system

Abstract

We provide the system. [Solution] A means for monitoring the mental health status of employees, A method for analyzing employee communication data using natural language processing technology, A means for calculating an emotional score and a stress score based on the analysis results, A means of notifying the administrator and the employee themselves when the stress score exceeds a set threshold, Means of providing specific mental care suggestions, 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 method for controlling a persona chatbot, which is performed by at least one processor, 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] In a remote work environment, the opportunity for employees to consult their managers about their mental health problems decreases, and they often have to deal with problems alone. Also, it is difficult for managers to grasp the mental state of employees in a remote environment, and it is difficult to provide appropriate support. Such a situation may cause a decrease in employees' productivity and health damage.

Means for Solving the Problems

[0005] To solve the above problems, the present invention provides a system that includes means for monitoring the mental health status of employees, means for analyzing employee communication data using natural language processing technology, means for calculating an emotion score and a stress score based on the analysis results, means for notifying managers and the employee themselves when the stress score exceeds a set threshold, and means for providing specific mental care suggestions. This system makes it possible to grasp the mental health of employees in real time and take appropriate action, even when they are working remotely.

[0006] 1. "Mental health status" refers to the mental and emotional health status of an employee.

[0007] 2. “Monitoring means” refers to equipment or methods for continuously observing the condition of employees and collecting data.

[0008] 3. "Natural language processing technology" refers to computer technology used to analyze text and audio data to extract emotions and meanings.

[0009] 4. "Communication data" refers to all communication data used by employees for work purposes, including emails, chats, meeting recordings, and voice memos.

[0010] 5. An "emotion score" is a numerical representation of the degree and type of emotion obtained from data analyzed using natural language processing technology.

[0011] 6. A "stress score" is a numerical representation of the recipient's stress level, derived from data analyzed using natural language processing technology.

[0012] 7. A "threshold" refers to a set threshold value that is used to trigger a specific action.

[0013] 8. "Means of notification" refers to methods or devices for informing administrators and individuals of detected anomalies or important information.

[0014] 9. "Mental care suggestions" refer to specific actions or behaviors aimed at improving or maintaining employees' mental health. [Brief explanation of the drawing]

[0015] [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]It is a sequence diagram showing the processing flow of a data processing system in Application Example 2 when a sentiment engine is combined.

Embodiments for Carrying Out the Invention

[0016] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0017] First, the terms used in the following description will be explained.

[0018] In the following embodiments, a numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple 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), etc.

[0019] In the following embodiments, a numbered RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0020] In the following embodiments, a 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, etc.

[0021] 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).

[0022] 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."

[0023] [First Embodiment]

[0024] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0025] 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.

[0026] 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).

[0027] 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.

[0028] 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.

[0029] 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.

[0030] 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.

[0031] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0032] 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.

[0033] 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.

[0034] 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.

[0035] 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".

[0036] The system of this invention monitors the mental health of employees working remotely in real time and detects alerts for stress levels. Specific embodiments are described below.

[0037] Data collection

[0038] 1. Terminal: Employees log in to the system using a work-issued PC or mobile device. Based on the login information, the system prepares to collect data.

[0039] 2. Terminal: A script starts running to collect communication data such as employee emails, chats, meeting recordings, and voice memos. The collected data is saved to a temporary file.

[0040] 3. Terminal: Communication data stored in temporary files is encrypted and transferred to the server using a secure protocol.

[0041] Data Analysis

[0042] 1. Server: Receives the transferred data and stores it in a secure database. The stored data is pre-processed (e.g., noise reduction, formatting standardization).

[0043] 2. Server: Applies natural language processing algorithms to perform sentiment analysis on text and audio data. The analysis includes emotional tone, frequency of positive and negative language use, and specific stress signs (e.g., frequent use of expressions like "tired" or "irritated").

[0044] 3. Server: Calculates emotional and stress scores to assess the mental health status of each employee.

[0045] Alert generation

[0046] 1. Server: Compares the emotional score and stress score to a threshold and generates an alert if the stress score exceeds the threshold. The alert includes specific care suggestions or suggestions for taking a break.

[0047] 2. Server: Records generated alerts in the logging system for future analysis and feedback.

[0048] Notifications and suggestions

[0049] 1. Server: Sends alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[0050] 2. Devices: Notifications are displayed on employees' PCs and mobile devices, along with a message such as, "Your stress level is high; we recommend you take a break."

[0051] 3. User: Employees can choose to accept or reject the proposed mental health care action. They submit their decision to the system as feedback.

[0052] Feedback and continuous monitoring

[0053] 1. Server: Collects the results of the proposal's implementation and feedback, and records them in the system log.

[0054] 2. Server: Based on the collected feedback, the server will make timely improvements to the natural language processing algorithm and threshold settings.

[0055] 3. Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[0056] Specific example

[0057] This section describes a specific example of monitoring employee A's mental health.

[0058] 1. Device: Collect one week's worth of email and chat data from employee A's PC and securely transfer it to the cloud.

[0059] 2. Server: The server analyzes the data using natural language processing technology and determines that "negative tones" account for 50% of the responses. The stress score is calculated to be 75 (threshold is 70).

[0060] 3. Server: Generates an alert such as "Take a break" because the stress score exceeds the threshold.

[0061] 4. Server: Send an alert to the administrator and employee A, including specific care suggestions such as "Consult a mental health professional."

[0062] 5. Terminal: A notification will appear on employee A's PC, allowing employee A to submit feedback.

[0063] This system allows for proper management of employees' mental health and provision of necessary care, even in a remote work environment.

[0064] The following describes the processing flow.

[0065] Step 1:

[0066] Terminal: An employee logs into the system using a work-issued PC or mobile device. The system verifies the login information and begins preparing to collect data.

[0067] Step 2:

[0068] Terminal: A script runs that collects communication data such as employee emails, chats, meeting recordings, and voice memos. This data is stored in temporary files.

[0069] Step 3:

[0070] Terminal: Data stored in temporary files is encrypted and sent to the server using a secure protocol (e.g., HTTPS).

[0071] Step 4:

[0072] Server: Receives the transferred data and stores it in a secure database. The stored data is preprocessed for analysis (e.g., noise reduction, formatting standardization).

[0073] Step 5:

[0074] Server: Applies natural language processing (NLP) algorithms to analyze collected text and audio data. The analysis extracts emotional tones, the frequency of positive and negative language use, and specific stress signs.

[0075] Step 6:

[0076] Server: Based on the NLP results, it calculates emotional and stress scores to assess the mental health status of each employee.

[0077] Step 7:

[0078] Server: Compares the emotion score and stress score to pre-set thresholds and generates an alert if the stress score exceeds the threshold.

[0079] Step 8:

[0080] Server: Alerts include specific care suggestions and break suggestions. These alerts are recorded in the logging system.

[0081] Step 9:

[0082] Server: Sends generated alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[0083] Step 10:

[0084] Device: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a break" may be displayed.

[0085] Step 11:

[0086] User: Employees choose whether to accept the proposed mental health care action. The choice is sent to the system as feedback.

[0087] Step 12:

[0088] Server: Collects feedback from employees and administrators and uses it to make timely improvements to the system's natural language processing algorithms and threshold settings.

[0089] Step 13:

[0090] Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[0091] (Example 1)

[0092] 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."

[0093] With the widespread adoption of remote work, managing employees' mental health, in addition to monitoring work progress, has become a crucial issue. However, traditional methods make it difficult to grasp employees' stress levels and emotional states in real time and provide appropriate countermeasures, often resulting in mental health problems being left unaddressed. This leads to problems such as decreased work efficiency and increased employee turnover, making an effective mental health monitoring system necessary.

[0094] 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.

[0095] In this invention, the server includes means for monitoring the mental health status of employees in real time, means for collecting and analyzing employee communication data using natural language processing technology, and means for calculating and evaluating emotion scores and stress scores based on the analysis results. This makes it possible to effectively monitor the mental health of employees and provide timely and appropriate alerts and care suggestions.

[0096] "Mental health" refers to an individual's psychological, emotional, and social well-being. Mental health is a crucial factor that influences one's ability to cope with everyday stress, manage work and relationships, and more.

[0097] "Real-time monitoring" is a process of monitoring and evaluating data and conditions almost instantly. This process allows for a rapid response when problems occur.

[0098] "Natural language processing technology" refers to the technology that enables computers to understand and process human language. This includes text analysis and sentiment assessment.

[0099] "Communication data" refers to information such as emails, chats, meeting recordings, and voice memos exchanged by employees in the course of their work. This information is important data for evaluating the mental state of employees.

[0100] An "emotion score" is a numerical representation of an employee's emotional state, based on data analyzed using natural language processing technology. The score is used to evaluate emotions such as positive and negative.

[0101] A "stress score" is a numerical representation of an employee's stress level, based on data analyzed using natural language processing technology. This score is used to assess the mental burden on employees.

[0102] A "threshold" is a set score limit that triggers a specific alert or action. If the stress score exceeds this value, the system will take a specific action.

[0103] "Alert generation" is the process by which a system issues a warning or notification when set conditions or thresholds are met. This allows relevant parties to take immediate action.

[0104] "Notifying administrators and employees" refers to the process of sending analysis results and alert information to administrators' and employees' devices. Notifications are sent via methods such as email or applications.

[0105] "Mental health suggestions" refer to specific advice and action plans provided to improve employees' mental health. Examples include suggestions for breaks or consultations with mental health professionals.

[0106] "Collecting feedback" is the process by which employees submit to the system the results and opinions they have had regarding the mental health care suggestions they have received. The system then uses this feedback to make improvements.

[0107] "Improving analysis algorithms and threshold settings" refers to using collected feedback data to promptly update the system's analysis methods and the conditions for triggering actions to the latest and most optimal state.

[0108] The system of this invention monitors the mental health of employees working remotely in real time and detects alerts for stress levels. Specific embodiments are described below.

[0109] System Overview

[0110] This system uses natural language processing technology to collect and analyze employee communication data in order to monitor employees' mental health status in real time, and to calculate and evaluate emotion scores and stress scores. The server generates alerts based on stress scores that exceed a threshold and notifies both administrators and the employees themselves. It also provides specific mental care suggestions and collects employee feedback to improve the analysis algorithms and threshold settings.

[0111] Data collection

[0112] When an employee logs into the system using their work PC or mobile device, they enter their login information (e.g., user ID, password), and the system prepares to collect data. A script automatically starts on the terminal, collecting communication data such as emails, chats, meeting recordings, and voice memos, and saving it to a temporary file. The collected data is encrypted using AES (Advanced Encryption Standard) or similar and transferred to a secure server using the SSL / TLS protocol.

[0113] Data Analysis

[0114] The server receives the transferred data and stores it in a secure database such as AWS® RDS. The stored data undergoes a preprocessing process (e.g., noise reduction, formatting standardization) and then sentiment analysis is performed using natural language processing algorithms (e.g., GPT-3®, Python's NLTK library). This analysis calculates each employee's sentiment score (0-100) and stress score (1-10). This allows for an assessment of each employee's mental health status.

[0115] Alert generation and notification

[0116] The server compares the emotion score and stress score to a threshold (e.g., stress score of 70 or higher) and generates an alert if the threshold is exceeded. The generated alert is recorded in a logging system (e.g., MongoDB) and notified to the administrator and the employee. The notification includes specific mental health care suggestions (e.g., "Take a break," "Consult a mental health professional") via email (e.g., Amazon SES) or a dedicated application.

[0117] Employee feedback and system improvements

[0118] Employees, as users, choose to accept or reject proposed mental health care actions and provide feedback to the system. The server collects this feedback information and records it in the system log (e.g., Redis). Based on the collected feedback, the analysis algorithms and threshold settings are improved.

[0119] Continuous monitoring

[0120] The server periodically collects, analyzes, and generates alerts to continuously monitor employees' mental health. This makes it possible to properly manage employee mental health and provide necessary care, even in a remote work environment.

[0121] Specific example

[0122] For example, employee A collects a week's worth of email and chat data from their PC and securely transfers it to the cloud. The server then analyzes the data using natural language processing technology and determines that 50% of it has a "negative tone." If the stress score is calculated to be 75 (the threshold is 70), the system generates an alert such as "Take a break" and notifies both the administrator and employee A. It also includes specific care suggestions such as "Consult a mental health professional." Finally, employee A can choose whether to take the suggested action and provide feedback on the result.

[0123] Prompt example

[0124] "Create a program that uses employee email and chat data to monitor their mental health in real time. It should generate an alert and send a notification if the stress score exceeds a threshold. The technologies to use are the Python NLTK library and GPT-3."

[0125] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0126] Step 1:

[0127] Terminal: Employees log in to the system using a work-issued PC or mobile device.

[0128] Input: Authentication information such as user ID and password.

[0129] Data processing: Authentication processing of login information.

[0130] Output: A message indicating successful or failed login. The specific action involves entering a user ID and password and sending a request to the system's authentication server.

[0131] Step 2:

[0132] Terminal: Upon successful login, the script will automatically launch to collect communication data such as employee emails, chats, meeting recordings, and voice memos.

[0133] Input: Employee communication data (emails, chat logs, voice memos, etc.).

[0134] Data processing: Data collection and saving to temporary files.

[0135] Output: Communication data saved in a temporary file. Specifically, data is retrieved from each application (Outlook, Teams, etc.) using APIs and saved to a temporary file.

[0136] Step 3:

[0137] Terminal: Encrypts data stored in temporary files using AES or similar encryption methods and transfers it to the server using the SSL / TLS protocol.

[0138] Input: Communication data stored in a temporary file.

[0139] Data processing: Data encryption and secure transfer.

[0140] Output: An encrypted temporary file is transferred to the server. Specifically, the data is encrypted using an encryption library and uploaded to the server via HTTPS.

[0141] Step 4:

[0142] Server: Receives the transferred data and stores it in a secure database.

[0143] Input: Encrypted data that arrived on the server.

[0144] Data processing: Decrypting data and saving it to a database.

[0145] Output: Communication data stored in a secure database. Specifically, the data is decrypted and inserted into a database (e.g., AWS RDS).

[0146] Step 5:

[0147] Server: Preprocesses the stored data (noise reduction, format standardization).

[0148] Input: Communication data stored in the database.

[0149] Data processing: Data cleaning and formatting standardization.

[0150] Output: Preprocessed dataset. Specifically, data cleaning is performed using a script such as Python to remove noise and unwanted parts.

[0151] Step 6:

[0152] Server: Applies natural language processing algorithms to perform sentiment analysis.

[0153] Input: Preprocessed dataset.

[0154] Data processing: Sentiment analysis of text and audio data.

[0155] Output: Sentiment score and stress score. Specifically, the positive and negative sentiment of each document is analyzed using GPT-3 and the Python NLTK library.

[0156] Step 7:

[0157] Server: Evaluates each employee's mental health status based on the calculated emotion score and stress score.

[0158] Input: Emotion score and stress score.

[0159] Data processing: Assessment of mental health status.

[0160] Output: Evaluation result (normal, caution needed, dangerous, etc.). Specifically, the system compares the score with a threshold and classifies the state.

[0161] Step 8:

[0162] Server: Generates an alert when the stress score exceeds a threshold.

[0163] Input: Stress score for each employee.

[0164] Data processing: Comparison with thresholds and alert generation.

[0165] Output: Alert message. Specifically, if the stress score exceeds a specified threshold, an alert message such as "Take a break" is generated.

[0166] Step 9:

[0167] Server: Logs generated alerts and notifies administrators and employees themselves.

[0168] Input: Alert message.

[0169] Data processing: logging and sending notifications.

[0170] Output: Log entries and notification messages. Specifically, alert information is recorded in a logging system such as MongoDB, and administrators and employees are notified via email or a dedicated application.

[0171] Step 10:

[0172] User: Employees receive a notification and can choose whether to take the suggested mental health care action.

[0173] Input: Alert notification.

[0174] Data processing: Implementation and feedback of mental care actions.

[0175] Output: Feedback data. Specifically, an employee clicks the "Take Break" button and sends the result to the system.

