9:

[1119] The server extracts health information from text data and emotional states. Specifically, it extracts information such as "headache" and "took medicine," as well as the emotional state of "sadness."

[1120] Step 10:

[1121] The server updates the user's profile with the extracted health information and emotional state. The following data is added to Person A's profile: "Had a headache on 2023-10-01 and took medication. Emotional state: Sad."

[1122] Step 11:

[1123] If the server detects an anomaly, it will send a notification to the designated contacts. Because the emotional state "sadness" was detected, the server will send a notification to Person A's family stating, "Person A complained of a headache and took medication. Also, their emotional state was analyzed as sadness."

[1124] By processing the data step by step in this way, it becomes possible to grasp the health and emotional state of elderly individuals in real time and respond quickly. This system is expected to further enhance the physical and mental care of the elderly and improve their quality of life.

[1125] (Example 2)

[1126] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1127] Monitoring the health and emotional state of elderly individuals in real time and responding promptly is a challenging task. Especially with the increasing number of elderly people living alone, there is a need to accurately monitor their daily health and emotional changes and provide appropriate support when needed. Furthermore, given the limited means available to medical institutions and families to monitor the condition of elderly individuals, a system is needed that allows for monitoring their health remotely.

[1128] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1129] In this invention, the server includes speech recognition means, speech synthesis means, generation model means, emotion analysis means, health information extraction means, profile update means, and notification means. This allows for real-time monitoring of the health and emotional state of elderly individuals, even from a remote location, by analyzing their speech content to determine their health and emotional state and generating and outputting appropriate responses. Furthermore, it enables rapid notification in the event of an abnormality, prompting appropriate action.

[1130] "Speech recognition means" refers to a device or software for converting speech data into text data.

[1131] "Speech synthesis means" refers to a device or software for converting text data into speech data.

[1132] "Generative model means" refers to an AI model that generates an appropriate response based on input data.

[1133] "Emotional analysis means" refers to technology for analyzing and identifying a user's emotional state from voice data and text data.

[1134] A "health information extraction means" is a program or device for extracting health-related information from a user's statements or actions.

[1135] A "profile update method" is a system that records and updates extracted health information and emotional states in the user profile.

[1136] A "notification means" is a device or software that sends a notification to a designated contact when an anomaly is detected.

[1137] This invention provides a system that uses a speech recognition system combined with emotion analysis to monitor the health and emotional state of elderly individuals in real time and enable prompt responses as needed. This system consists of speech recognition means, speech synthesis means, generative model means, emotion analysis means, health information extraction means, profile update means, and notification means.

[1138] Specific examples of hardware and software to be used

[1139] hardware

[1140] Smart speakers: Used as devices to collect everyday conversations of the elderly. Specific examples include Amazon Echo and Google Nest speakers.

[1141] Server: A computer server used for data analysis and processing. Cloud-based servers (e.g., AWS, Google Cloud) can also be used.

[1142] software

[1143] Speech recognition software: Used to convert speech data into text data. Specific examples include Google Speech-to-Text and Microsoft Azure Speech.

[1144] Text-to-speech software: Used to convert text data into speech data. Specific examples include Amazon Polly and Google Text-to-Speech.

[1145] Sentiment analysis engine: Used to analyze emotions from text and audio data. Specific examples include IBM Watson Tone Analyzer and Amazon Comprehend.

[1146] Generative AI models: Used to generate appropriate responses based on user input. A specific example is OpenAI GPT-4.

[1147] Database system: Used to record and update health information and emotional status in user profiles. Specific examples include MySQL and MongoDB.

[1148] Notification system: Used to send notifications when an anomaly is detected. Specific examples include Twilio and SMTP servers.

[1149] Specific details of the actions

[1150] Collection of audio data

[1151] The user speaks to the smart speaker, saying, "I have a headache today, so I took some medicine." The smart speaker collects this voice data.

[1152] Speech-to-text conversion

[1153] The device uses voice recognition to convert the voice data into text data that reads, "I had a headache today, so I took some medicine."

[1154] Emotion analysis

[1155] The server sends text and audio data to an emotion analysis engine to analyze emotions such as "sadness" and "anxiety."

[1156] Response generation

[1157] The server generates an appropriate response using a generative model based on the analyzed emotional state. The generated response text will read something like, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[1158] Audio output

[1159] The terminal converts the generated response into voice data using speech synthesis technology and outputs it to the user via the smart speaker.

[1160] Extraction of health information

[1161] The server uses a health information extraction method to extract health information such as "headache" and "took medicine" from the user's speech.

[1162] Profile update

[1163] The server records the extracted health information and emotional state in the user profile. For example, it adds data such as, "Had a headache on 2023-10-01 and took medication. Emotional state: Sad."

[1164] Notifications and alerts

[1165] If an anomaly is detected, the server will send a notification to the designated contact. The message will include something like, "Person A complained of a headache and took medication. Their emotional state was analyzed as sadness."

[1166] Examples of prompt statements

[1167] If someone says, "I have a headache today, so I took some medicine," a specific example of a response like, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone," is generated and output by the speech synthesis system.

[1168] By combining the above elements, the present invention makes it possible to monitor the health and emotional state of elderly people in real time and to take prompt and appropriate action when necessary.

[1169] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1170] Step 1:

[1171] Collection of audio data

[1172] User: An elderly person says, "I have a headache today, so I took some medicine."

[1173] Specific action: An elderly person speaks into a smart speaker.

[1174] Input: User's voice.

[1175] Output: Audio data captured by the smart speaker.

[1176] Step 2:

[1177] Speech-to-text conversion

[1178] Terminal: The smart speaker collects voice data and converts it into text data using speech recognition technology.

[1179] Specific operation: Speech recognition software (e.g., Google Speech-to-Text API) analyzes the audio data and generates a corresponding text string.

[1180] Input: Audio data.

[1181] Output: Text data "I had a headache today, so I took some medicine."

[1182] Step 3:

[1183] Emotion analysis

[1184] Server: The converted text and audio data are sent to the sentiment analysis engine.

[1185] Specific operation: A sentiment analysis engine on the server (e.g., IBM Watson Tone Analyzer) analyzes text data and voice tone to identify the user's emotional state.

[1186] Input: Text data and audio data.

[1187] Output: Emotional analysis results such as "sadness" and "anxiety".

[1188] Step 4:

[1189] Response generation

[1190] Server: Generates an appropriate response using a generative model based on the analyzed emotional state.

[1191] Specific operation: Based on the text input as a prompt and sentiment data, a generative AI model (e.g., OpenAI GPT-4) generates an appropriate response.

[1192] Input: Sentiment analysis results and text data.

[1193] Output: A text response saying, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[1194] Step 5:

[1195] Audio output

[1196] Terminal: The generated response text is converted into speech data using speech synthesis means.

[1197] Specific operation: Text-to-speech software (e.g., Amazon Polly) converts the response text into speech data.

[1198] Input: Text response data.

[1199] Output: Audio data of the response.

[1200] Device: Outputs voice messages to the user from the smart speaker.

[1201] Specific operation: Synthesized audio is played from the speaker and heard by the user.

[1202] Input: Audio data of the response.

[1203] Output: "You have a headache, please take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone." (Voice)

[1204] Step 6:

[1205] Extraction of health information

[1206] Server: Using a health information extraction method, it extracts health information such as "headache" or "took medicine" from the user's statements.

[1207] Specific operation: A text analysis engine (e.g., SpaCy or NLTK) identifies and extracts medical-related keywords from the text.

[1208] Input: Text data.

[1209] Output: Health information such as "headache" or "took medicine".

[1210] Step 7:

[1211] Profile update

[1212] Server: Records extracted health information and emotional state in each user's profile.

[1213] Specific action: The database system (e.g., MySQL or MongoDB) adds the information "Had a headache on 2023-10-01 and took medication. Emotional state: Sad" to the user profile.

[1214] Input: Health information and emotional state.

[1215] Output: Updated user profile.

[1216] Step 8:

[1217] Notifications and alerts

[1218] Server: If an anomaly is detected, it will send a notification to the configured contact.

[1219] Specific action: The notification system (e.g., Twilio or an SMTP server) sends a message to the configured contact stating, "Person A complained of a headache and took medication. Their emotional state was analyzed as sad."

[1220] Input: Health information and emotional state in which abnormalities were detected.

[1221] Output: Sent notification message.

[1222] (Application Example 2)

[1223] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1224] It is extremely important to understand the health and emotional state of the elderly in real time and to respond quickly and appropriately. However, conventional systems lacked the ability to integrate emotion analysis, autonomous response generation, and real-time health monitoring and notification, leaving them insufficient. In particular, the risk of delayed response in emergencies is high for elderly people living alone, so a new system is needed to solve this problem.

[1225] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes a speech recognition means, a speech synthesis means, a generation model means, an emotion analysis means, a health information extraction means, a profile update means, a real-time monitoring means, a push notification system, and a cloud database. This makes it possible to collect voice data from elderly people, perform emotion analysis and extract health information in real time, generate appropriate responses based on the content of speech, and send notifications quickly in emergencies.

[1226] "Speech recognition means" refers to technology that collects speech data and converts that speech into text.

[1227] "Speech synthesis means" refers to a technology that converts text data into speech data and outputs it as speech.

[1228] "Generative model means" refers to AI model technology that generates appropriate responses based on input data.

[1229] A "health information extraction method" is a technology that extracts health-related information from everyday conversations.

[1230] "Profile update means" refers to a technology that adds or updates extracted health information to the user's individual profile.

[1231] "Notification means" refers to a technology that sends notifications to designated contacts when an anomaly is detected.

[1232] "Emotional analysis methods" refer to technologies that analyze a user's emotional state from voice and text data.

[1233] "Real-time monitoring means" refers to technology that collects, analyzes, and processes the speech of elderly people in real time.

[1234] An "autonomous response generation means" is a technology that generates an appropriate response based on the results of an analysis of emotional states and other factors.

[1235] A "cloud database" is an online database that stores and manages data via the internet.

[1236] A "push notification system" is a technology that automatically sends important information and warnings to recipients in real time.

[1237] To implement this invention, a system is needed to collect and analyze voice data of elderly people in real time. This system consists of voice recognition means, voice synthesis means, generative model means, emotion analysis means, health information extraction means, profile update means, real-time monitoring means, push notification system, and cloud database.

[1238] First, the user speaks through a device such as a smartphone or a pet-like robot. The device's speech recognition system collects this speech data and converts it into text data using a speech recognition engine such as Google Cloud Speech-to-Text. For example, if the user says, "I have a headache today, so I took some medicine," that statement is converted into the text data "I have a headache today, so I took some medicine."

[1239] Next, the converted text data is analyzed using sentiment analysis tools such as IBM Watson Tone Analyzer to clarify the user's emotional state. As a result of the analysis, emotions such as "sadness" or "anxiety" may be detected.

[1240] The converted text data and analyzed emotional state are sent to a generative AI model such as OpenAI's GPT-4 to generate an appropriate response. The generated response consists of kind words directed at the user, such as, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[1241] The generated responses are converted into audio data via a speech synthesis engine such as Amazon Polly and played back through a smartphone or robot's speaker. Users can gain a sense of reassurance by listening to the generated audio.

[1242] Furthermore, the server extracts health-related information from text data using a health information extraction mechanism. For example, specific health information such as "headache" and "took medicine" is extracted. The extracted health information and emotional state are then updated in the user profile stored in a cloud database (e.g., Firebase).

[1243] The real-time monitoring system constantly monitors user speech and, if an anomaly is detected, immediately sends a notification to a designated contact (family or medical staff) using a push notification system (e.g., Firebase Cloud Messaging). This notification might contain a message such as, "Person A complained of a headache and took medication. Their emotional state was analyzed as sadness."

[1244] As a concrete example, here is an example of a prompt message:

[1245] User comment: "I had a headache today, so I took some medicine."

[1246] Detected emotion: "sadness"

[1247] Generate an appropriate response.

[1248] As described above, the system for implementing this invention can grasp the health and emotional state of elderly people in real time and respond quickly. This makes it possible to ensure the safety and security of elderly people living alone.

[1249] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1250] Step 1:

[1251] The user speaks. The user uses a device such as a smartphone or a pet-like robot to say something like, "I have a headache today, so I took some medicine." This voice data becomes the system's input.

[1252] Step 2:

[1253] The device collects audio data. Using speech recognition technology (e.g., Google Cloud Speech-to-Text), the collected audio data is converted into text data in real time. Here, the input is audio data, and the output is text data.

[1254] Step 3:

[1255] The server sends text and audio data to an emotion analysis device (e.g., IBM Watson Tone Analyzer) to analyze the emotional state. The input is text and audio data, and the output is the user's emotional state (e.g., "sadness").

[1256] Step 4:

[1257] The server sends text data and analyzed emotional states to a generative model (e.g., OpenAI GPT-4) to generate an appropriate response. The input is text data and emotional states, and the output is the response text (e.g., "You have a headache, take care of yourself.").

[1258] Step 5:

[1259] The server sends the generated response text to a speech synthesis system (e.g., Amazon Polly) to convert it into speech data. The input is the response text, and the output is the response speech data.

[1260] Step 6:

[1261] The device plays a voice response and responds to the user. The user can hear a voice message that says, "You have a headache, please take care of yourself."

[1262] Step 7:

[1263] The server uses a health information extraction method to extract health information from text data. The input is text data, and the output is health information (e.g., "headache," "took medicine").

[1264] Step 8:

[1265] The server accesses a cloud database (e.g., Firebase) and updates the user profile with extracted health and emotional status information. The input is health and emotional status, and the output is the updated profile information.

[1266] Step 9:

[1267] The server continuously monitors user speech using real-time monitoring methods. If an anomaly is detected (e.g., "sadness" or "headache"), a push notification system (e.g., Firebase Cloud Messaging) is used to send a notification to the configured contacts. Inputs are profile information and anomaly detection results, and output is a notification message.