[0176] Step 11:

[0177] Server: Collects feedback data and records it in the system log.

[0178] Input: Employee feedback.

[0179] Data processing: Recording feedback.

[0180] Output: Log entries. Specifically, feedback information is saved to a database such as Redis.

[0181] Step 12:

[0182] Server: Based on the collected feedback data, improve the analysis algorithm and threshold settings.

[0183] Input: Feedback data.

[0184] Data processing: Updating algorithms and settings.

[0185] Output: Improved analysis algorithm and threshold settings. Specifically, the parameters of the machine learning algorithm are readjusted based on the collected feedback.

[0186] Step 13:

[0187] Server: Regularly collects, analyzes, and generates alerts to continuously monitor employees' mental health.

[0188] Input: Communication data and feedback data from all employees.

[0189] Data processing: Iterative monitoring and alert generation process.

[0190] Output: Continuously updated mental health status and alerts. Specifically, the system periodically runs scripts to collect and analyze new data to monitor employee mental health.

[0191] (Application Example 1)

[0192] 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."

[0193] Monitoring the mental health of employees working remotely presents a challenge. In particular, there is a lack of technology to detect stress and fatigue in real time from employee voice and facial expressions, and to provide effective support. This can increase employee health risks and reduce work efficiency.

[0194] 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.

[0195] In this invention, the server includes means for monitoring the mental health status of employees, means for analyzing employee communication data using natural language processing technology, means for collecting voice and visual data using sensors in smart glasses, means for analyzing the employee's stress level from the collected visual and voice data, means for calculating an emotional score and a stress score based on the analysis results, means for notifying the administrator and the employee themselves when the stress score exceeds a set threshold, and means for providing specific mental care suggestions. This makes it possible to monitor the mental health status of employees in real time and provide appropriate care quickly.

[0196] An "employee" refers to an individual who belongs to an organization or company and engages in its operations.

[0197] "Means of monitoring mental health status" refers to techniques or methods for assessing and tracking the psychological and emotional health status of employees.

[0198] "Natural language processing technology" refers to computer technologies and algorithms used to understand and analyze human language.

[0199] "Communication data" refers to information collected in the form of emails, chats, meeting recordings, voice memos, and other similar formats.

[0200] "Smart glasses" are wearable devices that are used for data collection and analysis, and are capable of augmented reality and voice control.

[0201] "Means for collecting audio and visual data using sensors" refers to methods of acquiring audio and video information using devices such as microphones and cameras built into smart glasses.

[0202] "Methods for analyzing stress levels" refers to technologies that analyze collected audio and visual data to evaluate the stress levels of employees.

[0203] "Means for calculating emotional and stress scores" refers to techniques or methods for quantifying and evaluating indicators related to emotions and stress.

[0204] A "threshold" refers to a numerical value or condition that serves as the basis for generating an alert.

[0205] "Means of notification" refers to technologies and methods for communicating alerts and information to administrators and employees themselves.

[0206] "Mental health suggestions" refer to specific actions and advice recommended to improve employees' mental health.

[0207] A "server" refers to a computer system that handles data processing and storage.

[0208] The system of this invention monitors employees' mental health in real time, detects stress levels, and provides appropriate care. Specific embodiments are described below.

[0209] Data collection

[0210] Terminal (Smart Glasses): When an employee wears smart glasses, the device collects audio and visual data. The microphone and camera built into the smart glasses record the employee's voice and facial expressions in real time and temporarily store them in storage.

[0211] Data transfer and encryption

[0212] Device (smart glasses): The collected data is encrypted and transferred to the server using a secure protocol (e.g., TLS).

[0213] Data Analysis

[0214] Server: The server receives the transferred data and stores it in a secure database (e.g., PostgreSQL). The data is then preprocessed, and audio data is converted to text using the Google® Speech-to-Text API. For visual data, sentiment analysis is performed using an emotion recognition API (e.g., Amazon Comprehend).

[0215] Stress level analysis and score calculation

[0216] Server: Analyzes employee stress levels from collected audio and visual data. Natural language processing techniques and speech analysis libraries are used for the analysis. Based on the analysis results, emotion scores and stress scores are calculated.

[0217] Alert generation

[0218] Server: Generates an alert when the stress score exceeds a set threshold. The alert includes specific mental health care suggestions such as "Take a break" or "Consult a mental health professional."

[0219] notification

[0220] Server: Generated alerts are notified to both the administrator and the employee themselves. The notification is displayed on the smart glasses' screen, visually indicating the content of the alert.

[0221] Feedback and continuous monitoring

[0222] User: Employees can choose to accept or reject proposed care actions. The choice is sent to the system as feedback. This allows the system to continuously monitor the employee's mental health status and improve the analysis algorithms and threshold settings as needed.

[0223] Specific example: Stress detection during meetings

[0224] While employee B is participating in an important meeting, smart glasses analyze B's voice and facial expressions in real time. If the analysis determines that B is experiencing high stress levels, the smart glasses display a message saying, "Take a deep breath and relax." After the meeting, B is suggested to take a break, and if necessary, recommended to consult a mental health professional.

[0225] Prompt example:

[0226] Prompt text to input to the generative AI model:

[0227] "Please describe in detail the design of a smart glasses application that monitors employee stress levels in real time and suggests necessary care. Specifically, explain how voice and visual data will be collected, analyzed, and alerts will be generated. Also, describe in detail how employee feedback will be provided to the system regarding whether they accept the suggestions."

[0228] While the embodiments for carrying out the present invention have been described in detail, the invention is not limited thereto, and various modifications and improvements are possible.

[0229] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0230] Step 1: Data Collection

[0231] The device (smart glasses) collects audio and visual data when worn by an employee. Input is data from the microphone and camera built into the smart glasses, and output is raw audio and visual data temporarily stored in storage. Specifically, audio data is captured from the microphone every second, and visual data is periodically captured by the camera as snapshots and videos.

[0232] Step 2: Data encryption and transfer

[0233] The device (smart glasses) encrypts the collected audio and visual data and transfers it to the server via a secure protocol. The input is unencrypted collected data, and the output is encrypted data packets. Specifically, secure protocols such as TLS are used to protect the data from hacking and unauthorized access.

[0234] Step 3: Convert speech to text

[0235] The server receives the transferred data and stores it in a database. It then converts the audio data to text using the Google Speech-to-Text API. The input is encrypted audio data, and the output is text data. Specifically, an audio file is sent via an API call and returned as text.

[0236] Step 4: Emotion Analysis

[0237] The server uses text and visual data to perform sentiment analysis with an emotion recognition API (e.g., Amazon Comprehend). The input is text and image data, and the output is sentiment scores. Specifically, text data is sent to the sentiment recognition API, which then returns positive, negative, and neutral sentiment scores. Visual data is analyzed using a face recognition algorithm, and sentiment scores are similarly generated.

[0238] Step 5: Stress level analysis and score calculation

[0239] The server analyzes employee stress levels from collected visual and audio data. Inputs are emotion scores and features from audio data, while output is a stress score. Specifically, natural language processing techniques and speech analysis libraries are used to comprehensively evaluate stress indicators (e.g., "irritation," "fatigue") extracted from text and audio.

[0240] Step 6: Generate Alerts

[0241] The server generates an alert when the stress score exceeds a set threshold. The input is the stress score, and the output is an alert notification and specific mental health care suggestions. Specifically, when the score exceeds the set threshold, an alert message is generated, including suggestions such as "Take a break" or "Consult a mental health professional."

[0242] Step 7: Notification

[0243] The server notifies administrators and employees of the generated alerts. The input is the alert message, and the output is the notification displayed on the smart glasses' screen. Specifically, notification data is sent to the smart glasses via a network protocol, and the notification is sent to the administrator via email or chat application.

[0244] Step 8: Gathering Feedback

[0245] The user (employee) chooses whether to accept or reject the proposed care action. The input is the user's response to the proposal, and the output is feedback data sent to the system. Specifically, the option to "accept" or "reject" is displayed through the smart glasses interface, and the selection result is fed back to the server.

[0246] Step 9: Continuous Monitoring

[0247] The server continuously improves the algorithm and updates threshold settings based on the collected feedback. The input is feedback data, and the output is the updated algorithm parameters. Specifically, the feedback is used to optimize the machine learning model parameters, which are then reflected in the next analysis.

[0248] 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.

[0249] This invention relates to a system for monitoring the mental health of employees working remotely in real time and detecting their stress levels. In particular, by combining it with an emotion engine, it becomes possible to more accurately understand the emotional state of employees and propose appropriate care.

[0250] Data collection

[0251] 1. Terminal: Employees log in to the system using a work-issued PC or mobile device. The system verifies the login information and prepares to collect data.

[0252] 2. Terminal: A script is executed to collect communication data such as employee emails, chats, meeting recordings, and voice memos. This data is stored in temporary files.

[0253] 3. Terminal: Data stored in temporary files is encrypted and sent to the server using a secure protocol (e.g., HTTPS).

[0254] Data analysis and emotion recognition

[0255] 1. Server: Receives the transferred data and stores it in a secure database. The stored data is preprocessed for analysis (e.g., noise reduction, formatting standardization).

[0256] 2. Server: Applies natural language processing (NLP) algorithms to analyze the collected text and audio data. From the analysis results, it extracts emotional tone, frequency of positive and negative language use, and specific stress signs.

[0257] 3. Server: Furthermore, an emotion engine is used to recognize employees' emotions from communication data. Specifically, the emotion engine analyzes voice and text data to recognize emotions in real time.

[0258] Calculation of emotional score and stress score

[0259] 1. Server: Based on the results of NLP and the emotion engine, it calculates emotion scores and stress scores to assess the mental health status of each employee.

[0260] Alert generation

[0261] 1. Server: Compares the emotion score and stress score with pre-set thresholds and generates an alert if the stress score exceeds the threshold.

[0262] 2. Server: Alerts include specific care suggestions and break suggestions. These alerts are recorded in the logging system.

[0263] Notifications and suggestions

[0264] 1. Server: Sends generated alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[0265] 2. Devices: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a break" may be displayed.

[0266] 3. User: Employees choose whether to accept the proposed mental health care action. The choice is sent to the system as feedback.

[0267] Feedback and continuous monitoring

[0268] 1. Server: Collects feedback from employees and administrators and uses it to make timely improvements to the system's natural language processing algorithms, sentiment engine, and threshold settings.

[0269] 2. Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[0270] Specific example

[0271] This section describes a specific example of monitoring employee B's mental health.

[0272] 1. Device: Collect one week's worth of email and chat data from employee B's PC and securely transfer it to the cloud.

[0273] 2. Server: The server analyzes the collected data using natural language processing technology and an emotion engine. For example, "negative tones" are detected frequently, and emotions such as "anxiety" are identified from the audio data. The stress score becomes 80 (threshold is 70).

[0274] 3. Server: Generate an alert because the stress score exceeds the threshold, and include a suggestion such as "Take a break."

[0275] 4. Server: Send an alert to the administrator and employee B, including specific care suggestions such as "Consult a mental health professional."

[0276] 5. Terminal: A notification will appear on employee B's PC, allowing employee B to submit feedback.

[0277] This system allows for accurate and real-time monitoring of employees' emotional states, even in a remote work environment, and enables the provision of appropriate care.

[0278] The following describes the processing flow.

[0279] Step 1:

[0280] Terminal: The employee logs in to the system using a business PC or mobile device. Based on the login information, the system starts preparing for data collection.

[0281] Step 2:

[0282] Terminal: A script that collects the employee's communication data such as emails, chats, meeting recordings, and voice memos is automatically executed. These data are saved in a temporary file.

[0283] Step 3:

[0284] Terminal: The data saved in the temporary file is subjected to an encryption process and sent to the server using a secure protocol (e.g., HTTPS).

[0285] Step 4:

[0286] Server: The received data is saved in a secure database. The saved data is pre - processed for analysis (e.g., noise removal, format unification).

[0287] Step 5:

[0288] Server: Apply natural language processing (NLP) algorithms to analyze the collected text data and voice data. The analysis results include the emotional tone, positive expressions, negative expressions, and specific stress signs.

[0289] Step 6:

[0290] Server: Use an emotion engine to perform real - time emotion recognition from the same communication data. For example, emotions such as "uneasy" or "angry" are detected from voice data, and emotions such as "sad" or "stress" are identified from text data.

[0291] Step 7:

[0292] Server: Calculates employee emotional scores and stress scores based on the analysis results of NLP and the emotion engine.

[0293] Step 8:

[0294] Server: Compares the emotion score and stress score to the set threshold. If the stress score exceeds the threshold, an alert is generated.

[0295] Step 9:

[0296] Server: The generated alerts include specific care suggestions (e.g., suggestions for taking a break or advice on deep breathing) and links to consult with professionals. These alerts are recorded in the logging system.

[0297] Step 10:

[0298] Server: Sends alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[0299] Step 11:

[0300] Device: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a break" may be displayed.

[0301] Step 12:

[0302] User: Employees can choose whether or not to accept the proposed mental health care action. The choice is sent to the system as feedback.

[0303] Step 13:

[0304] Server: Receive feedback and improve natural language processing algorithms, sentiment engines, and threshold settings in a timely manner based on it.

[0305] Step 14:

[0306] Server: Regularly perform data collection, analysis, and alert generation, and continuously monitor the mental health of employees during remote work.

[0307] Specific Example

[0308] Describe a specific example of monitoring the mental health of Employee C.

[0309] Step 1:

[0310] Terminal: Collect one week's worth of email and chat data from Employee C's PC and securely transfer it to the cloud.

[0311] Step 2:

[0312] Server: Analyze the data using natural language processing technology and a sentiment engine. As a result of the analysis, there are many negative tones, and emotions such as anxiety are detected from the voice data. The stress score is calculated to be 85 (the threshold is 70).

[0313] Step 3:

[0314] Server: Since the stress score exceeds the threshold, generate an alert such as "Let's take a break." Specific suggestions and consultation links to experts are also included.

[0315] Step 4:

[0316] Server: Send the alert to the administrator and Employee C. ​​​​​​Terminal: A notification appears on employee C's PC, and when employee C submits feedback, the system receives it and improves the algorithm.

[0319] This system allows for accurate and real-time monitoring of employees' emotional states, even in a remote work environment, and enables the provision of appropriate care.

[0320] (Example 2)

[0321] 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 device 14 as the "terminal".

[0322] In a remote work environment, it is difficult to monitor employees' mental health in real time and accurately grasp their stress levels and emotional fluctuations. Furthermore, there is a lack of means to quickly provide appropriate care suggestions when employees are experiencing excessive stress. In such circumstances, it is difficult to prevent the deterioration of employees' mental health, potentially leading to decreased productivity and increased employee turnover.

[0323] 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.

[0324] In this invention, the server includes means for monitoring the mental health status of employees, means for analyzing employee communication data using natural language processing technology, and means for recognizing employee emotions from text and voice data using an emotion engine. This makes it possible to accurately grasp the mental health status of employees in real time and to quickly provide appropriate care suggestions.

[0325] "Mental health status" refers to an employee's mental and emotional well-being. This includes stress levels and emotional stability.

[0326] "Monitoring methods" refer to technologies and devices for continuously observing and recording employees' mental health status over time.

[0327] "Natural language processing technology" refers to artificial intelligence technology used to understand, analyze, and generate human language. This includes analyzing text data and classifying emotions.

[0328] "Communication data" refers to data obtained from communication methods used by employees during work, such as email, text chat, audio conference recordings, and voice memos.

[0329] An "emotion engine" refers to algorithms and technologies used to analyze and recognize emotions from text and audio data. This often includes generative AI models.

[0330] An "emotion score" is an indicator that numerically represents an employee's emotional state, calculated from the results of analysis by the emotion engine.

[0331] A "stress score" is an indicator that quantifies how much stress an employee is experiencing. It is calculated based on the results of an emotional analysis.

[0332] A "threshold" refers to a pre-set baseline value for emotion and stress scores. An alert is generated when this value is exceeded.