[1268] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1269] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1270] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[1271] [Fourth Embodiment]

[1272] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1273] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1274] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1275] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[1276] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1277] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1278] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1279] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[1280] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1281] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1282] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1283] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1284] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1285] This invention provides a system that enables the understanding of the health status of elderly people living alone through natural conversation and allows for prompt action as needed. This system consists of speech recognition means, speech synthesis means, generative model means, health information extraction means, profile update means, and notification means. Details of each component and specific examples of systems combining them are described below.

[1286] Main processing flow

[1287] 1. Collection of audio data (device):

[1288] A voice recognition device, such as a smart speaker, is constantly on standby, waiting for the elderly person (user) to begin speaking. When the user says, "I have a headache today, so I took some medicine," the voice data is collected.

[1289] 2. Speech-to-text conversion (device):

[1290] The collected audio data is sent to a speech recognition system and converted into text data. For example, the text "I had a headache today, so I took some medicine" is generated.

[1291] 3. Response generation (server):

[1292] The converted text data is sent to the server, where a generative model analyzes it to generate an appropriate response. An example of a generated response is, "You have a headache, take care of yourself. Don't forget to take your medicine."

[1293] 4. Audio output (device):

[1294] The generated response text is converted into speech data by a speech synthesis system, and the voice is output from the smart speaker. The user hears the response, "You have a headache, take care of yourself. Don't forget to take your medicine."

[1295] 5. Extraction of health information (server):

[1296] The server extracts health-related information from the user's statements using a health information extraction method. In this example, the health information extracted includes "headache" and "took medicine."

[1297] 6. Profile update (server):

[1298] The extracted health information is updated in each user's profile. The server adds information to the profile such as "Had a headache on 2023-10-01 and took medication."

[1299] 7. Notifications and alerts (server):

[1300] If the server detects an anomaly in the profile's health information, it sends a notification to a pre-configured contact. For example, a message such as "Mr. / Ms. A complained of a headache and has taken medication" might be sent to a family member.

[1301] Specific system examples

[1302] Example: In the case of "Mr. A"

[1303] 1. Collection of audio data:

[1304] User A says, "I had a headache today, so I took some medicine."

[1305] Device: A smart speaker collects this statement.

[1306] 2. Text conversion:

[1307] Terminal: Converts the collected audio data into text data: "I had a headache today, so I took some medicine."

[1308] 3. Response generation:

[1309] Server: The generative model analyzes the text data and generates the response, "You have a headache, take care of yourself. Don't forget to take your medicine."

[1310] 4. Audio output:

[1311] Terminal: The speech synthesis device converts the generated response into speech and outputs the following message from the speaker: "You have a headache, take care of yourself. Don't forget to take your medicine."

[1312] 5. Extraction of health information:

[1313] Server: The health information extraction method extracts information such as "headache" and "took medicine" from the text data.

[1314] 6. Profile Update:

[1315] Server: Records the extracted health information in Person A's profile and updates the data to "Had a headache on 2023-10-01 and took medication."

[1316] 7. Notifications and alerts:

[1317] Server: Detects an anomaly in the health profile (in this case, "headache") and sends a notification to the designated contact (A's family) stating, "A is complaining of a headache and has taken medication."

[1318] This specific workflow allows for real-time monitoring of the health status of elderly individuals and prompt provision of necessary information to family members and care staff. This system can improve the quality of life for elderly people living alone.

[1319] The following describes the processing flow.

[1320] Step 1:

[1321] The user speaks. For example, they might say, "I had a headache today, so I took some medicine."

[1322] Step 2:

[1323] The device collects the user's speech as audio data. This is done by smart speakers and voice recognition devices.

[1324] Step 3:

[1325] The device calls a speech recognition engine to convert the collected audio data into text data. For example, it might convert the audio data into the text "I had a headache today, so I took some medicine."

[1326] Step 4:

[1327] The terminal sends the converted text data to the server. This initiates the analysis of the text data.

[1328] Step 5:

[1329] The server uses a generative model to analyze the received text data and generate an appropriate response. For example, a response such as "You have a headache, take care of yourself. Don't forget to take your medicine." might be generated.

[1330] Step 6:

[1331] The server sends the generated response text to the terminal.

[1332] Step 7:

[1333] The terminal sends the received response text to a speech synthesis engine, which converts it into speech data. The audio is then output through the speaker. Specifically, the following message is played: "You have a headache, please take care of yourself. Don't forget to take your medicine."

[1334] Step 8:

[1335] The server extracts health information from text data. For example, it might extract health information such as "headache" and "took medication."

[1336] Step 9:

[1337] The server updates the user's profile with the extracted health information. For example, the information "Had a headache on 2023-10-01 and took medication" will be recorded in person A's profile.

[1338] Step 10:

[1339] If the server detects an anomaly, it will send a notification to the designated contact. For example, a notification stating, "Mr. / Ms. A complained of a headache and has taken medication," will be sent to the user's family.

[1340] This processing flow allows for real-time monitoring of the health status of elderly individuals and enables rapid response. The sequential execution of each step ensures the safety and security of the user.

[1341] (Example 1)

[1342] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1343] The aim is to solve the problem of difficulty in monitoring and responding promptly to the daily health conditions of elderly people living alone. In particular, there is a need to provide appropriate responses in real time when elderly people report any abnormalities.

[1344] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1345] In this invention, the server includes a generation model means, a health information extraction means, a profile update means, and a notification means. This makes it possible to analyze the content of elderly people's statements, extract and record health-related information, and send notifications quickly when an abnormality is detected.

[1346] "Speech recognition means" refers to a device and program that collects user speech as audio data and converts it into text data.

[1347] "Speech synthesis means" refers to a device and program for converting text data into speech data and playing it back.

[1348] A "generative model means" is a program that analyzes collected text data and generates an appropriate response.

[1349] "Health information extraction means" refers to a device and program that extracts health-related information from text data.

[1350] "Profile update means" refers to a device and program that records and updates extracted health information in each user's profile.

[1351] "Notification means" refers to a device and program that sends a notification to a pre-configured contact when an anomaly is detected.

[1352] This invention provides a system that enables the understanding of the health status of elderly people living alone through natural conversation and enables prompt action as needed. This system consists of a speech recognition means, a speech synthesis means, a generation model means, a health information extraction means, a profile update means, and a notification means.

[1353] Hardware and software to be used

[1354] 1. Speech recognition means

[1355] Hardware: Smart speaker

[1356] Software: Speech recognition API (e.g., Google Speech-to-Text API)

[1357] 2. Speech synthesis means

[1358] Software: Text-to-speech API (e.g., Google Text-to-Speech API)

[1359] 3. Generative Model Means

[1360] Software: Generative AI models (e.g., GPT-3)

[1361] 4. Health information extraction means

[1362] Software: Natural Language Processing Algorithms

[1363] 5. Profile update method

[1364] Software: Database management systems (e.g., MySQL)

[1365] 6. Means of notification

[1366] Software: Notification API (e.g., Twilio)

[1367] Specific example: In the case of "Mr. A"

[1368] Collection of audio data

[1369] User A says, "I had a headache today, so I took some medicine."

[1370] Device: A smart speaker collects this statement.

[1371] Speech-to-text conversion

[1372] Terminal: The collected audio data is converted into text data, "I had a headache today, so I took some medicine," using a speech recognition API.