[0333] "Notification methods" refer to technologies and devices used to communicate alerts and care suggestions to administrators and employees themselves. This includes email and dedicated applications.

[0334] "Specific mental health support suggestions" refer to concrete actions and resources to improve employees' mental health. Examples include suggestions for breaks or links to mental health professionals.

[0335] This invention is a system that monitors the mental health of employees working remotely in real time and detects their stress levels. In particular, by combining it with an emotion engine, it is possible to more accurately understand the emotional state of employees and propose appropriate care.

[0336] Data collection

[0337] 1. Terminal: Employees log in to the system using a work PC or mobile device. The system verifies the login information and prepares to collect data.

[0338] 2. Device: After logging in, a script is automatically executed to collect communication data such as emails, chats, meeting recordings, and voice memos. For example, email data is collected from the mailbox by gathering the subject and body text, and chat data is retrieved using an API. Meeting recordings and voice memos are recorded via the device's microphone and saved as temporary files.

[0339] 3. Terminal: The collected data is stored in a temporary file and then encrypted using the AES-256 encryption algorithm. The encrypted data is then sent to the server using the secure HTTPS protocol.

[0340] Data analysis and emotion recognition

[0341] 1. Server: Receives data transferred from terminals and stores it in a secure database. After storage, the data undergoes preprocessing such as noise reduction and formatting standardization.

[0342] 2. Server: Applies natural language processing (NLP) algorithms to analyze text and audio data. For NLP analysis, libraries such as spaCy or NLTK are used. Audio data is converted to text using the Google Cloud Speech-to-Text API and integrated with the text data for analysis.

[0343] 3. Server: Recognizes emotions from text and audio data using an emotion engine. The emotion engine uses generative AI models such as BERT or GPT-3 to perform emotion classification tasks.

[0344] Calculation of emotional score and stress score

[0345] 1. Server: Based on the results of NLP and the emotion engine, it calculates an emotion score and a stress score. A weighted average or rule-based scoring system is used to calculate the scores. A higher emotion score is assigned when there is a high frequency of positive language use, and a higher stress score is assigned when there is a high frequency of negative language use.

[0346] Alert generation

[0347] 1. Server: Compares the emotion score and stress score to a threshold, and generates an alert if the stress score exceeds the threshold. For example, if the threshold is 70 and the stress score is 80, an alert will be generated.

[0348] 2. Server: Alerts include specific care suggestions. These suggestions may include consultation with a psychological counselor, use of relaxation apps, or suggestions for taking a break. These alerts are recorded in the logging system and stored as a dataset that can be used for analysis.

[0349] Notifications and suggestions

[0350] 1. Server: Notifies administrators and employees of generated alerts. Notifications are primarily made via email or a dedicated application (e.g., a company-specific mental health care app).

[0351] 2. Devices: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a 10-minute break" may pop up.

[0352] 3. User: Employees choose whether to accept the proposed care action. The choice is sent to the system as feedback and recorded in the database. This will be reflected in future analyses.

[0353] Feedback and continuous monitoring

[0354] 1. Server: Collects feedback from employees and administrators to improve the system's NLP algorithms, emotion engine, and threshold settings. Feedback is used to improve the accuracy of the emotion recognition model and readjust thresholds.

[0355] 2. Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely. For example, it performs weekly data analysis and provides administrators with monthly reports.

[0356] Specific examples and prompt statements

[0357] As a concrete example of monitoring employee B's mental health, the following operations are performed: Email and chat data for one week are collected from employee B's PC and securely transferred to the cloud. The collected data is analyzed using natural language processing technology and an emotion engine. If negative tones are detected frequently and emotions such as "anxiety" are identified from the voice data, the stress score becomes 80 (the threshold is 70). In this case, an alert is generated and a suggestion such as "Take a break" is made. The alert is sent to the administrator and employee B, and employee B can submit feedback.

[0358] Example prompt: "Collect one week's worth of emails, chats, and meeting recordings from employee B and analyze them using natural language processing and a sentiment engine. If the results show a high proportion of negative tones, generate specific care suggestions and notify the employee."

[0359] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0360] Step 1:

[0361] Terminal: Employees log in to the system using a work PC or mobile device.

[0362] Input: Employee login information (User ID, password)

[0363] Operation: The system authenticates login information and prepares for data collection.

[0364] Output: Login success message, instruction to start data collection script.

[0365] Step 2:

[0366] Terminal: After logging in, a script will run that automatically collects communication data such as emails, chats, meeting recordings, and voice memos.

[0367] Input: Employee account information after login authentication

[0368] Operation: Email data is collected from the subject and body text of employee mailboxes. Chat data is retrieved using an API. Meeting recordings and voice memos are recorded via the device's microphone and stored in temporary files.

[0369] Output: Text data (emails and chats), audio data (meeting recordings and voice memos), temporary files

[0370] Step 3:

[0371] Terminal: The collected data is saved to a temporary file and then encrypted using the AES-256 encryption algorithm.

[0372] Input: Raw data stored in a temporary file

[0373] Operation: Encrypts data using the AES-256 encryption algorithm. Sends the encrypted data to the server using the HTTPS protocol.

[0374] Output: Encrypted data, HTTPS request

[0375] Step 4:

[0376] Server: Receives data transferred from terminals and stores it in a secure database.

[0377] Input: Encrypted data included in the HTTPS request

[0378] Operation: Saves data to a database. After saving, preprocessing is performed on the data, such as noise reduction and formatting standardization.

[0379] Output: Preprocessed data

[0380] Step 5:

[0381] Server: Applies natural language processing (NLP) algorithms to analyze text and audio data.

[0382] Input: Preprocessed text data and audio data

[0383] Operation: For NLP analysis, libraries such as spaCy or NLTK are used to tokenize text data and tag parts of speech. Audio data is converted to text using the Google Cloud Speech-to-Text API and integrated with the text data for analysis.

[0384] Output: Analyzed text data, text data converted from audio data

[0385] Step 6:

[0386] Server: Uses an emotion engine to recognize emotions from text and audio data.

[0387] Input: Analyzed text data, text data converted from audio data

[0388] Operation: The emotion engine uses generative AI models such as BERT or GPT-3 to perform emotion classification tasks. It assigns emotion labels such as positive, negative, and neutral to the data.

[0389] Output: Data with emotion labels

[0390] Step 7:

[0391] Server: Calculates emotion scores and stress scores based on the results of NLP and the emotion engine.

[0392] Input: Data with emotion labels

[0393] Operation: The scoring system uses weighted averages or rule-based scoring systems to calculate scores. A higher sentiment score is assigned when positive language usage is frequent, and a higher stress score is assigned when negative language usage is frequent.

[0394] Output: Recorded sentiment score and stress score

[0395] Step 8:

[0396] Server: Compares the emotion score and stress score to a threshold, and generates an alert if the stress score exceeds the threshold.

[0397] Input: Calculated sentiment score and stress score, threshold setting

[0398] Operation: Determines whether the stress score exceeds a threshold and generates an alert if it does.

[0399] Output: Alert Information

[0400] Step 9:

[0401] Server: Based on alerts recorded in the database, it notifies administrators and employees themselves.

[0402] Input: Alert Information

[0403] Operation: Notifies administrators and employees via email or a dedicated application. Sends messages such as, "Your stress level is high. We recommend you take a break."

[0404] Output: Notification message

[0405] Step 10:

[0406] Device: Notifications will be displayed on employees' PCs and mobile devices.

[0407] Input: Notification message

[0408] Action: A notification message is displayed. A pop-up appears saying something like, "Your stress level is high. We recommend you take a 10-minute break."

[0409] Output: Displayed notification message

[0410] Step 11:

[0411] User: The employee chooses whether to accept the proposed care action.

[0412] Input: Displayed notification message

[0413] Operation: Employees can choose to "accept" or "reject" a notification, and the system receives feedback on their choice.

[0414] Output: Acceptance or rejection feedback information

[0415] Step 12:

[0416] Server: Collects feedback from employees and administrators to improve the system's NLP algorithms, sentiment engine, and threshold settings.

[0417] Input: Feedback information

[0418] Operation: Feedback is recorded in a database and used to adjust algorithms, emotion engines, and threshold settings for future analyses.

[0419] Output: Improved NLP algorithm, emotion engine, threshold settings

[0420] Step 13:

[0421] Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[0422] Input: Continuously collected communication data

[0423] Operation: Performs weekly data analysis. Provides administrators with monthly reports summarizing the data.

[0424] Output: Weekly data analysis report, monthly report

[0425] (Application Example 2)

[0426] 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".

[0427] The mental health of factory workers is susceptible to the effects of harsh working conditions and stress, making appropriate monitoring and care essential. However, current systems lack the ability to grasp workers' stress and emotions in real time and provide appropriate care suggestions. In particular, there is a need to quickly detect changes in emotions and provide workers with effective improvement suggestions. Therefore, there is a demand for a system that efficiently and accurately monitors the mental health of employees working on-site and provides appropriate care suggestions in real time.

[0428] 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.

[0429] In this invention, the server includes means for monitoring the mental health status of employees, means for analyzing employee communication data using natural language processing technology, means for calculating an emotion score and a stress score based on the analysis results, means for notifying managers and the employee themselves when the stress score exceeds a set threshold, means for providing specific mental care suggestions, means for collecting voice and text communications of factory workers, means for securely transferring the collected data, means for notifying workers of alerts using factory robots, head-mounted displays, or smart glasses, and means for collecting feedback from managers and workers and improving the system's algorithms and threshold settings. This makes it possible to accurately monitor the mental health of factory workers in real time and to quickly provide appropriate care suggestions.

[0430] "Employee mental health status" refers to the psychological and emotional health of workers.

[0431] "Monitoring methods" refer to technologies and devices used to monitor the status of employees by collecting and analyzing data in real time or periodically.

[0432] "Natural language processing technology" is a technology that enables computers to analyze, understand, and generate human language.

[0433] "Communication data" refers to data in the form of emails, chats, meeting recordings, voice memos, work reports, conversations, etc., that workers exchange in their daily work.

[0434] "Emotion score" refers to a numerical value that quantitatively represents the emotional state of a worker, calculated using natural language processing and emotion recognition technologies.

[0435] A "stress score" refers to a numerical value that evaluates the stress level of workers based on their emotional scores.

[0436] A "threshold" refers to a value that serves as a benchmark for indicating that an emotional score or stress score has reached a specific state.

[0437] "Means of notification" refers to technologies and devices used to transmit information to relevant parties using alerts or messages.

[0438] "Mental care suggestions" refer to specific advice and action proposals aimed at improving the mental health of workers.

[0439] "Means of collecting voice and text communications" refers to technologies and devices for acquiring voice and text data from workers.

[0440] "Secure transfer" refers to encrypting data and transmitting it using a secure protocol.

[0441] "Factory robots" refer to mechanical devices designed to automatically perform tasks and operations within a factory.

[0442] A "head-mounted display (HMD)" refers to a device that displays visual information when worn on the head.

[0443] "Smart glasses" refer to glasses-type wearable devices that are capable of displaying visual information and collecting data.

[0444] An "alert" refers to a warning message that is sent when a specific situation or condition occurs.

[0445] "Means of collecting feedback" refers to technologies and equipment used to gather opinions and evaluations from workers and managers.

[0446] "Means of improving system algorithms and threshold settings" refers to techniques and processes for optimizing system performance and settings based on feedback and data.

[0447] This invention relates to a system for monitoring the mental health of factory workers in real time and detecting stress levels. In particular, by combining it with an emotion engine, it is possible to accurately grasp the emotional state of workers and propose appropriate care. The following hardware and software are required to implement the system of this invention.

[0448] 1. Hardware

[0449] Factory robot: A robotic device that performs tasks within a factory.

[0450] Head-mounted displays (HMDs): For example, Microsoft HoloLens® can be used to display visual information to workers and deliver notifications.

[0451] Smart glasses: For example, wearable devices that use Google Glass® to collect voice and text data.

[0452] 2. Software

[0453] TENSORFLOW®: Data analysis using natural language processing (NLP) technology.

[0454] IBM Watson®: Used as an emotion engine to calculate emotion scores from collected data.

[0455] AWS: Manage databases and perform secure data transfer in the cloud.

[0456] Python: Creating programs for data preprocessing, analysis, and alert generation.

[0457] The system operates as follows:

[0458] 1. Data Collection

[0459] The system collects voice and text data from the terminal through the worker's HMD or smart glasses. For example, conversations and work reports made by workers during their work are collected as data.

[0460] 2. Data Transfer

[0461] The collected data is stored as a temporary file, encrypted, and securely transmitted to the AWS cloud using the HTTPS protocol.

[0462] 3. Data Analysis

[0463] Data is preprocessed on the cloud by a server, and text data is analyzed using NLP analysis with TensorFlow. In addition, IBM Watson's emotion engine analyzes audio data to calculate emotion scores and stress scores. For example, if many negative emotions such as "tired" or "irritated" are detected during work, the stress score will be high.

[0464] 4. Alert generation and notification

[0465] The server generates alerts based on emotion and stress scores, and sends alert messages containing appropriate care suggestions to the worker's HMD or smart glasses if the threshold is exceeded. For example, if the stress score exceeds the threshold, a message such as "We recommend you take a break" will be displayed.

[0466] 5. Gathering feedback and making improvements

[0467] The server collects feedback from workers and administrators, and uses it to improve the system's algorithms and threshold settings. This feedback allows the system to continuously improve its performance.

[0468] As a concrete example, consider a case where worker A at a factory wears an HMD while working. Daily voice and text data collected from this worker is transferred to the AWS cloud and analyzed by TensorFlow and IBM Watson. If the emotion score indicates "fatigue" or "dissatisfaction," and the stress score exceeds a threshold, an alert is displayed on the HMD. Subsequently, feedback from worker A is sent to the server, and the system is further optimized.

[0469] Examples of prompts to input into a generative AI model are as follows:

[0470] "Develop a system to monitor the mental health of factory workers and provide real-time care suggestions based on their stress levels. This system will use factory robots, HMDs, and smart glasses to collect and analyze employee voice and text data. It should calculate employee emotion and stress scores using natural language processing algorithms and an emotion recognition engine, and then provide appropriate care suggestions."

[0471] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0472] Step 1:

[0473] The terminal collects voice and text data from the worker's head-mounted display (HMD) or smart glasses. The input at this stage is voice and text communication data generated during the worker's daily work. The output is the collected data, which is stored as a temporary file.

[0474] Step 2:

[0475] The device encrypts the collected temporary file data and sends it to the AWS cloud using a secure protocol (e.g., HTTPS). The input at this stage is the unencrypted data stored in the temporary file, and the output is a notification that the transfer to the secure cloud is complete.

[0476] Step 3:

[0477] The server decrypts the data received in the cloud and performs preprocessing (noise reduction, format standardization). The input is encrypted collected data, and the output is preprocessed data. Specifically, background noise is removed from the audio data, and the format of the text data is standardized.

[0478] Step 4:

[0479] The server applies natural language processing (NLP) to preprocessed data to analyze the text data. The software used is TensorFlow; the input is preprocessed text data, and the output is the result of the text analysis. Specifically, it extracts emotional tone and the frequency of positive and negative language use.

[0480] Step 5:

[0481] The server uses an emotion engine (IBM Watson) to analyze voice data and calculate an emotion score. The input is pre-processed voice data, and the output is the emotion score. Specifically, it analyzes the intonation and tone of the voice to determine the emotional state.

[0482] Step 6:

[0483] The server calculates a stress score based on NLP analysis results and emotion engine results. The input is text analysis results and emotion score, and the output is the stress score. Specifically, it integrates the emotional tone of text and voice to assess the employee's stress level.

[0484] Step 7:

[0485] The server checks whether the calculated stress score exceeds a set threshold. The input is the stress score, and the output is a flag that instructs the server to generate an alert if the threshold is exceeded.