[1373] Response generation

[1374] Server: The generative model analyzes the text data and generates the response, "You have a headache, take care of yourself. Don't forget to take your medicine."

[1375] Audio output

[1376] Terminal: The response generated using a speech synthesis API is converted into speech and output from the speaker the following message: "You have a headache, take care of yourself. Don't forget to take your medicine."

[1377] Extraction of health information

[1378] Server: The health information extraction method extracts information such as "headache" and "took medicine" from the text data.

[1379] Profile Update

[1380] Server: Records the extracted health information in Person A's profile and updates the data to "Had a headache on 2023-10-01 and took medication."

[1381] Notifications and alerts

[1382] Server: Detects an anomaly in the health profile (in this case, "headache") and sends a notification to the configured contacts stating, "Person A complained of a headache and has taken medication."

[1383] Example of a prompt

[1384] An elderly person says, "I had a headache today, so I took some medicine."

[1385] The server uses a generated AI model to respond, "You have a headache, take care of yourself. Don't forget to take your medicine."

[1386] This system allows for real-time monitoring of the health status of elderly people living alone and promptly provides necessary information to family members and care staff. This, in turn, can improve the quality of life for elderly people living alone.

[1387] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1388] Step 1:

[1389] Collection of audio data (device)

[1390] Input: The user speaks into the smart speaker.

[1391] Action: The user says, "I have a headache today, so I took some medicine." The smart speaker, acting as the device, uses its microphone to collect this statement.

[1392] Data processing: The microphone converts the audio signal into a digital signal and temporarily stores it in a buffer.

[1393] Output: Digital audio data.

[1394] Step 2:

[1395] Speech-to-text conversion (device)

[1396] Input: Digital audio data.

[1397] Operation: The device sends the collected audio data to a speech recognition API (e.g., Google Speech-to-Text API).

[1398] Data processing: The speech recognition API analyzes the audio data and identifies phonemes and context. This converts the audio data into text data: "I had a headache today, so I took some medicine."

[1399] Output: Text data "I had a headache today, so I took some medicine."

[1400] Step 3:

[1401] Response generation (server)

[1402] Input: Text data "I had a headache today, so I took some medicine."

[1403] Operation: The server inputs the received text data into a generating AI model (e.g., GPT-3) and generates an appropriate response.

[1404] Data processing: The generative AI model analyzes text data and generates a response based on the context. Specifically, it generates a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine."

[1405] Output: Response text "You have a headache, please take care of yourself. Don't forget to take your medicine."

[1406] Step 4:

[1407] Audio output (device)

[1408] Input: Response text: "You have a headache, please take care of yourself. Don't forget to take your medicine."

[1409] Operation: The device sends the response text to a text-to-speech API (e.g., Google Text-to-Speech API) and converts it into speech data.

[1410] Data processing: The speech synthesis API analyzes the text data and generates an audio waveform. This audio waveform is then prepared as audio data.

[1411] Output: Response voice data.

[1412] Step 5:

[1413] Extraction of health information (server)

[1414] Input: Text data "I had a headache today, so I took some medicine."

[1415] Operation: The server uses a health information extraction method to identify health-related information from text data.

[1416] Data processing: Analyze the text using natural language processing algorithms and extract health-related keywords such as "headache" and "took medicine."

[1417] Output: Health information "headache" and "took medication".

[1418] Step 6:

[1419] Profile update (server)

[1420] Input: Health information "headache" and "took medication".

[1421] Operation: The server records and updates the extracted health information in the user profile.

[1422] Data processing: Access the user profile database and add the information, "Had a headache on 2023-10-01 and took medication."

[1423] Output: Updated user profile.

[1424] Step 7:

[1425] Notifications and alerts (server)

[1426] Input: Updated user profile.

[1427] Operation: The server detects anomalies from the health profile and sends notifications to the configured contacts.

[1428] Data processing: Anomaly detection algorithms are used to extract anomalous data that matches the specified conditions. Specifically, a notification is generated stating, "Person A complained of a headache and took medication."

[1429] Output: A notification message sent to the contact.

[1430] This system's processing flow makes it possible to monitor the health status of elderly individuals in real time and quickly provide necessary information to their families and care staff.

[1431] (Application Example 1)

[1432] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1433] For elderly people living alone, monitoring their health and ensuring a rapid response in emergencies are major challenges. However, conventional systems have struggled to naturally monitor health conditions through voice conversations and to quickly detect and respond to abnormalities. In particular, the lack of adequate means for appropriate responses in emergencies and for promptly notifying family members and medical professionals has made ensuring the safety of the elderly a significant concern.

[1434] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1435] In this invention, the server includes speech recognition means, speech synthesis means, generation model means, health information extraction means, profile update means, notification means, speech response generation means, and health information transmission means. This makes it possible to analyze the voice spoken by elderly people on a daily basis in real time and constantly monitor their health status. Furthermore, when an abnormality is detected, it is possible to quickly generate an appropriate response, enabling emergency response and immediate notification to family members and medical professionals through remote transmission of health information.

[1436] "Speech recognition means" refers to a technological element that collects speech from elderly people as audio data and converts that audio data into text data.

[1437] "Speech synthesis means" refers to a technological element that converts generated text responses into speech data and transmits it to elderly people.

[1438] A "generative model means" is a technical element that analyzes text data of everyday conversations of elderly people and generates appropriate responses.

[1439] A "health information extraction method" is a technological element that extracts health-related information from everyday conversations.

[1440] A "profile update method" is a technological element that records and updates extracted health information in each user's profile.

[1441] A "notification method" is a technical element that sends a notification to a designated contact when an anomaly is detected.

[1442] A "voice response generation means" is a technological element that automatically generates an appropriate response in the event of an abnormality, thereby addressing the needs of the elderly.

[1443] A "health information transmission means" is a technical element that transmits extracted health information to a remote server.

[1444] The system of this invention provides support for elderly people to live safely and healthily in their homes or care facilities. This system is composed of the following hardware and software.

[1445] Hardware configuration

[1446] 1. Voice recognition device: Use a microphone built into the robot (e.g., USB microphone, built-in microphone of a smart speaker). This will pick up the speech of elderly people.

[1447] 2. Speech synthesis device: Uses the robot's built-in speaker to output the generated response as speech.

[1448] 3. Computing Unit: A server or high-performance computer that manages the entire system.

[1449] Software Configuration

[1450] 1. Speech recognition software: The speech_recognition library in Python is used to convert the speech of elderly people into text data.

[1451] 2. Speech synthesis software: Use the pyttsx3 library in Python to convert text responses into speech.

[1452] 3. Generative Models: AI models that use natural language processing (NLP) to analyze text data and generate appropriate responses. For example, OpenAI's GPT-3 is used.

[1453] 4. Communication method: Use the Python requests library to send health information to a remote server.

[1454] Explanation of the process

[1455] The server first receives audio data transmitted from the speech recognition device and converts it into text data. This process is performed using the speech_recognition library. Next, the generative model analyzes the text data and generates an appropriate response from what the elderly person is saying. The generated response is converted into audio data using the pyttsx3 library and output by the speech synthesis device.

[1456] Furthermore, health-related information is extracted from the generated text data using a health information extraction method. This extracted information is then updated in the user's profile by the server. If an abnormality is detected, a notification is sent to the contacts set by the notification method. In addition, by sending the extracted health information to a remote server using the health information transmission method, the health status of elderly individuals can be continuously monitored.

[1457] Explanation of specific examples

[1458] For example, if person A says, "I have a headache today, so I took some medicine," the system collects this utterance as audio data and converts it into text data. Next, a generative model analyzes this and generates a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine." This response is then spoken by a speech synthesis system and conveyed to person A. Furthermore, the information "headache" and "took medicine" is extracted by a health information extraction system and updated in person A's profile. If an abnormality is detected, person A's family is notified by email.

[1459] Example of a prompt

[1460] "In the following conversation, please provide an explanation automatically. If the user says, 'I have a headache today, so I took some medicine,' the robot should respond, 'You have a headache, please take care of yourself. Don't forget to take your medicine,' and send the health information to the server."

[1461] This will allow for continuous monitoring of the health status of elderly individuals and enable a swift response when abnormalities occur.

[1462] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1463] Step 1:

[1464] The device collects speech from elderly individuals as audio data through a speech recognition device (microphone). The input is the elderly person's voice, and the output is that audio data. This audio data is stored in internal memory for speech recognition processing.

[1465] Step 2:

[1466] The terminal uses the Python `speech_recognition` library to convert collected audio data into text data. The input is audio data, and the output is text data. This conversion process is performed using a speech recognition engine, and the text data will contain content such as "I had a headache today, so I took some medicine."

[1467] Step 3:

[1468] The server analyzes the transformed text data using a generative model (e.g., GPT-3) and generates an appropriate response. The input is text data, and the output is the response text. The generative model understands the context based on the input text data and outputs a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine."

[1469] Step 4:

[1470] The terminal uses a speech synthesis system (pyttsx3 library) to generate response text, which is then converted into audio data and output to the elderly person through a speaker. The input is response text, and the output is audio data. This audio data is played back through the speaker so that the elderly person can hear it.

[1471] Step 5:

[1472] The server uses a health information extraction method to extract health-related information from text data of elderly people's daily conversations. The input is text data of conversations, and the output is health information (e.g., "headache" and "took medicine"). This extraction process is performed using natural language processing techniques.

[1473] Step 6:

[1474] The server updates the user profile with extracted health information using a profile update mechanism. The input is health information, and the output is the updated profile data. This update process is carried out by adding the new health information to the user-specific profile database.

[1475] Step 7:

[1476] The server uses a notification system to send notifications to configured contacts when an anomaly is detected. Inputs are profile data and anomaly detection information, and output is a notification message. This notification is sent via methods such as email or SMS.

[1477] Step 8:

[1478] The server uses a health information transmission method to send extracted health information to a remote server. The input is health information, and the output is the transmitted data. This transmission process is carried out, for example, using an HTTP request, and the data is stored on a remote management server.

[1479] Through these steps, it becomes possible to monitor the health status of elderly individuals in real time and take necessary actions quickly.

[1480] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1481] This invention relates to a system that uses a speech recognition system combined with an emotion engine to grasp the emotions and health status of elderly people in real time and to take prompt action as needed. This system consists of a speech recognition means, a speech synthesis means, a generation model means, a health information extraction means, a profile update means, a notification means, and an emotion engine. The details of each component and a specific example of a system combining them are described below.

[1482] Main processing flow

[1483] 1. Collection of audio data (device):

[1484] A voice recognition device, such as a smart speaker, is constantly on standby, waiting for the elderly person (user) to begin speaking. When the user says, "I have a headache today, so I took some medicine," the voice data is collected.

[1485] 2. Speech-to-text conversion (device):

[1486] The collected audio data is sent to a speech recognition system and converted into text data. For example, the text "I had a headache today, so I took some medicine" is generated.

[1487] 3. Sentiment analysis (server):

[1488] Text and audio data are sent to the emotion engine, which analyzes the user's emotional state. For example, emotions such as "sadness" and "anxiety" may be detected.

[1489] 4. Response generation (server):

[1490] The converted text data and analyzed emotional state are sent to a generative model to generate an appropriate response. Based on the emotional state, a response such as "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone." is generated.

[1491] 5. Audio output (device):

[1492] The generated response text is converted into speech data by a speech synthesis system, and the voice is output from the smart speaker. The user hears the response, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[1493] 6. Extraction of health information (server):

[1494] The server extracts health-related information from the user's statements using a health information extraction method. In this example, the health information extracted includes "headache" and "took medicine."

[1495] 7. Profile update (server):

[1496] The extracted health information and emotional state are updated in each user's profile. The server adds the following information to the profile: "Had a headache on 2023-10-01 and took medication. Emotional state: Sad."

[1497] 8. Notifications and alerts (server):

[1498] If the server detects an anomaly in the profile's health information, it sends a notification to a pre-configured contact. For example, a message such as, "Person A complained of a headache and took medication. Their emotional state was analyzed as sadness," might be sent to family members.

[1499] Specific system examples

[1500] Example: In the case of "Mr. A"

[1501] 1. Collection of audio data:

[1502] User A says, "I had a headache today, so I took some medicine."

[1503] Device: A smart speaker collects this statement.

[1504] 2. Text conversion:

[1505] Terminal: Converts the collected audio data into text data: "I had a headache today, so I took some medicine."

[1506] 3. Emotion analysis:

[1507] Server: Text and audio data are sent to the emotion engine, and the emotion "sadness" is analyzed.

[1508] 4. Response generation:

[1509] Server: The generative model analyzes the text data and emotional state to generate a response such as, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[1510] 5. Audio output:

[1511] Terminal: The speech synthesis system converts the generated response into speech and outputs the following message from the speaker: "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[1512] 6. Extraction of health information:

[1513] Server: The health information extraction method extracts information such as "headache" and "took medicine" from the text data.

[1514] 7. Profile Update:

[1515] Server: Records the extracted health information and the emotional state of "sadness" in Person A's profile and updates the data to "Had a headache on 2023-10-01 and took medicine. Emotional state: sadness."

[1516] 8. Notifications and alerts:

[1517] Server: Detects anomalies in the health profile (in this case, the emotional states of "headache" and "sadness") and sends a notification to the designated contact (A's family) stating, "A complained of a headache and took medication. Their emotional state was also analyzed as sadness."

[1518] This specific workflow allows for real-time monitoring of the health and emotional state of elderly individuals, enabling prompt responses. By combining it with an emotional intelligence engine, more accurate monitoring and psychological care can be achieved.

[1519] The following describes the processing flow.

[1520] Step 1:

[1521] The user speaks. For example, they might say, "I had a headache today, so I took some medicine."

[1522] Step 2:

[1523] The device collects the user's speech as audio data. This collection is performed by smart speakers or voice recognition devices.

[1524] Step 3:

[1525] The device sends the collected voice data to a speech recognition engine, which converts it into text data. Specifically, the voice data is converted into the string "I had a headache today, so I took some medicine."

[1526] Step 4:

[1527] The terminal sends the converted text data to the server. This data transmission initiates the analysis of the text data.