[0486] Step 8:

[0487] The server generates an alert when a threshold is exceeded and sends an alert message, including specific care suggestions, to the worker's HMD or smart glasses. The input is the result of the threshold being exceeded, and the output is the alert notification to the worker. Specifically, a message such as "We recommend you take a break" might be displayed.

[0488] Step 9:

[0489] The server collects feedback from workers and managers through HMDs and smart glasses, and uses this feedback to improve the system's algorithms and threshold settings. The input is feedback data, and the output is the improved algorithms and settings. Specifically, it analyzes the feedback and adjusts the algorithm parameters.

[0490] The following example prompts illustrate the specific actions required to perform each step:

[0491] "Develop a system to monitor the mental health of factory workers and provide real-time care suggestions based on their stress levels. This system will use factory robots, HMDs, and smart glasses to collect and analyze employee voice and text data. It should calculate employee emotion and stress scores using natural language processing algorithms and an emotion recognition engine, and then provide appropriate care suggestions."

[0492] 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.

[0493] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (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.

[0494] 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.

[0495] [Second Embodiment]

[0496] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0497] 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.

[0498] 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).

[0499] 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.

[0500] 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.

[0501] 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).

[0502] 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.

[0503] 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.

[0504] 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.

[0505] 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.

[0506] 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.

[0507] 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".

[0508] The system of this invention monitors the mental health of employees working remotely in real time and detects alerts for stress levels. Specific embodiments are described below.

[0509] Data collection

[0510] 1. Terminal: Employees log in to the system using a work-issued PC or mobile device. Based on the login information, the system prepares to collect data.

[0511] 2. Terminal: A script starts running to collect communication data such as employee emails, chats, meeting recordings, and voice memos. The collected data is saved to a temporary file.

[0512] 3. Terminal: Communication data stored in temporary files is encrypted and transferred to the server using a secure protocol.

[0513] Data Analysis

[0514] 1. Server: Receives the transferred data and stores it in a secure database. The stored data is pre-processed (e.g., noise reduction, formatting standardization).

[0515] 2. Server: Applies natural language processing algorithms to perform sentiment analysis on text and audio data. The analysis includes emotional tone, frequency of positive and negative language use, and specific stress signs (e.g., frequent use of expressions like "tired" or "irritated").

[0516] 3. Server: Calculates emotional and stress scores to assess the mental health status of each employee.

[0517] Alert generation

[0518] 1. Server: Compares the emotional score and stress score to a threshold and generates an alert if the stress score exceeds the threshold. The alert includes specific care suggestions or suggestions for taking a break.

[0519] 2. Server: Records generated alerts in the logging system for future analysis and feedback.

[0520] Notifications and suggestions

[0521] 1. Server: Sends alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[0522] 2. Devices: Notifications are displayed on employees' PCs and mobile devices, along with a message such as, "Your stress level is high; we recommend you take a break."

[0523] 3. User: Employees can choose to accept or reject the proposed mental health care action. They submit their decision to the system as feedback.

[0524] Feedback and continuous monitoring

[0525] 1. Server: Collects the results of the proposal's implementation and feedback, and records them in the system log.

[0526] 2. Server: Based on the collected feedback, the server will make timely improvements to the natural language processing algorithm and threshold settings.

[0527] 3. Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[0528] Specific example

[0529] This section describes a specific example of monitoring employee A's mental health.

[0530] 1. Device: Collect one week's worth of email and chat data from employee A's PC and securely transfer it to the cloud.

[0531] 2. Server: The server analyzes the data using natural language processing technology and determines that "negative tones" account for 50% of the responses. The stress score is calculated to be 75 (threshold is 70).

[0532] 3. Server: Generates an alert such as "Take a break" because the stress score exceeds the threshold.

[0533] 4. Server: Send an alert to the administrator and employee A, including specific care suggestions such as "Consult a mental health professional."

[0534] 5. Terminal: A notification will appear on employee A's PC, allowing employee A to submit feedback.

[0535] This system allows for proper management of employees' mental health and provision of necessary care, even in a remote work environment.

[0536] The following describes the processing flow.

[0537] Step 1:

[0538] Terminal: An employee logs into the system using a work-issued PC or mobile device. The system verifies the login information and begins preparing to collect data.

[0539] Step 2:

[0540] Terminal: A script runs that collects communication data such as employee emails, chats, meeting recordings, and voice memos. This data is stored in temporary files.

[0541] Step 3:

[0542] Terminal: Data stored in temporary files is encrypted and sent to the server using a secure protocol (e.g., HTTPS).

[0543] Step 4:

[0544] Server: Receives the transferred data and stores it in a secure database. The stored data is preprocessed for analysis (e.g., noise reduction, formatting standardization).

[0545] Step 5:

[0546] Server: Applies natural language processing (NLP) algorithms to analyze collected text and audio data. The analysis extracts emotional tones, the frequency of positive and negative language use, and specific stress signs.

[0547] Step 6:

[0548] Server: Based on the NLP results, it calculates emotional and stress scores to assess the mental health status of each employee.

[0549] Step 7:

[0550] Server: Compares the emotion score and stress score to pre-set thresholds and generates an alert if the stress score exceeds the threshold.

[0551] Step 8:

[0552] Server: Alerts include specific care suggestions and break suggestions. These alerts are recorded in the logging system.

[0553] Step 9:

[0554] Server: Sends generated alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[0555] Step 10:

[0556] Device: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a break" may be displayed.

[0557] Step 11:

[0558] User: Employees choose whether to accept the proposed mental health care action. The choice is sent to the system as feedback.

[0559] Step 12:

[0560] Server: Collects feedback from employees and administrators and uses it to make timely improvements to the system's natural language processing algorithms and threshold settings.

[0561] Step 13:

[0562] Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[0563] (Example 1)

[0564] 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."

[0565] With the widespread adoption of remote work, managing employees' mental health, in addition to monitoring work progress, has become a crucial issue. However, traditional methods make it difficult to grasp employees' stress levels and emotional states in real time and provide appropriate countermeasures, often resulting in mental health problems being left unaddressed. This leads to problems such as decreased work efficiency and increased employee turnover, making an effective mental health monitoring system necessary.

[0566] 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.

[0567] In this invention, the server includes means for monitoring the mental health status of employees in real time, means for collecting and analyzing employee communication data using natural language processing technology, and means for calculating and evaluating emotion scores and stress scores based on the analysis results. This makes it possible to effectively monitor the mental health of employees and provide timely and appropriate alerts and care suggestions.

[0568] "Mental health" refers to an individual's psychological, emotional, and social well-being. Mental health is a crucial factor that influences one's ability to cope with everyday stress, manage work and relationships, and more.

[0569] "Real-time monitoring" is a process of monitoring and evaluating data and conditions almost instantly. This process allows for a rapid response when problems occur.

[0570] "Natural language processing technology" refers to the technology that enables computers to understand and process human language. This includes text analysis and sentiment assessment.

[0571] "Communication data" refers to information such as emails, chats, meeting recordings, and voice memos exchanged by employees in the course of their work. This information is important data for evaluating the mental state of employees.

[0572] An "emotion score" is a numerical representation of an employee's emotional state, based on data analyzed using natural language processing technology. The score is used to evaluate emotions such as positive and negative.

[0573] A "stress score" is a numerical representation of an employee's stress level, based on data analyzed using natural language processing technology. This score is used to assess the mental burden on employees.

[0574] A "threshold" is a set score limit that triggers a specific alert or action. If the stress score exceeds this value, the system will take a specific action.

[0575] "Alert generation" is the process by which a system issues a warning or notification when set conditions or thresholds are met. This allows relevant parties to take immediate action.

[0576] "Notifying administrators and employees" refers to the process of sending analysis results and alert information to administrators' and employees' devices. Notifications are sent via methods such as email or applications.

[0577] "Mental health suggestions" refer to specific advice and action plans provided to improve employees' mental health. Examples include suggestions for breaks or consultations with mental health professionals.

[0578] "Collecting feedback" is the process by which employees submit to the system the results and opinions they have had regarding the mental health care suggestions they have received. The system then uses this feedback to make improvements.

[0579] "Improving analysis algorithms and threshold settings" refers to using collected feedback data to promptly update the system's analysis methods and the conditions for triggering actions to the latest and most optimal state.

[0580] The system of this invention monitors the mental health of employees working remotely in real time and detects alerts for stress levels. Specific embodiments are described below.

[0581] System Overview

[0582] This system uses natural language processing technology to collect and analyze employee communication data in order to monitor employees' mental health status in real time, and to calculate and evaluate emotion scores and stress scores. The server generates alerts based on stress scores that exceed a threshold and notifies both administrators and the employees themselves. It also provides specific mental care suggestions and collects employee feedback to improve the analysis algorithms and threshold settings.

[0583] Data collection

[0584] When an employee logs into the system using their work PC or mobile device, they enter their login information (e.g., user ID, password), and the system prepares to collect data. A script automatically starts on the terminal, collecting communication data such as emails, chats, meeting recordings, and voice memos, and saving it to a temporary file. The collected data is encrypted using AES (Advanced Encryption Standard) or similar and transferred to a secure server using the SSL / TLS protocol.

[0585] Data Analysis

[0586] The server receives the transferred data and stores it in a secure database such as AWS RDS. The stored data undergoes a preprocessing process (e.g., noise reduction, formatting standardization) and then sentiment analysis is performed using natural language processing algorithms (e.g., GPT-3, Python's NLTK library). This analysis calculates each employee's sentiment score (0-100) and stress score (1-10). This allows for an assessment of each employee's mental health status.

[0587] Alert generation and notification

[0588] The server compares the emotion score and stress score to a threshold (e.g., stress score of 70 or higher) and generates an alert if the threshold is exceeded. The generated alert is recorded in a logging system (e.g., MongoDB) and notified to the administrator and the employee. The notification includes specific mental health care suggestions (e.g., "Take a break," "Consult a mental health professional") via email (e.g., Amazon SES) or a dedicated application.

[0589] Employee feedback and system improvements

[0590] Employees, as users, choose to accept or reject proposed mental health care actions and provide feedback to the system. The server collects this feedback information and records it in the system log (e.g., Redis). Based on the collected feedback, the analysis algorithms and threshold settings are improved.

[0591] Continuous monitoring

[0592] The server periodically collects, analyzes, and generates alerts to continuously monitor employees' mental health. This makes it possible to properly manage employee mental health and provide necessary care, even in a remote work environment.

[0593] Specific example

[0594] For example, employee A collects a week's worth of email and chat data from their PC and securely transfers it to the cloud. The server then analyzes the data using natural language processing technology and determines that 50% of it has a "negative tone." If the stress score is calculated to be 75 (the threshold is 70), the system generates an alert such as "Take a break" and notifies both the administrator and employee A. It also includes specific care suggestions such as "Consult a mental health professional." Finally, employee A can choose whether to take the suggested action and provide feedback on the result.

[0595] Prompt example

[0596] "Create a program that uses employee email and chat data to monitor their mental health in real time. It should generate an alert and send a notification if the stress score exceeds a threshold. The technologies to use are the Python NLTK library and GPT-3."

[0597] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0598] Step 1:

[0599] Terminal: Employees log in to the system using a work-issued PC or mobile device.

[0600] Input: Authentication information such as user ID and password.

[0601] Data processing: Authentication processing of login information.

[0602] Output: A message indicating successful or failed login. The specific action involves entering a user ID and password and sending a request to the system's authentication server.

[0603] Step 2:

[0604] Terminal: If login is successful, the script will automatically start and collect communication data such as employee emails, chats, meeting recordings, and voice memos.

[0605] Input: Employee communication data (emails, chat logs, voice memos, etc.).

[0606] Data processing: Data collection and saving to temporary files.

[0607] Output: Communication data saved in a temporary file. Specifically, data is retrieved from each application (Outlook, Teams, etc.) using APIs and saved to a temporary file.

[0608] Step 3:

[0609] Terminal: Encrypts data stored in temporary files using AES or similar encryption methods and transfers it to the server using the SSL / TLS protocol.

[0610] Input: Communication data stored in a temporary file.

[0611] Data processing: Data encryption and secure transfer.

[0612] Output: An encrypted temporary file is transferred to the server. Specifically, the data is encrypted using an encryption library and uploaded to the server via HTTPS.

[0613] Step 4:

[0614] Server: Receives the transferred data and stores it in a secure database.

[0615] Input: Encrypted data that arrived on the server.

[0616] Data processing: Decrypting data and saving it to a database.

[0617] Output: Communication data stored in a secure database. Specifically, the data is decrypted and inserted into a database (e.g., AWS RDS).

[0618] Step 5:

[0619] Server: Preprocesses the stored data (noise reduction, format standardization).

[0620] Input: Communication data stored in the database.

[0621] Data processing: Data cleaning and formatting standardization.

[0622] Output: Preprocessed dataset. Specifically, data cleaning is performed using a script such as Python to remove noise and unwanted parts.

[0623] Step 6:

[0624] Server: Applies natural language processing algorithms to perform sentiment analysis.

[0625] Input: Preprocessed dataset.

[0626] Data processing: Sentiment analysis of text and audio data.

[0627] Output: Sentiment score and stress score. Specifically, the positive and negative sentiment of each document is analyzed using GPT-3 and the Python NLTK library.

[0628] Step 7:

[0629] Server: Evaluates each employee's mental health status based on the calculated emotion score and stress score.

[0630] Input: Emotion score and stress score.

[0631] Data processing: Assessment of mental health status.

[0632] Output: Evaluation result (normal, caution needed, dangerous, etc.). Specifically, the system compares the score with a threshold and classifies the state.

[0633] Step 8:

[0634] Server: Generates an alert when the stress score exceeds a threshold.

[0635] Input: Stress score for each employee.

[0636] Data processing: Comparison with thresholds and alert generation.

[0637] Output: Alert message. Specifically, if the stress score exceeds a specified threshold, an alert message such as "Take a break" is generated.

[0638] Step 9:

[0639] Server: Logs generated alerts and notifies administrators and employees themselves.

[0640] Input: Alert message.

[0641] Data processing: logging and sending notifications.

[0642] Output: Log entries and notification messages. Specifically, alert information is recorded in a logging system such as MongoDB, and administrators and employees are notified via email or a dedicated application.

[0643] Step 10:

[0644] User: Employees receive a notification and can choose whether to take the suggested mental health care action.

[0645] Input: Alert notification.

[0646] Data processing: Implementation and feedback of mental care actions.

[0647] Output: Feedback data. Specifically, an employee clicks the "Take Break" button and sends the result to the system.

[0648] Step 11:

[0649] Server: Collects feedback data and records it in the system log.

[0650] Input: Employee feedback.

[0651] Data processing: Recording feedback.

[0652] Output: Log entries. Specifically, feedback information is saved to a database such as Redis.

[0653] Step 12:

[0654] Server: Based on the collected feedback data, improve the analysis algorithm and threshold settings.

[0655] Input: Feedback data.

[0656] Data processing: Updating algorithms and settings.

[0657] Output: Improved analysis algorithm and threshold settings. Specifically, the parameters of the machine learning algorithm are readjusted based on the collected feedback.

[0658] Step 13:

[0659] Server: Regularly collects, analyzes, and generates alerts to continuously monitor employees' mental health.

[0660] Input: Communication data and feedback data from all employees.

[0661] Data processing: Iterative monitoring and alert generation process.

[0662] Output: Continuously updated mental health status and alerts. Specifically, the system periodically runs scripts to collect and analyze new data to monitor employee mental health.

[0663] (Application Example 1)

[0664] 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."

[0665] Monitoring the mental health of employees working remotely presents a challenge. In particular, there is a lack of technology to detect stress and fatigue in real time from employee voice and facial expressions, and to provide effective support. This can increase employee health risks and reduce work efficiency.

[0666] 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.

[0667] In this invention, the server includes means for monitoring the mental health status of employees, means for analyzing employee communication data using natural language processing technology, means for collecting voice and visual data using sensors in smart glasses, means for analyzing the employee's stress level from the collected visual and voice data, means for calculating an emotional score and a stress score based on the analysis results, means for notifying the administrator and the employee themselves when the stress score exceeds a set threshold, and means for providing specific mental care suggestions. This makes it possible to monitor the mental health status of employees in real time and provide appropriate care quickly.