[1528] Step 5:

[1529] The server sends the received text and audio data to the emotion engine, which then analyzes the user's emotional state. For example, emotions such as "sadness" can be detected from the content of speech and tone of voice.

[1530] Step 6:

[1531] The server sends the output of the emotion engine and text data to the generation model to generate an appropriate response. Based on the emotional state, the generated response will be, "You have a headache, take care of yourself. If you have any concerns, please talk to someone."

[1532] Step 7:

[1533] The server sends the generated response text to the terminal. The terminal then processes the response.

[1534] Step 8:

[1535] The device sends the received response text to a speech synthesis engine, which converts it into speech data. Then, the smart speaker outputs the following voice message: "You have a headache, please take care of yourself. If you have any concerns, please talk to someone."

[1536] Step 9:

[1537] The server extracts health information from text data and emotional states. Specifically, it extracts information such as "headache" and "took medicine," as well as the emotional state of "sadness."

[1538] Step 10:

[1539] The server updates the user's profile with the extracted health information and emotional state. The following data is added to Person A's profile: "Had a headache on 2023-10-01 and took medication. Emotional state: Sad."

[1540] Step 11:

[1541] If the server detects an anomaly, it will send a notification to the designated contacts. Because the emotional state "sadness" was detected, the server will send a notification to Person A's family stating, "Person A complained of a headache and took medication. Also, their emotional state was analyzed as sadness."

[1542] By processing the data step by step in this way, it becomes possible to grasp the health and emotional state of elderly individuals in real time and respond quickly. This system is expected to further enhance the physical and mental care of the elderly and improve their quality of life.

[1543] (Example 2)

[1544] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1545] Monitoring the health and emotional state of elderly individuals in real time and responding promptly is a challenging task. Especially with the increasing number of elderly people living alone, there is a need to accurately monitor their daily health and emotional changes and provide appropriate support when needed. Furthermore, given the limited means available to medical institutions and families to monitor the condition of elderly individuals, a system is needed that allows for monitoring their health remotely.

[1546] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1547] In this invention, the server includes speech recognition means, speech synthesis means, generation model means, emotion analysis means, health information extraction means, profile update means, and notification means. This allows for real-time monitoring of the health and emotional state of elderly individuals, even from a remote location, by analyzing their speech content to determine their health and emotional state and generating and outputting appropriate responses. Furthermore, it enables rapid notification in the event of an abnormality, prompting appropriate action.

[1548] "Speech recognition means" refers to a device or software for converting speech data into text data.

[1549] "Speech synthesis means" refers to a device or software for converting text data into speech data.

[1550] "Generative model means" refers to an AI model that generates an appropriate response based on input data.

[1551] "Emotional analysis means" refers to technology for analyzing and identifying a user's emotional state from voice data and text data.

[1552] A "health information extraction means" is a program or device for extracting health-related information from a user's statements or actions.

[1553] A "profile update method" is a system that records and updates extracted health information and emotional states in the user profile.

[1554] A "notification means" is a device or software that sends a notification to a designated contact when an anomaly is detected.

[1555] This invention provides a system that uses a speech recognition system combined with emotion analysis to monitor the health and emotional state of elderly individuals in real time and enable prompt responses as needed. This system consists of speech recognition means, speech synthesis means, generative model means, emotion analysis means, health information extraction means, profile update means, and notification means.

[1556] Specific examples of hardware and software to be used

[1557] hardware

[1558] Smart speakers: Used as devices to collect everyday conversations of the elderly. Specific examples include Amazon Echo and Google Nest speakers.

[1559] Server: A computer server used for data analysis and processing. Cloud-based servers (e.g., AWS, Google Cloud) can also be used.

[1560] software

[1561] Speech recognition software: Used to convert speech data into text data. Specific examples include Google Speech-to-Text and Microsoft Azure Speech.

[1562] Text-to-speech software: Used to convert text data into speech data. Specific examples include Amazon Polly and Google Text-to-Speech.

[1563] Sentiment analysis engine: Used to analyze emotions from text and audio data. Specific examples include IBM Watson Tone Analyzer and Amazon Comprehend.

[1564] Generative AI models: Used to generate appropriate responses based on user input. A specific example is OpenAI GPT-4.

[1565] Database system: Used to record and update health information and emotional status in user profiles. Specific examples include MySQL and MongoDB.

[1566] Notification system: Used to send notifications when an anomaly is detected. Specific examples include Twilio and SMTP servers.

[1567] Specific details of the actions

[1568] Collection of audio data

[1569] The user speaks to the smart speaker, saying, "I have a headache today, so I took some medicine." The smart speaker collects this voice data.

[1570] Speech-to-text conversion

[1571] The device uses voice recognition to convert the voice data into text data that reads, "I had a headache today, so I took some medicine."

[1572] Emotion analysis

[1573] The server sends text and audio data to an emotion analysis engine to analyze emotions such as "sadness" and "anxiety."

[1574] Response generation

[1575] The server generates an appropriate response using a generative model based on the analyzed emotional state. The generated response text will read something like, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[1576] Audio output

[1577] The terminal converts the generated response into voice data using speech synthesis technology and outputs it to the user via the smart speaker.

[1578] Extraction of health information

[1579] The server uses a health information extraction method to extract health information such as "headache" and "took medicine" from the user's speech.

[1580] Profile update

[1581] The server records the extracted health information and emotional state in the user profile. For example, it adds data such as, "Had a headache on 2023-10-01 and took medication. Emotional state: Sad."

[1582] Notifications and alerts

[1583] If an anomaly is detected, the server will send a notification to the designated contact. The message will include something like, "Person A complained of a headache and took medication. Their emotional state was analyzed as sadness."

[1584] Examples of prompt statements

[1585] If someone says, "I have a headache today, so I took some medicine," a specific example of a response like, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone," is generated and output by the speech synthesis system.

[1586] By combining the above elements, the present invention makes it possible to monitor the health and emotional state of elderly people in real time and to take prompt and appropriate action when necessary.

[1587] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1588] Step 1:

[1589] Collection of audio data

[1590] User: An elderly person says, "I have a headache today, so I took some medicine."

[1591] Specific action: An elderly person speaks into a smart speaker.

[1592] Input: User's voice.

[1593] Output: Audio data captured by the smart speaker.

[1594] Step 2:

[1595] Speech-to-text conversion

[1596] Terminal: The smart speaker collects voice data and converts it into text data using speech recognition technology.

[1597] Specific operation: Speech recognition software (e.g., Google Speech-to-Text API) analyzes the audio data and generates a corresponding text string.

[1598] Input: Audio data.

[1599] Output: Text data "I had a headache today, so I took some medicine."

[1600] Step 3:

[1601] Emotion analysis

[1602] Server: The converted text and audio data are sent to the sentiment analysis engine.

[1603] Specific operation: A sentiment analysis engine on the server (e.g., IBM Watson Tone Analyzer) analyzes text data and voice tone to identify the user's emotional state.

[1604] Input: Text data and audio data.

[1605] Output: Emotional analysis results such as "sadness" and "anxiety".