[0668] An "employee" refers to an individual who belongs to an organization or company and engages in its operations.

[0669] "Means of monitoring mental health status" refers to techniques or methods for assessing and tracking the psychological and emotional health status of employees.

[0670] "Natural language processing technology" refers to computer technologies and algorithms used to understand and analyze human language.

[0671] "Communication data" refers to information collected in the form of emails, chats, meeting recordings, voice memos, and other similar formats.

[0672] "Smart glasses" are wearable devices that are used for data collection and analysis, and are capable of augmented reality and voice control.

[0673] "Means for collecting audio and visual data using sensors" refers to methods of acquiring audio and video information using devices such as microphones and cameras built into smart glasses.

[0674] "Methods for analyzing stress levels" refers to technologies that analyze collected audio and visual data to evaluate the stress levels of employees.

[0675] "Means for calculating emotional and stress scores" refers to techniques or methods for quantifying and evaluating indicators related to emotions and stress.

[0676] A "threshold" refers to a numerical value or condition that serves as the basis for generating an alert.

[0677] "Means of notification" refers to technologies and methods for communicating alerts and information to administrators and employees themselves.

[0678] "Mental health suggestions" refer to specific actions and advice recommended to improve employees' mental health.

[0679] A "server" refers to a computer system that handles data processing and storage.

[0680] The system of this invention monitors employees' mental health in real time, detects stress levels, and provides appropriate care. Specific embodiments are described below.

[0681] Data collection

[0682] Terminal (Smart Glasses): When an employee wears smart glasses, the device collects audio and visual data. The microphone and camera built into the smart glasses record the employee's voice and facial expressions in real time and temporarily store them in storage.

[0683] Data transfer and encryption

[0684] Device (smart glasses): The collected data is encrypted and transferred to the server using a secure protocol (e.g., TLS).

[0685] Data Analysis

[0686] Server: The server receives the transferred data and stores it in a secure database (e.g., PostgreSQL). The data is then preprocessed, and audio data is converted to text using the Google Speech-to-Text API. For visual data, sentiment analysis is performed using an emotion recognition API (e.g., Amazon Comprehend).

[0687] Stress level analysis and score calculation

[0688] Server: Analyzes employee stress levels from collected audio and visual data. Natural language processing techniques and speech analysis libraries are used for the analysis. Based on the analysis results, emotion scores and stress scores are calculated.

[0689] Alert generation

[0690] Server: Generates an alert when the stress score exceeds a set threshold. The alert includes specific mental health care suggestions such as "Take a break" or "Consult a mental health professional."

[0691] notification

[0692] Server: Generated alerts are notified to both the administrator and the employee themselves. The notification is displayed on the smart glasses' screen, visually indicating the content of the alert.

[0693] Feedback and continuous monitoring

[0694] User: Employees can choose to accept or reject proposed care actions. The choice is sent to the system as feedback. This allows the system to continuously monitor the employee's mental health status and improve the analysis algorithms and threshold settings as needed.

[0695] Specific example: Stress detection during meetings

[0696] While employee B is participating in an important meeting, smart glasses analyze B's voice and facial expressions in real time. If the analysis determines that B is experiencing high stress levels, the smart glasses display a message saying, "Take a deep breath and relax." After the meeting, B is suggested to take a break, and if necessary, recommended to consult a mental health professional.

[0697] Prompt example:

[0698] Prompt text to input to the generative AI model:

[0699] "Please describe in detail the design of a smart glasses application that monitors employee stress levels in real time and suggests necessary care. Specifically, explain how voice and visual data will be collected, analyzed, and alerts will be generated. Also, describe in detail how employee feedback will be provided to the system regarding whether they accept the suggestions."

[0700] While the embodiments for carrying out the present invention have been described in detail, the invention is not limited thereto, and various modifications and improvements are possible.

[0701] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0702] Step 1: Data Collection

[0703] The device (smart glasses) collects audio and visual data when worn by an employee. Input is data from the microphone and camera built into the smart glasses, and output is raw audio and visual data temporarily stored in storage. Specifically, audio data is captured from the microphone every second, and visual data is periodically captured by the camera as snapshots and videos.

[0704] Step 2: Data encryption and transfer

[0705] The device (smart glasses) encrypts the collected audio and visual data and transfers it to the server via a secure protocol. The input is unencrypted collected data, and the output is encrypted data packets. Specifically, secure protocols such as TLS are used to protect the data from hacking and unauthorized access.

[0706] Step 3: Convert speech to text

[0707] The server receives the transferred data and stores it in a database. It then converts the audio data to text using the Google Speech-to-Text API. The input is encrypted audio data, and the output is text data. Specifically, an audio file is sent via an API call and returned as text.

[0708] Step 4: Emotion Analysis

[0709] The server uses text and visual data to perform sentiment analysis with an emotion recognition API (e.g., Amazon Comprehend). The input is text and image data, and the output is sentiment scores. Specifically, text data is sent to the sentiment recognition API, which then returns positive, negative, and neutral sentiment scores. Visual data is analyzed using a face recognition algorithm, and sentiment scores are similarly generated.

[0710] Step 5: Stress level analysis and score calculation

[0711] The server analyzes employee stress levels from collected visual and audio data. Inputs are emotion scores and features from audio data, while output is a stress score. Specifically, natural language processing techniques and speech analysis libraries are used to comprehensively evaluate stress indicators (e.g., "irritation," "fatigue") extracted from text and audio.

[0712] Step 6: Generate Alerts

[0713] The server generates an alert when the stress score exceeds a set threshold. The input is the stress score, and the output is an alert notification and specific mental health care suggestions. Specifically, when the score exceeds the set threshold, an alert message is generated, including suggestions such as "Take a break" or "Consult a mental health professional."

[0714] Step 7: Notification

[0715] The server notifies administrators and employees of the generated alerts. The input is the alert message, and the output is the notification displayed on the smart glasses' screen. Specifically, notification data is sent to the smart glasses via a network protocol, and the notification is sent to the administrator via email or chat application.

[0716] Step 8: Gathering Feedback

[0717] The user (employee) chooses whether to accept or reject the proposed care action. The input is the user's response to the proposal, and the output is feedback data sent to the system. Specifically, the option to "accept" or "reject" is displayed through the smart glasses interface, and the selection result is fed back to the server.

[0718] Step 9: Continuous Monitoring

[0719] The server continuously improves the algorithm and updates threshold settings based on the collected feedback. The input is feedback data, and the output is the updated algorithm parameters. Specifically, the feedback is used to optimize the machine learning model parameters, which are then reflected in the next analysis.

[0720] 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.

[0721] This invention relates to a system for monitoring the mental health of employees working remotely in real time and detecting their stress levels. In particular, by combining it with an emotion engine, it becomes possible to more accurately understand the emotional state of employees and propose appropriate care.

[0722] Data collection

[0723] 1. Terminal: Employees log in to the system using a work-issued PC or mobile device. The system verifies the login information and prepares to collect data.

[0724] 2. Terminal: A script is executed to collect communication data such as employee emails, chats, meeting recordings, and voice memos. This data is stored in a temporary file.

[0725] 3. Terminal: Data stored in temporary files is encrypted and sent to the server using a secure protocol (e.g., HTTPS).

[0726] Data analysis and emotion recognition

[0727] 1. Server: Receives the transferred data and stores it in a secure database. The stored data is preprocessed for analysis (e.g., noise reduction, formatting standardization).

[0728] 2. Server: Applies natural language processing (NLP) algorithms to analyze the collected text and audio data. From the analysis results, it extracts emotional tone, frequency of positive and negative language use, and specific stress signs.

[0729] 3. Server: Furthermore, an emotion engine is used to recognize employees' emotions from communication data. Specifically, the emotion engine analyzes voice and text data to recognize emotions in real time.

[0730] Calculation of emotional score and stress score

[0731] 1. Server: Based on the results of NLP and the emotion engine, it calculates emotion scores and stress scores to assess the mental health status of each employee.

[0732] Alert generation

[0733] 1. Server: Compares the emotion score and stress score with pre-set thresholds and generates an alert if the stress score exceeds the threshold.

[0734] 2. Server: Alerts include specific care suggestions and break suggestions. These alerts are recorded in the logging system.

[0735] Notifications and suggestions

[0736] 1. Server: Sends generated alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[0737] 2. Devices: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a break" may be displayed.

[0738] 3. User: Employees choose whether to accept the proposed mental health care action. The choice is sent to the system as feedback.

[0739] Feedback and continuous monitoring

[0740] 1. Server: Collects feedback from employees and administrators and uses it to make timely improvements to the system's natural language processing algorithms, sentiment engine, and threshold settings.

[0741] 2. Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[0742] Specific example

[0743] This section describes a specific example of monitoring employee B's mental health.

[0744] 1. Device: Collect one week's worth of email and chat data from employee B's PC and securely transfer it to the cloud.

[0745] 2. Server: The server analyzes the collected data using natural language processing technology and an emotion engine. For example, "negative tones" are detected frequently, and emotions such as "anxiety" are identified from the audio data. The stress score becomes 80 (threshold is 70).

[0746] 3. Server: Generate an alert because the stress score exceeds the threshold, and include a suggestion such as "Take a break."

[0747] 4. Server: Send an alert to the administrator and employee B, including specific care suggestions such as "Consult a mental health professional."

[0748] 5. Terminal: A notification will appear on employee B's PC, allowing employee B to submit feedback.

[0749] This system allows for accurate and real-time monitoring of employees' emotional states, even in a remote work environment, and enables the provision of appropriate care.

[0750] The following describes the processing flow.

[0751] Step 1:

[0752] Terminal: An employee logs into the system using a work-issued PC or mobile device. Based on the login information, the system begins preparing to collect data.

[0753] Step 2:

[0754] Terminal: A script automatically runs to collect communication data such as employee emails, chats, meeting recordings, and voice memos. This data is stored in temporary files.

[0755] Step 3:

[0756] Terminal: Encrypts data stored in temporary files and sends it to the server using a secure protocol (e.g., HTTPS).

[0757] Step 4:

[0758] Server: Receives data and stores it in a secure database. Stored data is preprocessed for analysis (e.g., noise reduction, formatting standardization).

[0759] Step 5:

[0760] Server: Applies natural language processing (NLP) algorithms to analyze collected text and audio data. The analysis results include emotional tone, positive and negative expressions, and specific stress signs.

[0761] Step 6:

[0762] Server: Uses an emotion engine to perform real-time emotion recognition from the same communication data. For example, emotions such as "anxiety" and "anger" are detected from voice data, while emotions such as "sadness" and "stress" are identified from text data.

[0763] Step 7:

[0764] Server: Calculates employee emotional scores and stress scores based on the analysis results of NLP and the emotion engine.

[0765] Step 8:

[0766] Server: Compares the emotion score and stress score to the set threshold. If the stress score exceeds the threshold, an alert is generated.

[0767] Step 9:

[0768] Server: The generated alerts include specific care suggestions (e.g., suggestions for taking a break or advice on deep breathing) and links to consult with professionals. These alerts are recorded in the logging system.

[0769] Step 10:

[0770] Server: Sends alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[0771] Step 11:

[0772] Device: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a break" may be displayed.

[0773] Step 12:

[0774] User: Employees can choose whether or not to accept the proposed mental health care action. The choice is sent to the system as feedback.

[0775] Step 13:

[0776] Server: Receives feedback and uses it to make timely improvements to natural language processing algorithms, sentiment engines, and threshold settings.

[0777] Step 14:

[0778] Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[0779] Specific example

[0780] This section describes a specific example of monitoring employee C's mental health.

[0781] Step 1:

[0782] Terminal: Collect one week's worth of email and chat data from employee C's PC and securely transfer it to the cloud.

[0783] Step 2:

[0784] Server: Analyzes data using natural language processing technology and an emotion engine. The analysis reveals a high proportion of negative tones, and emotions such as anxiety are detected from the audio data. A stress score of 85 (threshold is 70) was calculated.

[0785] Step 3:

[0786] Server: Generates an alert such as "Take a break" because the stress score exceeds the threshold. This alert also includes specific suggestions and links to consult with experts.

[0787] Step 4:

[0788] Server: Sends alerts to administrator and employee C.

[0789] Step 5:

[0790] Terminal: A notification appears on employee C's PC, and when employee C submits feedback, the system receives it and improves the algorithm.

[0791] This system allows for accurate and real-time monitoring of employees' emotional states, even in a remote work environment, and enables the provision of appropriate care.

[0792] (Example 2)

[0793] 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".

[0794] In a remote work environment, it is difficult to monitor employees' mental health in real time and accurately grasp their stress levels and emotional fluctuations. Furthermore, there is a lack of means to quickly provide appropriate care suggestions when employees are experiencing excessive stress. In such circumstances, it is difficult to prevent the deterioration of employees' mental health, potentially leading to decreased productivity and increased employee turnover.

[0795] 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.

[0796] In this invention, the server includes means for monitoring the mental health status of employees, means for analyzing employee communication data using natural language processing technology, and means for recognizing employee emotions from text and voice data using an emotion engine. This makes it possible to accurately grasp the mental health status of employees in real time and to quickly provide appropriate care suggestions.

[0797] "Mental health status" refers to an employee's mental and emotional well-being. This includes stress levels and emotional stability.

[0798] "Monitoring methods" refer to technologies and devices for continuously observing and recording employees' mental health status over time.

[0799] "Natural language processing technology" refers to artificial intelligence technology used to understand, analyze, and generate human language. This includes analyzing text data and classifying emotions.

[0800] "Communication data" refers to data obtained from communication methods used by employees during work, such as email, text chat, audio conference recordings, and voice memos.

[0801] An "emotion engine" refers to algorithms and technologies used to analyze and recognize emotions from text and audio data. This often includes generative AI models.

[0802] An "emotion score" is an indicator that numerically represents an employee's emotional state, calculated from the results of analysis by the emotion engine.

[0803] A "stress score" is an indicator that quantifies how much stress an employee is experiencing. It is calculated based on the results of an emotional analysis.

[0804] A "threshold" refers to a pre-set baseline value for emotion and stress scores. An alert is generated when this value is exceeded.

[0805] "Notification methods" refer to technologies and devices used to communicate alerts and care suggestions to administrators and employees themselves. This includes email and dedicated applications.

[0806] "Specific mental health support suggestions" refer to concrete actions and resources to improve employees' mental health. Examples include suggestions for breaks or links to mental health professionals.

[0807] This invention is a system that monitors the mental health of employees working remotely in real time and detects their stress levels. In particular, by combining it with an emotion engine, it is possible to more accurately understand the emotional state of employees and propose appropriate care.

[0808] Data collection

[0809] 1. Terminal: Employees log in to the system using a work PC or mobile device. The system verifies the login information and prepares to collect data.

[0810] 2. Device: After logging in, a script is automatically executed to collect communication data such as emails, chats, meeting recordings, and voice memos. For example, email data is collected from the mailbox by gathering the subject and body text, and chat data is retrieved using an API. Meeting recordings and voice memos are recorded via the device's microphone and saved as temporary files.

[0811] 3. Terminal: The collected data is stored in a temporary file and then encrypted using the AES-256 encryption algorithm. The encrypted data is then sent to the server using the secure HTTPS protocol.

[0812] Data analysis and emotion recognition

[0813] 1. Server: Receives data transferred from terminals and stores it in a secure database. After storage, the data undergoes preprocessing such as noise reduction and formatting standardization.

[0814] 2. Server: Applies natural language processing (NLP) algorithms to analyze text and audio data. For NLP analysis, libraries such as spaCy or NLTK are used. Audio data is converted to text using the Google Cloud Speech-to-Text API and integrated with the text data for analysis.

[0815] 3. Server: Recognizes emotions from text and audio data using an emotion engine. The emotion engine uses generative AI models such as BERT or GPT-3 to perform emotion classification tasks.