[1606] Step 4:

[1607] Response generation

[1608] Server: Generates an appropriate response using a generative model based on the analyzed emotional state.

[1609] Specific operation: Based on the text input as a prompt and sentiment data, a generative AI model (e.g., OpenAI GPT-4) generates an appropriate response.

[1610] Input: Sentiment analysis results and text data.

[1611] Output: A text response saying, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[1612] Step 5:

[1613] Audio output

[1614] Terminal: The generated response text is converted into speech data using speech synthesis means.

[1615] Specific operation: Text-to-speech software (e.g., Amazon Polly) converts the response text into speech data.

[1616] Input: Text response data.

[1617] Output: Audio data of the response.

[1618] Device: Outputs voice messages to the user from the smart speaker.

[1619] Specific operation: Synthesized audio is played from the speaker and heard by the user.

[1620] Input: Audio data of the response.

[1621] Output: "You have a headache, please take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone." (Voice)

[1622] Step 6:

[1623] Extraction of health information

[1624] Server: Using a health information extraction method, it extracts health information such as "headache" or "took medicine" from the user's statements.

[1625] Specific operation: A text analysis engine (e.g., SpaCy or NLTK) identifies and extracts medical-related keywords from the text.

[1626] Input: Text data.

[1627] Output: Health information such as "headache" or "took medicine".

[1628] Step 7:

[1629] Profile update

[1630] Server: Records extracted health information and emotional state in each user's profile.

[1631] Specific action: The database system (e.g., MySQL or MongoDB) adds the information "Had a headache on 2023-10-01 and took medication. Emotional state: Sad" to the user profile.

[1632] Input: Health information and emotional state.

[1633] Output: Updated user profile.

[1634] Step 8:

[1635] Notifications and alerts

[1636] Server: If an anomaly is detected, it will send a notification to the configured contact.

[1637] Specific action: The notification system (e.g., Twilio or an SMTP server) sends a message to the configured contact stating, "Person A complained of a headache and took medication. Their emotional state was analyzed as sad."

[1638] Input: Health information and emotional state in which abnormalities were detected.

[1639] Output: Sent notification message.

[1640] (Application Example 2)

[1641] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1642] It is extremely important to understand the health and emotional state of the elderly in real time and to respond quickly and appropriately. However, conventional systems lacked the ability to integrate emotion analysis, autonomous response generation, and real-time health monitoring and notification, leaving them insufficient. In particular, the risk of delayed response in emergencies is high for elderly people living alone, so a new system is needed to solve this problem.

[1643] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes a speech recognition means, a speech synthesis means, a generation model means, an emotion analysis means, a health information extraction means, a profile update means, a real-time monitoring means, a push notification system, and a cloud database. This makes it possible to collect voice data from elderly people, perform emotion analysis and extract health information in real time, generate appropriate responses based on the content of speech, and send notifications quickly in emergencies.

[1644] "Speech recognition means" refers to technology that collects speech data and converts that speech into text.

[1645] "Speech synthesis means" refers to a technology that converts text data into speech data and outputs it as speech.

[1646] "Generative model means" refers to AI model technology that generates appropriate responses based on input data.

[1647] A "health information extraction method" is a technology that extracts health-related information from everyday conversations.

[1648] "Profile update means" refers to a technology that adds or updates extracted health information to the user's individual profile.

[1649] "Notification means" refers to a technology that sends notifications to designated contacts when an anomaly is detected.

[1650] "Emotional analysis methods" refer to technologies that analyze a user's emotional state from voice and text data.

[1651] "Real-time monitoring means" refers to technology that collects, analyzes, and processes the speech of elderly people in real time.

[1652] An "autonomous response generation means" is a technology that generates an appropriate response based on the results of an analysis of emotional states and other factors.

[1653] A "cloud database" is an online database that stores and manages data via the internet.

[1654] A "push notification system" is a technology that automatically sends important information and warnings to recipients in real time.

[1655] To implement this invention, a system is needed to collect and analyze voice data of elderly people in real time. This system consists of voice recognition means, voice synthesis means, generative model means, emotion analysis means, health information extraction means, profile update means, real-time monitoring means, push notification system, and cloud database.

[1656] First, the user speaks through a device such as a smartphone or a pet-like robot. The device's speech recognition system collects this speech data and converts it into text data using a speech recognition engine such as Google Cloud Speech-to-Text. For example, if the user says, "I have a headache today, so I took some medicine," that statement is converted into the text data "I have a headache today, so I took some medicine."

[1657] Next, the converted text data is analyzed using sentiment analysis tools such as IBM Watson Tone Analyzer to clarify the user's emotional state. As a result of the analysis, emotions such as "sadness" or "anxiety" may be detected.

[1658] The converted text data and analyzed emotional state are sent to a generative AI model such as OpenAI's GPT-4 to generate an appropriate response. The generated response consists of kind words directed at the user, such as, "You have a headache, take care of yourself. Don't forget to take your medicine. If you have any concerns, please talk to someone."

[1659] The generated responses are converted into audio data via a speech synthesis engine such as Amazon Polly and played back through a smartphone or robot's speaker. Users can gain a sense of reassurance by listening to the generated audio.

[1660] Furthermore, the server extracts health-related information from text data using a health information extraction mechanism. For example, specific health information such as "headache" and "took medicine" is extracted. The extracted health information and emotional state are then updated in the user profile stored in a cloud database (e.g., Firebase).

[1661] The real-time monitoring system constantly monitors user speech and, if an anomaly is detected, immediately sends a notification to a designated contact (family or medical staff) using a push notification system (e.g., Firebase Cloud Messaging). This notification might contain a message such as, "Person A complained of a headache and took medication. Their emotional state was analyzed as sadness."

[1662] As a concrete example, here is an example of a prompt message:

[1663] User comment: "I had a headache today, so I took some medicine."

[1664] Detected emotion: "sadness"

[1665] Generate an appropriate response.

[1666] As described above, the system for implementing this invention can grasp the health and emotional state of elderly people in real time and respond quickly. This makes it possible to ensure the safety and security of elderly people living alone.

[1667] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1668] Step 1:

[1669] The user speaks. The user uses a device such as a smartphone or a pet-like robot to say something like, "I have a headache today, so I took some medicine." This voice data becomes the system's input.

[1670] Step 2:

[1671] The device collects audio data. Using speech recognition technology (e.g., Google Cloud Speech-to-Text), the collected audio data is converted into text data in real time. Here, the input is audio data, and the output is text data.

[1672] Step 3:

[1673] The server sends text and audio data to an emotion analysis device (e.g., IBM Watson Tone Analyzer) to analyze the emotional state. The input is text and audio data, and the output is the user's emotional state (e.g., "sadness").

[1674] Step 4:

[1675] The server sends text data and analyzed emotional states to a generative model (e.g., OpenAI GPT-4) to generate an appropriate response. The input is text data and emotional states, and the output is the response text (e.g., "You have a headache, take care of yourself.").

[1676] Step 5:

[1677] The server sends the generated response text to a speech synthesis system (e.g., Amazon Polly) to convert it into speech data. The input is the response text, and the output is the response speech data.

[1678] Step 6:

[1679] The device plays a voice response and responds to the user. The user can hear a voice message that says, "You have a headache, please take care of yourself."