[0816] Calculation of emotional score and stress score

[0817] 1. Server: Based on the results of NLP and the emotion engine, it calculates an emotion score and a stress score. A weighted average or rule-based scoring system is used to calculate the scores. A higher emotion score is assigned when there is a high frequency of positive language use, and a higher stress score is assigned when there is a high frequency of negative language use.

[0818] Alert generation

[0819] 1. Server: Compares the emotion score and stress score to a threshold, and generates an alert if the stress score exceeds the threshold. For example, if the threshold is 70 and the stress score is 80, an alert will be generated.

[0820] 2. Server: Alerts include specific care suggestions. These suggestions may include consultation with a psychological counselor, use of relaxation apps, or suggestions for taking a break. These alerts are recorded in the logging system and stored as a dataset that can be used for analysis.

[0821] Notifications and suggestions

[0822] 1. Server: Notifies administrators and employees of generated alerts. Notifications are primarily made via email or a dedicated application (e.g., a company-specific mental health care app).

[0823] 2. Devices: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a 10-minute break" may pop up.

[0824] 3. User: Employees choose whether to accept the proposed care action. The choice is sent to the system as feedback and recorded in the database. This will be reflected in future analyses.

[0825] Feedback and continuous monitoring

[0826] 1. Server: Collects feedback from employees and administrators to improve the system's NLP algorithms, emotion engine, and threshold settings. Feedback is used to improve the accuracy of the emotion recognition model and readjust thresholds.

[0827] 2. Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely. For example, it performs weekly data analysis and provides administrators with monthly reports.

[0828] Specific examples and prompt statements

[0829] As a concrete example of monitoring employee B's mental health, the following operations are performed: Email and chat data for one week are collected from employee B's PC and securely transferred to the cloud. The collected data is analyzed using natural language processing technology and an emotion engine. If negative tones are detected frequently and emotions such as "anxiety" are identified from the voice data, the stress score becomes 80 (the threshold is 70). In this case, an alert is generated and a suggestion such as "Take a break" is made. The alert is sent to the administrator and employee B, and employee B can submit feedback.

[0830] Example prompt: "Collect one week's worth of emails, chats, and meeting recordings from employee B and analyze them using natural language processing and a sentiment engine. If the results show a high proportion of negative tones, generate specific care suggestions and notify the employee."

[0831] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0832] Step 1:

[0833] Terminal: Employees log in to the system using a work PC or mobile device.

[0834] Input: Employee login information (User ID, password)

[0835] Operation: The system authenticates login information and prepares for data collection.

[0836] Output: Login success message, instruction to start data collection script.

[0837] Step 2:

[0838] Terminal: After logging in, a script will run that automatically collects communication data such as emails, chats, meeting recordings, and voice memos.

[0839] Input: Employee account information after login authentication

[0840] Operation: Email data is collected from the subject and body text of employee mailboxes. Chat data is retrieved using an API. Meeting recordings and voice memos are recorded via the device's microphone and stored in temporary files.

[0841] Output: Text data (emails and chats), audio data (meeting recordings and voice memos), temporary files

[0842] Step 3:

[0843] Terminal: The collected data is saved to a temporary file and then encrypted using the AES-256 encryption algorithm.

[0844] Input: Raw data stored in a temporary file

[0845] Operation: Encrypts data using the AES-256 encryption algorithm. Sends the encrypted data to the server using the HTTPS protocol.

[0846] Output: Encrypted data, HTTPS request

[0847] Step 4:

[0848] Server: Receives data transferred from terminals and stores it in a secure database.

[0849] Input: Encrypted data included in the HTTPS request

[0850] Operation: Saves data to a database. After saving, preprocessing is performed on the data, such as noise reduction and formatting standardization.

[0851] Output: Preprocessed data

[0852] Step 5:

[0853] Server: Applies natural language processing (NLP) algorithms to analyze text and audio data.

[0854] Input: Preprocessed text data and audio data

[0855] Operation: For NLP analysis, libraries such as spaCy or NLTK are used to tokenize text data and tag parts of speech. Audio data is converted to text using the Google Cloud Speech-to-Text API and integrated with the text data for analysis.

[0856] Output: Analyzed text data, text data converted from audio data

[0857] Step 6:

[0858] Server: Uses an emotion engine to recognize emotions from text and audio data.

[0859] Input: Analyzed text data, text data converted from audio data

[0860] Operation: The emotion engine uses generative AI models such as BERT or GPT-3 to perform emotion classification tasks. It assigns emotion labels such as positive, negative, and neutral to the data.

[0861] Output: Data with emotion labels

[0862] Step 7:

[0863] Server: Calculates emotion scores and stress scores based on the results of NLP and the emotion engine.

[0864] Input: Data with emotion labels

[0865] Operation: The scoring system uses weighted averages or rule-based scoring systems to calculate scores. A higher sentiment score is assigned when positive language usage is frequent, and a higher stress score is assigned when negative language usage is frequent.

[0866] Output: Recorded sentiment score and stress score

[0867] Step 8:

[0868] Server: Compares the emotion score and stress score to a threshold, and generates an alert if the stress score exceeds the threshold.

[0869] Input: Calculated sentiment score and stress score, threshold setting

[0870] Operation: Determines whether the stress score exceeds a threshold and generates an alert if it does.

[0871] Output: Alert Information

[0872] Step 9:

[0873] Server: Based on alerts recorded in the database, it notifies administrators and employees themselves.

[0874] Input: Alert Information

[0875] Operation: Notifies administrators and employees via email or a dedicated application. Sends messages such as, "Your stress level is high. We recommend you take a break."

[0876] Output: Notification message

[0877] Step 10:

[0878] Device: Notifications will be displayed on employees' PCs and mobile devices.

[0879] Input: Notification message

[0880] Action: A notification message is displayed. A pop-up appears saying something like, "Your stress level is high. We recommend you take a 10-minute break."

[0881] Output: Displayed notification message

[0882] Step 11:

[0883] User: The employee chooses whether to accept the proposed care action.

[0884] Input: Displayed notification message

[0885] Operation: Employees can choose to "accept" or "reject" a notification, and the system receives feedback on their choice.

[0886] Output: Acceptance or rejection feedback information

[0887] Step 12:

[0888] Server: Collects feedback from employees and administrators to improve the system's NLP algorithms, sentiment engine, and threshold settings.

[0889] Input: Feedback information

[0890] Operation: Feedback is recorded in a database and used to adjust algorithms, emotion engines, and threshold settings for future analyses.

[0891] Output: Improved NLP algorithm, emotion engine, threshold settings

[0892] Step 13:

[0893] Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[0894] Input: Continuously collected communication data

[0895] Operation: Performs weekly data analysis. Provides administrators with monthly reports summarizing the data.

[0896] Output: Weekly data analysis report, monthly report

[0897] (Application Example 2)

[0898] 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."

[0899] The mental health of factory workers is susceptible to the effects of harsh working conditions and stress, making appropriate monitoring and care essential. However, current systems lack the ability to grasp workers' stress and emotions in real time and provide appropriate care suggestions. In particular, there is a need to quickly detect changes in emotions and provide workers with effective improvement suggestions. Therefore, there is a demand for a system that efficiently and accurately monitors the mental health of employees working on-site and provides appropriate care suggestions in real time.

[0900] 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.

[0901] In this invention, the server includes means for monitoring the mental health status of employees, means for analyzing employee communication data using natural language processing technology, means for calculating an emotion score and a stress score based on the analysis results, means for notifying managers and the employee themselves when the stress score exceeds a set threshold, means for providing specific mental care suggestions, means for collecting voice and text communications of factory workers, means for securely transferring the collected data, means for notifying workers of alerts using factory robots, head-mounted displays, or smart glasses, and means for collecting feedback from managers and workers and improving the system's algorithms and threshold settings. This makes it possible to accurately monitor the mental health of factory workers in real time and to quickly provide appropriate care suggestions.

[0902] "Employee mental health status" refers to the psychological and emotional health of workers.

[0903] "Monitoring methods" refer to technologies and devices used to monitor the status of employees by collecting and analyzing data in real time or periodically.

[0904] "Natural language processing technology" is a technology that enables computers to analyze, understand, and generate human language.

[0905] "Communication data" refers to data in the form of emails, chats, meeting recordings, voice memos, work reports, conversations, etc., that workers exchange in their daily work.

[0906] "Emotion score" refers to a numerical value that quantitatively represents the emotional state of a worker, calculated using natural language processing and emotion recognition technologies.

[0907] A "stress score" refers to a numerical value that evaluates the stress level of workers based on their emotional scores.

[0908] A "threshold" refers to a value that serves as a benchmark for indicating that an emotional score or stress score has reached a specific state.

[0909] "Means of notification" refers to technologies and devices used to transmit information to relevant parties using alerts or messages.

[0910] "Mental care suggestions" refer to specific advice and action proposals aimed at improving the mental health of workers.

[0911] "Means of collecting voice and text communications" refers to technologies and devices for acquiring voice and text data from workers.

[0912] "Secure transfer" refers to encrypting data and transmitting it using a secure protocol.

[0913] "Factory robots" refer to mechanical devices designed to automatically perform tasks and operations within a factory.

[0914] A "head-mounted display (HMD)" refers to a device that displays visual information when worn on the head.

[0915] "Smart glasses" refer to glasses-type wearable devices that are capable of displaying visual information and collecting data.

[0916] An "alert" refers to a warning message that is sent when a specific situation or condition occurs.

[0917] "Means of collecting feedback" refers to technologies and equipment used to gather opinions and evaluations from workers and managers.

[0918] "Means of improving system algorithms and threshold settings" refers to techniques and processes for optimizing system performance and settings based on feedback and data.

[0919] This invention relates to a system for monitoring the mental health of factory workers in real time and detecting stress levels. In particular, by combining it with an emotion engine, it is possible to accurately grasp the emotional state of workers and propose appropriate care. The following hardware and software are required to implement the system of this invention.

[0920] 1. Hardware

[0921] Factory robot: A robotic device that performs tasks within a factory.

[0922] Head-mounted displays (HMDs): For example, Microsoft HoloLens can be used to display visual information to workers and deliver notifications.

[0923] Smart glasses: For example, wearable devices that use Google Glass to collect voice and text data.

[0924] 2. Software

[0925] TensorFlow: Data analysis using natural language processing (NLP) technology.

[0926] IBM Watson: Used as an emotion engine to calculate emotion scores from collected data.

[0927] AWS: Manage databases and perform secure data transfer in the cloud.

[0928] Python: Creating programs for data preprocessing, analysis, and alert generation.

[0929] The system operates as follows:

[0930] 1. Data Collection

[0931] The system collects voice and text data from the terminal through the worker's HMD or smart glasses. For example, conversations and work reports made by workers during their work are collected as data.

[0932] 2. Data Transfer

[0933] The collected data is stored as a temporary file, encrypted, and securely transmitted to the AWS cloud using the HTTPS protocol.

[0934] 3. Data Analysis

[0935] Data is preprocessed on the cloud by a server, and text data is analyzed using NLP analysis with TensorFlow. In addition, IBM Watson's emotion engine analyzes audio data to calculate emotion scores and stress scores. For example, if many negative emotions such as "tired" or "irritated" are detected during work, the stress score will be high.

[0936] 4. Alert generation and notification

[0937] The server generates alerts based on emotion and stress scores, and sends alert messages containing appropriate care suggestions to the worker's HMD or smart glasses if the threshold is exceeded. For example, if the stress score exceeds the threshold, a message such as "We recommend you take a break" will be displayed.

[0938] 5. Gathering feedback and making improvements

[0939] The server collects feedback from workers and administrators, and uses it to improve the system's algorithms and threshold settings. This feedback allows the system to continuously improve its performance.

[0940] As a concrete example, consider a case where worker A at a factory wears an HMD while working. Daily voice and text data collected from this worker is transferred to the AWS cloud and analyzed by TensorFlow and IBM Watson. If the emotion score indicates "fatigue" or "dissatisfaction," and the stress score exceeds a threshold, an alert is displayed on the HMD. Subsequently, feedback from worker A is sent to the server, and the system is further optimized.

[0941] Examples of prompts to input into a generative AI model are as follows:

[0942] "Develop a system to monitor the mental health of factory workers and provide real-time care suggestions based on their stress levels. This system will use factory robots, HMDs, and smart glasses to collect and analyze employee voice and text data. It should calculate employee emotion and stress scores using natural language processing algorithms and an emotion recognition engine, and then provide appropriate care suggestions."

[0943] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0944] Step 1:

[0945] The terminal collects voice and text data from the worker's head-mounted display (HMD) or smart glasses. The input at this stage is voice and text communication data generated during the worker's daily work. The output is the collected data, which is stored as a temporary file.

[0946] Step 2:

[0947] The device encrypts the collected temporary file data and sends it to the AWS cloud using a secure protocol (e.g., HTTPS). The input at this stage is the unencrypted data stored in the temporary file, and the output is a notification that the transfer to the secure cloud is complete.

[0948] Step 3:

[0949] The server decrypts the data received in the cloud and performs preprocessing (noise reduction, format standardization). The input is encrypted collected data, and the output is preprocessed data. Specifically, background noise is removed from the audio data, and the format of the text data is standardized.

[0950] Step 4:

[0951] The server applies natural language processing (NLP) to preprocessed data to analyze the text data. The software used is TensorFlow; the input is preprocessed text data, and the output is the result of the text analysis. Specifically, it extracts emotional tone and the frequency of positive and negative language use.

[0952] Step 5:

[0953] The server uses an emotion engine (IBM Watson) to analyze voice data and calculate an emotion score. The input is pre-processed voice data, and the output is the emotion score. Specifically, it analyzes the intonation and tone of the voice to determine the emotional state.

[0954] Step 6:

[0955] The server calculates a stress score based on NLP analysis results and emotion engine results. The input is text analysis results and emotion score, and the output is the stress score. Specifically, it integrates the emotional tone of text and voice to assess the employee's stress level.

[0956] Step 7:

[0957] The server checks whether the calculated stress score exceeds a set threshold. The input is the stress score, and the output is a flag that instructs the server to generate an alert if the threshold is exceeded.

[0958] Step 8:

[0959] The server generates an alert when a threshold is exceeded and sends an alert message, including specific care suggestions, to the worker's HMD or smart glasses. The input is the result of the threshold being exceeded, and the output is the alert notification to the worker. Specifically, a message such as "We recommend you take a break" might be displayed.

[0960] Step 9:

[0961] The server collects feedback from workers and managers through HMDs and smart glasses, and uses this feedback to improve the system's algorithms and threshold settings. The input is feedback data, and the output is the improved algorithms and settings. Specifically, it analyzes the feedback and adjusts the algorithm parameters.

[0962] The following example prompts illustrate the specific actions required to perform each step:

[0963] "Develop a system to monitor the mental health of factory workers and provide real-time care suggestions based on their stress levels. This system will use factory robots, HMDs, and smart glasses to collect and analyze employee voice and text data. It should calculate employee emotion and stress scores using natural language processing algorithms and an emotion recognition engine, and then provide appropriate care suggestions."

[0964] 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.

[0965] 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.

[0966] 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.

[0967] [Third Embodiment]

[0968] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0969] 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.

[0970] 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).

[0971] 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.

[0972] 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.

[0973] 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).

[0974] 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.

[0975] 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.

[0976] 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.

[0977] 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.

[0978] 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.

[0979] 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".

[0980] The system of this invention monitors the mental health of employees working remotely in real time and detects alerts for stress levels. Specific embodiments are described below.

[0981] Data collection

[0982] 1. Terminal: Employees log in to the system using a work-issued PC or mobile device. Based on the login information, the system prepares to collect data.

[0983] 2. Terminal: A script starts running to collect communication data such as employee emails, chats, meeting recordings, and voice memos. The collected data is saved to a temporary file.

[0984] 3. Terminal: Communication data stored in temporary files is encrypted and transferred to the server using a secure protocol.

[0985] Data Analysis

[0986] 1. Server: Receives the transferred data and stores it in a secure database. The stored data is pre-processed (e.g., noise reduction, formatting standardization).

[0987] 2. Server: Applies natural language processing algorithms to perform sentiment analysis on text and audio data. The analysis includes emotional tone, frequency of positive and negative language use, and specific stress signs (e.g., frequent use of expressions like "tired" or "irritated").

[0988] 3. Server: Calculates emotional and stress scores to assess the mental health status of each employee.

[0989] Alert generation

[0990] 1. Server: Compares the emotional score and stress score to a threshold and generates an alert if the stress score exceeds the threshold. The alert includes specific care suggestions or suggestions for taking a break.

[0991] 2. Server: Records generated alerts in the logging system for future analysis and feedback.

[0992] Notifications and suggestions

[0993] 1. Server: Sends alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[0994] 2. Devices: Notifications are displayed on employees' PCs and mobile devices, along with a message such as, "Your stress level is high; we recommend you take a break."

[0995] 3. User: Employees can choose to accept or reject the proposed mental health care action. They submit their decision to the system as feedback.

[0996] Feedback and continuous monitoring

[0997] 1. Server: Collects the results of the proposal's implementation and feedback, and records them in the system log.

[0998] 2. Server: Based on the collected feedback, the server will make timely improvements to the natural language processing algorithm and threshold settings.

[0999] 3. Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[1000] Specific example

[1001] This section describes a specific example of monitoring employee A's mental health.

[1002] 1. Device: Collect one week's worth of email and chat data from employee A's PC and securely transfer it to the cloud.

[1003] 2. Server: The server analyzes the data using natural language processing technology and determines that "negative tones" account for 50% of the responses. The stress score is calculated to be 75 (threshold is 70).

[1004] 3. Server: Generates an alert such as "Take a break" because the stress score exceeds the threshold.

[1005] 4. Server: Send an alert to the administrator and employee A, including specific care suggestions such as "Consult a mental health professional."

[1006] 5. Terminal: A notification will appear on employee A's PC, allowing employee A to submit feedback.

[1007] This system allows for proper management of employees' mental health and provision of necessary care, even in a remote work environment.

[1008] The following describes the processing flow.

[1009] Step 1:

[1010] Terminal: An employee logs into the system using a work-issued PC or mobile device. The system verifies the login information and begins preparing to collect data.

[1011] Step 2:

[1012] Terminal: A script runs that collects communication data such as employee emails, chats, meeting recordings, and voice memos. This data is stored in temporary files.

[1013] Step 3:

[1014] Terminal: Data stored in temporary files is encrypted and sent to the server using a secure protocol (e.g., HTTPS).

[1015] Step 4:

[1016] Server: Receives the transferred data and stores it in a secure database. The stored data is preprocessed for analysis (e.g., noise reduction, formatting standardization).

[1017] Step 5:

[1018] Server: Applies natural language processing (NLP) algorithms to analyze collected text and audio data. The analysis extracts emotional tones, the frequency of positive and negative language use, and specific stress signs.

[1019] Step 6:

[1020] Server: Based on the NLP results, it calculates emotional and stress scores to assess the mental health status of each employee.

[1021] Step 7:

[1022] Server: Compares the emotion score and stress score to pre-set thresholds and generates an alert if the stress score exceeds the threshold.

[1023] Step 8:

[1024] Server: Alerts include specific care suggestions and break suggestions. These alerts are recorded in the logging system.

[1025] Step 9:

[1026] Server: Sends generated alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[1027] Step 10:

[1028] Device: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a break" may be displayed.

[1029] Step 11:

[1030] User: Employees choose whether to accept the proposed mental health care action. The choice is sent to the system as feedback.

[1031] Step 12:

[1032] Server: Collects feedback from employees and administrators and uses it to make timely improvements to the system's natural language processing algorithms and threshold settings.

[1033] Step 13:

[1034] Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[1035] (Example 1)

[1036] 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."

[1037] With the widespread adoption of remote work, managing employees' mental health, in addition to monitoring work progress, has become a crucial issue. However, traditional methods make it difficult to grasp employees' stress levels and emotional states in real time and provide appropriate countermeasures, often resulting in mental health problems being left unaddressed. This leads to problems such as decreased work efficiency and increased employee turnover, making an effective mental health monitoring system necessary.

[1038] 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.

[1039] In this invention, the server includes means for monitoring the mental health status of employees in real time, means for collecting and analyzing employee communication data using natural language processing technology, and means for calculating and evaluating emotion scores and stress scores based on the analysis results. This makes it possible to effectively monitor the mental health of employees and provide timely and appropriate alerts and care suggestions.

[1040] "Mental health" refers to an individual's psychological, emotional, and social well-being. Mental health is a crucial factor that influences one's ability to cope with everyday stress, manage work and relationships, and more.

[1041] "Real-time monitoring" is a process of monitoring and evaluating data and conditions almost instantly. This process allows for a rapid response when problems occur.

[1042] "Natural language processing technology" refers to the technology that enables computers to understand and process human language. This includes text analysis and sentiment assessment.

[1043] "Communication data" refers to information such as emails, chats, meeting recordings, and voice memos exchanged by employees in the course of their work. This information is important data for evaluating the mental state of employees.

[1044] An "emotion score" is a numerical representation of an employee's emotional state, based on data analyzed using natural language processing technology. The score is used to evaluate emotions such as positive and negative.

[1045] A "stress score" is a numerical representation of an employee's stress level, based on data analyzed using natural language processing technology. This score is used to assess the mental burden on employees.

[1046] A "threshold" is a set score limit that triggers a specific alert or action. If the stress score exceeds this value, the system will take a specific action.

[1047] "Alert generation" is the process by which a system issues a warning or notification when set conditions or thresholds are met. This allows relevant parties to take immediate action.

[1048] "Notifying administrators and employees" refers to the process of sending analysis results and alert information to administrators' and employees' devices. Notifications are sent via methods such as email or applications.

[1049] "Mental health suggestions" refer to specific advice and action plans provided to improve employees' mental health. Examples include suggestions for breaks or consultations with mental health professionals.

[1050] "Collecting feedback" is the process by which employees submit to the system the results and opinions they have had regarding the mental health care suggestions they have received. The system then uses this feedback to make improvements.

[1051] "Improving analysis algorithms and threshold settings" refers to using collected feedback data to promptly update the system's analysis methods and the conditions for triggering actions to the latest and most optimal state.

[1052] The system of this invention monitors the mental health of employees working remotely in real time and detects alerts for stress levels. Specific embodiments are described below.

[1053] System Overview

[1054] This system uses natural language processing technology to collect and analyze employee communication data in order to monitor employees' mental health status in real time, and to calculate and evaluate emotion scores and stress scores. The server generates alerts based on stress scores that exceed a threshold and notifies both administrators and the employees themselves. It also provides specific mental care suggestions and collects employee feedback to improve the analysis algorithms and threshold settings.

[1055] Data collection

[1056] When an employee logs into the system using their work PC or mobile device, they enter their login information (e.g., user ID, password), and the system prepares to collect data. A script automatically starts on the terminal, collecting communication data such as emails, chats, meeting recordings, and voice memos, and saving it to a temporary file. The collected data is encrypted using AES (Advanced Encryption Standard) or similar and transferred to a secure server using the SSL / TLS protocol.

[1057] Data Analysis

[1058] The server receives the transferred data and stores it in a secure database such as AWS RDS. The stored data undergoes a preprocessing process (e.g., noise reduction, formatting standardization) and then sentiment analysis is performed using natural language processing algorithms (e.g., GPT-3, Python's NLTK library). This analysis calculates each employee's sentiment score (0-100) and stress score (1-10). This allows for an assessment of each employee's mental health status.

[1059] Alert generation and notification

[1060] The server compares the emotion score and stress score to a threshold (e.g., stress score of 70 or higher) and generates an alert if the threshold is exceeded. The generated alert is recorded in a logging system (e.g., MongoDB) and notified to the administrator and the employee. The notification includes specific mental health care suggestions (e.g., "Take a break," "Consult a mental health professional") via email (e.g., Amazon SES) or a dedicated application.

[1061] Employee feedback and system improvements

[1062] Employees, as users, choose to accept or reject proposed mental health care actions and provide feedback to the system. The server collects this feedback information and records it in the system log (e.g., Redis). Based on the collected feedback, the analysis algorithms and threshold settings are improved.

[1063] Continuous monitoring

[1064] The server periodically collects, analyzes, and generates alerts to continuously monitor employees' mental health. This makes it possible to properly manage employee mental health and provide necessary care, even in a remote work environment.

[1065] Specific example

[1066] For example, employee A collects a week's worth of email and chat data from their PC and securely transfers it to the cloud. The server then analyzes the data using natural language processing technology and determines that 50% of it has a "negative tone." If the stress score is calculated to be 75 (the threshold is 70), the system generates an alert such as "Take a break" and notifies both the administrator and employee A. It also includes specific care suggestions such as "Consult a mental health professional." Finally, employee A can choose whether to take the suggested action and provide feedback on the result.

[1067] Prompt example

[1068] "Create a program that uses employee email and chat data to monitor their mental health in real time. It should generate an alert and send a notification if the stress score exceeds a threshold. The technologies to use are the Python NLTK library and GPT-3."

[1069] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1070] Step 1:

[1071] Terminal: Employees log in to the system using a work-issued PC or mobile device.

[1072] Input: Authentication information such as user ID and password.

[1073] Data processing: Authentication processing of login information.

[1074] Output: A message indicating successful or failed login. The specific action involves entering a user ID and password and sending a request to the system's authentication server.

[1075] Step 2:

[1076] Terminal: Upon successful login, the script will automatically launch to collect communication data such as employee emails, chats, meeting recordings, and voice memos.

[1077] Input: Employee communication data (emails, chat logs, voice memos, etc.).

[1078] Data processing: Data collection and saving to temporary files.

[1079] Output: Communication data saved in a temporary file. Specifically, data is retrieved from each application (Outlook, Teams, etc.) using APIs and saved to a temporary file.

[1080] Step 3:

[1081] Terminal: Encrypts data stored in temporary files using AES or similar encryption methods and transfers it to the server using the SSL / TLS protocol.

[1082] Input: Communication data stored in a temporary file.

[1083] Data processing: Data encryption and secure transfer.

[1084] Output: An encrypted temporary file is transferred to the server. Specifically, the data is encrypted using an encryption library and uploaded to the server via HTTPS.

[1085] Step 4:

[1086] Server: Receives the transferred data and stores it in a secure database.

[1087] Input: Encrypted data that arrived on the server.

[1088] Data processing: Decrypting data and saving it to a database.

[1089] Output: Communication data stored in a secure database. Specifically, the data is decrypted and inserted into a database (e.g., AWS RDS).

[1090] Step 5:

[1091] Server: Preprocesses the stored data (noise reduction, format standardization).

[1092] Input: Communication data stored in the database.

[1093] Data processing: Data cleaning and formatting standardization.

[1094] Output: Preprocessed dataset. Specifically, data cleaning is performed using a script such as Python to remove noise and unwanted parts.

[1095] Step 6:

[1096] Server: Applies natural language processing algorithms to perform sentiment analysis.

[1097] Input: Preprocessed dataset.

[1098] Data processing: Sentiment analysis of text and audio data.

[1099] Output: Sentiment score and stress score. Specifically, the positive and negative sentiment of each document is analyzed using GPT-3 and the Python NLTK library.

[1100] Step 7:

[1101] Server: Evaluates each employee's mental health status based on the calculated emotion score and stress score.

[1102] Input: Emotion score and stress score.

[1103] Data processing: Assessment of mental health status.

[1104] Output: Evaluation result (normal, caution needed, dangerous, etc.). Specifically, the system compares the score with a threshold and classifies the state.

[1105] Step 8:

[1106] Server: Generates an alert when the stress score exceeds a threshold.

[1107] Input: Stress score for each employee.

[1108] Data processing: Comparison with thresholds and alert generation.

[1109] Output: Alert message. Specifically, if the stress score exceeds a specified threshold, an alert message such as "Take a break" is generated.

[1110] Step 9:

[1111] Server: Logs generated alerts and notifies administrators and employees themselves.

[1112] Input: Alert message.

[1113] Data processing: logging and sending notifications.

[1114] Output: Log entries and notification messages. Specifically, alert information is recorded in a logging system such as MongoDB, and administrators and employees are notified via email or a dedicated application.

[1115] Step 10:

[1116] User: Employees receive a notification and can choose whether to take the suggested mental health care action.

[1117] Input: Alert notification.

[1118] Data processing: Implementation and feedback of mental care actions.

[1119] Output: Feedback data. Specifically, an employee clicks the "Take Break" button and sends the result to the system.

[1120] Step 11:

[1121] Server: Collects feedback data and records it in the system log.

[1122] Input: Employee feedback.

[1123] Data processing: Recording feedback.

[1124] Output: Log entries. Specifically, feedback information is saved to a database such as Redis.

[1125] Step 12:

[1126] Server: Based on the collected feedback data, improve the analysis algorithm and threshold settings.

[1127] Input: Feedback data.

[1128] Data processing: Updating algorithms and settings.

[1129] Output: Improved analysis algorithm and threshold settings. Specifically, the parameters of the machine learning algorithm are readjusted based on the collected feedback.

[1130] Step 13:

[1131] Server: Regularly collects, analyzes, and generates alerts to continuously monitor employees' mental health.

[1132] Input: Communication data and feedback data from all employees.

[1133] Data processing: Iterative monitoring and alert generation process.

[1134] Output: Continuously updated mental health status and alerts. Specifically, the system periodically runs scripts to collect and analyze new data to monitor employee mental health.

[1135] (Application Example 1)

[1136] 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."

[1137] Monitoring the mental health of employees working remotely presents a challenge. In particular, there is a lack of technology to detect stress and fatigue in real time from employee voice and facial expressions, and to provide effective support. This can increase employee health risks and reduce work efficiency.

[1138] 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.

[1139] In this invention, the server includes means for monitoring the mental health status of employees, means for analyzing employee communication data using natural language processing technology, means for collecting voice and visual data using sensors in smart glasses, means for analyzing the employee's stress level from the collected visual and voice data, means for calculating an emotional score and a stress score based on the analysis results, means for notifying the administrator and the employee themselves when the stress score exceeds a set threshold, and means for providing specific mental care suggestions. This makes it possible to monitor the mental health status of employees in real time and provide appropriate care quickly.

[1140] An "employee" refers to an individual who belongs to an organization or company and engages in its operations.

[1141] "Means of monitoring mental health status" refers to techniques or methods for assessing and tracking the psychological and emotional health status of employees.

[1142] "Natural language processing technology" refers to computer technologies and algorithms used to understand and analyze human language.

[1143] "Communication data" refers to information collected in the form of emails, chats, meeting recordings, voice memos, and other similar formats.

[1144] "Smart glasses" are wearable devices that are used for data collection and analysis, and are capable of augmented reality and voice control.

[1145] "Means for collecting audio and visual data using sensors" refers to methods of acquiring audio and video information using devices such as microphones and cameras built into smart glasses.

[1146] "Methods for analyzing stress levels" refers to technologies that analyze collected audio and visual data to evaluate the stress levels of employees.

[1147] "Means for calculating emotional and stress scores" refers to techniques or methods for quantifying and evaluating indicators related to emotions and stress.

[1148] A "threshold" refers to a numerical value or condition that serves as the basis for generating an alert.

[1149] "Means of notification" refers to technologies and methods for communicating alerts and information to administrators and employees themselves.

[1150] "Mental health suggestions" refer to specific actions and advice recommended to improve employees' mental health.

[1151] A "server" refers to a computer system that handles data processing and storage.

[1152] The system of this invention monitors employees' mental health in real time, detects stress levels, and provides appropriate care. Specific embodiments are described below.