[1680] Step 7:

[1681] The server uses a health information extraction method to extract health information from text data. The input is text data, and the output is health information (e.g., "headache," "took medicine").

[1682] Step 8:

[1683] The server accesses a cloud database (e.g., Firebase) and updates the user profile with extracted health and emotional status information. The input is health and emotional status, and the output is the updated profile information.

[1684] Step 9:

[1685] The server continuously monitors user speech using real-time monitoring methods. If an anomaly is detected (e.g., "sadness" or "headache"), a push notification system (e.g., Firebase Cloud Messaging) is used to send a notification to the configured contacts. Inputs are profile information and anomaly detection results, and output is a notification message.

[1686] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1687] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1688] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[1689] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1690] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[1691] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[1692] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[1693] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[1694] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[1695] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[1696] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[1697] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[1698] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[1699] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1700] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[1701] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[1702] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[1703] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[1704] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[1705] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[1706] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted as being incorporated by reference.

[1707] The following is further disclosed regarding the embodiments described above.

[1708] (Claim 1)

[1709] Voice recognition means and

[1710] A speech synthesis method,

[1711] Generative model means,

[1712] A means for extracting health information,

[1713] Profile update method,

[1714] Notification methods,

[1715] A system that includes this.

[1716] (Claim 2)

[1717] A speech recognition method that collects voice data from elderly people and converts it into text data,

[1718] A generative model means that analyzes the converted text data to generate a response,

[1719] A speech synthesis means that converts the generated response into speech and outputs it,

[1720] A health information extraction method that extracts health information from the daily conversations of elderly people,

[1721] A profile update means for updating the user profile with extracted health information,

[1722] A notification method that sends a notification to a designated contact when an anomaly is detected,

[1723] The system according to claim 1, including the following:

[1724] (Claim 3)

[1725] The system according to claim 1, wherein the voice data collected by the voice recognition means is performed with the permission of the elderly person.

[1726] "Example 1"

[1727] (Claim 1)

[1728] Voice recognition means and

[1729] A speech synthesis method,

[1730] Generative model means,

[1731] A means for extracting health information,

[1732] Profile update method,

[1733] Notification methods,

[1734] A speech synthesis means that converts the generated response text into speech data,

[1735] A health information extraction method that analyzes the content of user statements and extracts health-related information,

[1736] A profile update means that records extracted health information in each user's profile,

[1737] A notification method that sends a notification when an abnormality is detected from the profile's health information,

[1738] A system that includes this.

[1739] (Claim 2)

[1740] A speech recognition method that collects voice data from elderly people and converts it into text data,

[1741] A generative model means that analyzes the converted text data to generate a response,

[1742] A speech synthesis means that converts the generated response into speech and outputs it,

[1743] A health information extraction method that extracts health information from the daily conversations of elderly people,

[1744] A profile update means for updating the user profile with extracted health information,

[1745] A notification method that sends a notification to a designated contact when an anomaly is detected,

[1746] The system according to claim 1, including the following:

[1747] (Claim 3)

[1748] The system according to claim 1, wherein the voice data collected by the voice recognition means is performed with the permission of the elderly person.

[1749] "Application Example 1"

[1750] (Claim 1)

[1751] Voice recognition means and

[1752] A speech synthesis method,

[1753] Generative model means,

[1754] A means for extracting health information,

[1755] Profile update method,

[1756] Notification methods,

[1757] Voice response generation means,

[1758] means of transmitting health information,

[1759] A system that includes this.

[1760] (Claim 2)

[1761] A speech recognition method that collects voice data from elderly people and converts it into text data,

[1762] A generative model means that analyzes the converted text data to generate a response,

[1763] A speech synthesis means that converts the generated response into speech and outputs it,

[1764] A health information extraction method that extracts health information from the daily conversations of elderly people,

[1765] A profile update means for updating the user profile with extracted health information,

[1766] A notification method that sends a notification to a designated contact when an anomaly is detected,

[1767] A voice response generation means for providing an emergency response in the event of an abnormality,

[1768] A means for transmitting health information to a remote server,

[1769] The system according to claim 1, including the following:

[1770] (Claim 3)

[1771] The system according to claim 1, wherein the voice data collected by the voice recognition means is performed with the permission of the elderly person.

[1772] "Example 2 of combining an emotion engine"

[1773] (Claim 1)

[1774] Voice recognition means and

[1775] A speech synthesis method,

[1776] Generative model means,

[1777] Emotion analysis methods,

[1778] A means for extracting health information,

[1779] Profile update method,

[1780] Notification methods,

[1781] A system that includes this.

[1782] (Claim 2)

[1783] A speech recognition method that collects voice data from elderly people and converts it into text data,

[1784] An emotion analysis means for analyzing emotional states using converted text data and audio data,

[1785] A generative model means that generates an appropriate response based on the analyzed emotional state,

[1786] A speech synthesis means that converts the generated response into speech and outputs it,

[1787] A health information extraction method that extracts health information from the daily conversations of elderly people,

[1788] A profile update means for updating the user profile with extracted health information and emotional state,

[1789] A notification method that sends a notification to a designated contact when an anomaly is detected,

[1790] The system according to claim 1, including the following:

[1791] (Claim 3)

[1792] The system according to claim 1, wherein the voice data collected by the voice recognition means is performed with the permission of the elderly person.

[1793] "Application example 2 when combining with an emotional engine"

[1794] (Claim 1)

[1795] Voice recognition means and

[1796] A speech synthesis method,

[1797] Generative model means,

[1798] A means for extracting health information,

[1799] Profile update method,

[1800] Notification methods,

[1801] Emotion analysis methods,

[1802] Real-time monitoring means,

[1803] Autonomous response generation means,

[1804] Cloud databases and

[1805] Push notification system and

[1806] A system that includes this.

[1807] (Claim 2)

[1808] A speech recognition method that collects voice data from elderly people and converts it into text data,

[1809] An emotion analysis means for analyzing converted text data and audio data to detect emotional states,

[1810] An autonomous response generation means that generates a response based on the analyzed emotional state,

[1811] A speech synthesis means that converts the generated response into speech and outputs it,

[1812] A health information extraction method that extracts health information from the daily conversations of elderly people,

[1813] A profile update means for updating the user profile with extracted health information,

[1814] A push notification system that sends notifications to designated contacts when an anomaly is detected,

[1815] A real-time monitoring method for monitoring user speech in real time,

[1816] The system according to claim 1, including the following:

[1817] (Claim 3)

[1818] The system according to claim 1, wherein the voice data collected by the voice recognition means is performed with the permission of the elderly person. [Explanation of Symbols]

[1819] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A speech recognition method that collects voice data from elderly people and converts it into text data, A generative model means that analyzes the converted text data to generate a response, A speech synthesis means that converts the generated response into speech and outputs it, A health information extraction method that extracts health information from the daily conversations of elderly people, A profile update means for updating the user profile with extracted health information, A notification method that sends a notification to a designated contact when an anomaly is detected, A system that includes this.

2. The system according to claim 1, wherein the voice data collected by the voice recognition means is performed with the permission of the elderly person.

Images