[1153] Data collection

[1154] Terminal (Smart Glasses): When an employee wears smart glasses, the device collects audio and visual data. The microphone and camera built into the smart glasses record the employee's voice and facial expressions in real time and temporarily store them in storage.

[1155] Data transfer and encryption

[1156] Device (smart glasses): The collected data is encrypted and transferred to the server using a secure protocol (e.g., TLS).

[1157] Data Analysis

[1158] Server: The server receives the transferred data and stores it in a secure database (e.g., PostgreSQL). The data is then preprocessed, and audio data is converted to text using the Google Speech-to-Text API. For visual data, sentiment analysis is performed using an emotion recognition API (e.g., Amazon Comprehend).

[1159] Stress level analysis and score calculation

[1160] Server: Analyzes employee stress levels from collected audio and visual data. Natural language processing techniques and speech analysis libraries are used for the analysis. Based on the analysis results, emotion scores and stress scores are calculated.

[1161] Alert generation

[1162] Server: Generates an alert when the stress score exceeds a set threshold. The alert includes specific mental health care suggestions such as "Take a break" or "Consult a mental health professional."

[1163] notification

[1164] Server: Generated alerts are notified to both the administrator and the employee themselves. The notification is displayed on the smart glasses' screen, visually indicating the content of the alert.

[1165] Feedback and continuous monitoring

[1166] User: Employees can choose to accept or reject proposed care actions. The choice is sent to the system as feedback. This allows the system to continuously monitor the employee's mental health status and improve the analysis algorithms and threshold settings as needed.

[1167] Specific example: Stress detection during meetings

[1168] While employee B is participating in an important meeting, smart glasses analyze B's voice and facial expressions in real time. If the analysis determines that B is experiencing high stress levels, the smart glasses display a message saying, "Take a deep breath and relax." After the meeting, B is suggested to take a break, and if necessary, recommended to consult a mental health professional.

[1169] Prompt example:

[1170] Prompt text to input to the generative AI model:

[1171] "Please describe in detail the design of a smart glasses application that monitors employee stress levels in real time and suggests necessary care. Specifically, explain how voice and visual data will be collected, analyzed, and alerts will be generated. Also, describe in detail how employee feedback will be provided to the system regarding whether they accept the suggestions."

[1172] While the embodiments for carrying out the present invention have been described in detail, the invention is not limited thereto, and various modifications and improvements are possible.

[1173] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1174] Step 1: Data Collection

[1175] The device (smart glasses) collects audio and visual data when worn by an employee. Input is data from the microphone and camera built into the smart glasses, and output is raw audio and visual data temporarily stored in storage. Specifically, audio data is captured from the microphone every second, and visual data is periodically captured by the camera as snapshots and videos.

[1176] Step 2: Data encryption and transfer

[1177] The device (smart glasses) encrypts the collected audio and visual data and transfers it to the server via a secure protocol. The input is unencrypted collected data, and the output is encrypted data packets. Specifically, secure protocols such as TLS are used to protect the data from hacking and unauthorized access.

[1178] Step 3: Convert speech to text

[1179] The server receives the transferred data and stores it in a database. It then converts the audio data to text using the Google Speech-to-Text API. The input is encrypted audio data, and the output is text data. Specifically, an audio file is sent via an API call and returned as text.

[1180] Step 4: Emotion Analysis

[1181] The server uses text and visual data to perform sentiment analysis with an emotion recognition API (e.g., Amazon Comprehend). The input is text and image data, and the output is sentiment scores. Specifically, text data is sent to the sentiment recognition API, which then returns positive, negative, and neutral sentiment scores. Visual data is analyzed using a face recognition algorithm, and sentiment scores are similarly generated.

[1182] Step 5: Stress level analysis and score calculation

[1183] The server analyzes employee stress levels from collected visual and audio data. Inputs are emotion scores and features from audio data, while output is a stress score. Specifically, natural language processing techniques and speech analysis libraries are used to comprehensively evaluate stress indicators (e.g., "irritation," "fatigue") extracted from text and audio.

[1184] Step 6: Generate Alerts

[1185] The server generates an alert when the stress score exceeds a set threshold. The input is the stress score, and the output is an alert notification and specific mental health care suggestions. Specifically, when the score exceeds the set threshold, an alert message is generated, including suggestions such as "Take a break" or "Consult a mental health professional."

[1186] Step 7: Notification

[1187] The server notifies administrators and employees of the generated alerts. The input is the alert message, and the output is the notification displayed on the smart glasses' screen. Specifically, notification data is sent to the smart glasses via a network protocol, and the notification is sent to the administrator via email or chat application.

[1188] Step 8: Gathering Feedback

[1189] The user (employee) chooses whether to accept or reject the proposed care action. The input is the user's response to the proposal, and the output is feedback data sent to the system. Specifically, the option to "accept" or "reject" is displayed through the smart glasses interface, and the selection result is fed back to the server.

[1190] Step 9: Continuous Monitoring

[1191] The server continuously improves the algorithm and updates threshold settings based on the collected feedback. The input is feedback data, and the output is the updated algorithm parameters. Specifically, the feedback is used to optimize the machine learning model parameters, which are then reflected in the next analysis.

[1192] 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.

[1193] This invention relates to a system for monitoring the mental health of employees working remotely in real time and detecting their stress levels. In particular, by combining it with an emotion engine, it becomes possible to more accurately understand the emotional state of employees and propose appropriate care.

[1194] Data collection

[1195] 1. Terminal: Employees log in to the system using a work-issued PC or mobile device. The system verifies the login information and prepares to collect data.

[1196] 2. Terminal: A script is executed to collect communication data such as employee emails, chats, meeting recordings, and voice memos. This data is stored in temporary files.

[1197] 3. Terminal: Data stored in temporary files is encrypted and sent to the server using a secure protocol (e.g., HTTPS).

[1198] Data analysis and emotion recognition

[1199] 1. Server: Receives the transferred data and stores it in a secure database. The stored data is preprocessed for analysis (e.g., noise reduction, formatting standardization).

[1200] 2. Server: Applies natural language processing (NLP) algorithms to analyze the collected text and audio data. From the analysis results, it extracts emotional tone, frequency of positive and negative language use, and specific stress signs.

[1201] 3. Server: Furthermore, an emotion engine is used to recognize employees' emotions from communication data. Specifically, the emotion engine analyzes voice and text data to recognize emotions in real time.

[1202] Calculation of emotional score and stress score

[1203] 1. Server: Based on the results of NLP and the emotion engine, it calculates emotion scores and stress scores to assess the mental health status of each employee.

[1204] Alert generation

[1205] 1. Server: Compares the emotion score and stress score with pre-set thresholds and generates an alert if the stress score exceeds the threshold.

[1206] 2. Server: Alerts include specific care suggestions and break suggestions. These alerts are recorded in the logging system.

[1207] Notifications and suggestions

[1208] 1. Server: Sends generated alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[1209] 2. Devices: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a break" may be displayed.

[1210] 3. User: Employees choose whether to accept the proposed mental health care action. The choice is sent to the system as feedback.

[1211] Feedback and continuous monitoring

[1212] 1. Server: Collects feedback from employees and administrators and uses it to make timely improvements to the system's natural language processing algorithms, sentiment engine, and threshold settings.

[1213] 2. Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[1214] Specific example

[1215] This section describes a specific example of monitoring employee B's mental health.

[1216] 1. Device: Collect one week's worth of email and chat data from employee B's PC and securely transfer it to the cloud.

[1217] 2. Server: The server analyzes the collected data using natural language processing technology and an emotion engine. For example, "negative tones" are detected frequently, and emotions such as "anxiety" are identified from the audio data. The stress score becomes 80 (threshold is 70).

[1218] 3. Server: Generate an alert because the stress score exceeds the threshold, and include a suggestion such as "Take a break."

[1219] 4. Server: Send an alert to the administrator and employee B, including specific care suggestions such as "Consult a mental health professional."

[1220] 5. Terminal: A notification will appear on employee B's PC, allowing employee B to submit feedback.

[1221] This system allows for accurate and real-time monitoring of employees' emotional states, even in a remote work environment, and enables the provision of appropriate care.

[1222] The following describes the processing flow.

[1223] Step 1:

[1224] Terminal: An employee logs into the system using a work-issued PC or mobile device. Based on the login information, the system begins preparing to collect data.

[1225] Step 2:

[1226] Terminal: A script automatically runs to collect communication data such as employee emails, chats, meeting recordings, and voice memos. This data is stored in temporary files.

[1227] Step 3:

[1228] Terminal: Encrypts data stored in temporary files and sends it to the server using a secure protocol (e.g., HTTPS).

[1229] Step 4:

[1230] Server: Receives data and stores it in a secure database. Stored data is preprocessed for analysis (e.g., noise reduction, formatting standardization).

[1231] Step 5:

[1232] Server: Applies natural language processing (NLP) algorithms to analyze collected text and audio data. The analysis results include emotional tone, positive and negative expressions, and specific stress signs.

[1233] Step 6:

[1234] Server: Uses an emotion engine to perform real-time emotion recognition from the same communication data. For example, emotions such as "anxiety" and "anger" are detected from voice data, while emotions such as "sadness" and "stress" are identified from text data.

[1235] Step 7:

[1236] Server: Calculates employee emotional scores and stress scores based on the analysis results of NLP and the emotion engine.

[1237] Step 8:

[1238] Server: Compares the emotion score and stress score to the set threshold. If the stress score exceeds the threshold, an alert is generated.

[1239] Step 9:

[1240] Server: The generated alerts include specific care suggestions (e.g., suggestions for taking a break or advice on deep breathing) and links to consult with professionals. These alerts are recorded in the logging system.

[1241] Step 10:

[1242] Server: Sends alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[1243] Step 11:

[1244] Device: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a break" may be displayed.

[1245] Step 12:

[1246] User: Employees can choose whether or not to accept the proposed mental health care action. The choice is sent to the system as feedback.

[1247] Step 13:

[1248] Server: Receives feedback and uses it to make timely improvements to natural language processing algorithms, sentiment engines, and threshold settings.

[1249] Step 14:

[1250] Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[1251] Specific example

[1252] This section describes a specific example of monitoring employee C's mental health.

[1253] Step 1:

[1254] Terminal: Collect one week's worth of email and chat data from employee C's PC and securely transfer it to the cloud.

[1255] Step 2:

[1256] Server: Analyzes data using natural language processing technology and an emotion engine. The analysis reveals a high proportion of negative tones, and emotions such as anxiety are detected from the audio data. A stress score of 85 (threshold is 70) was calculated.

[1257] Step 3:

[1258] Server: Generates an alert such as "Take a break" because the stress score exceeds the threshold. This alert also includes specific suggestions and links to consult with experts.

[1259] Step 4:

[1260] Server: Sends alerts to administrator and employee C.

[1261] Step 5:

[1262] Terminal: A notification appears on employee C's PC, and when employee C submits feedback, the system receives it and improves the algorithm.

[1263] This system allows for accurate and real-time monitoring of employees' emotional states, even in a remote work environment, and enables the provision of appropriate care.

[1264] (Example 2)

[1265] 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."

[1266] In a remote work environment, it is difficult to monitor employees' mental health in real time and accurately grasp their stress levels and emotional fluctuations. Furthermore, there is a lack of means to quickly provide appropriate care suggestions when employees are experiencing excessive stress. In such circumstances, it is difficult to prevent the deterioration of employees' mental health, potentially leading to decreased productivity and increased employee turnover.

[1267] 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.

[1268] In this invention, the server includes means for monitoring the mental health status of employees, means for analyzing employee communication data using natural language processing technology, and means for recognizing employee emotions from text and voice data using an emotion engine. This makes it possible to accurately grasp the mental health status of employees in real time and to quickly provide appropriate care suggestions.

[1269] "Mental health status" refers to an employee's mental and emotional well-being. This includes stress levels and emotional stability.

[1270] "Monitoring methods" refer to technologies and devices for continuously observing and recording employees' mental health status over time.

[1271] "Natural language processing technology" refers to artificial intelligence technology used to understand, analyze, and generate human language. This includes analyzing text data and classifying emotions.

[1272] "Communication data" refers to data obtained from communication methods used by employees during work, such as email, text chat, audio conference recordings, and voice memos.

[1273] An "emotion engine" refers to algorithms and technologies used to analyze and recognize emotions from text and audio data. This often includes generative AI models.

[1274] An "emotion score" is an indicator that numerically represents an employee's emotional state, calculated from the results of analysis by the emotion engine.

[1275] A "stress score" is an indicator that quantifies how much stress an employee is experiencing. It is calculated based on the results of an emotional analysis.

[1276] A "threshold" refers to a pre-set baseline value for emotion and stress scores. An alert is generated when this value is exceeded.

[1277] "Notification methods" refer to technologies and devices used to communicate alerts and care suggestions to administrators and employees themselves. This includes email and dedicated applications.

[1278] "Specific mental health support suggestions" refer to concrete actions and resources to improve employees' mental health. Examples include suggestions for breaks or links to mental health professionals.

[1279] This invention is a system that monitors the mental health of employees working remotely in real time and detects their stress levels. In particular, by combining it with an emotion engine, it is possible to more accurately understand the emotional state of employees and propose appropriate care.

[1280] Data collection

[1281] 1. Terminal: Employees log in to the system using a work PC or mobile device. The system verifies the login information and prepares to collect data.

[1282] 2. Device: After logging in, a script is automatically executed to collect communication data such as emails, chats, meeting recordings, and voice memos. For example, email data is collected from the mailbox by gathering the subject and body text, and chat data is retrieved using an API. Meeting recordings and voice memos are recorded via the device's microphone and saved as temporary files.

[1283] 3. Terminal: The collected data is stored in a temporary file and then encrypted using the AES-256 encryption algorithm. The encrypted data is then sent to the server using the secure HTTPS protocol.

[1284] Data analysis and emotion recognition

[1285] 1. Server: Receives data transferred from terminals and stores it in a secure database. After storage, the data undergoes preprocessing such as noise reduction and formatting standardization.

[1286] 2. Server: Applies natural language processing (NLP) algorithms to analyze text and audio data. For NLP analysis, libraries such as spaCy or NLTK are used. Audio data is converted to text using the Google Cloud Speech-to-Text API and integrated with the text data for analysis.

[1287] 3. Server: Recognizes emotions from text and audio data using an emotion engine. The emotion engine uses generative AI models such as BERT or GPT-3 to perform emotion classification tasks.

[1288] Calculation of emotional score and stress score

[1289] 1. Server: Based on the results of NLP and the emotion engine, it calculates an emotion score and a stress score. A weighted average or rule-based scoring system is used to calculate the scores. A higher emotion score is assigned when there is a high frequency of positive language use, and a higher stress score is assigned when there is a high frequency of negative language use.

[1290] Alert generation

[1291] 1. Server: Compares the emotion score and stress score to a threshold, and generates an alert if the stress score exceeds the threshold. For example, if the threshold is 70 and the stress score is 80, an alert will be generated.

[1292] 2. Server: Alerts include specific care suggestions. These suggestions may include consultation with a psychological counselor, use of relaxation apps, or suggestions for taking a break. These alerts are recorded in the logging system and stored as a dataset that can be used for analysis.

[1293] Notifications and suggestions

[1294] 1. Server: Notifies administrators and employees of generated alerts. Notifications are primarily made via email or a dedicated application (e.g., a company-specific mental health care app).

[1295] 2. Devices: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a 10-minute break" may pop up.

[1296] 3. User: Employees choose whether to accept the proposed care action. The choice is sent to the system as feedback and recorded in the database. This will be reflected in future analyses.

[1297] Feedback and continuous monitoring

[1298] 1. Server: Collects feedback from employees and administrators to improve the system's NLP algorithms, emotion engine, and threshold settings. Feedback is used to improve the accuracy of the emotion recognition model and readjust thresholds.

[1299] 2. Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely. For example, it performs weekly data analysis and provides administrators with monthly reports.

[1300] Specific examples and prompt statements

[1301] As a concrete example of monitoring employee B's mental health, the following operations are performed: Email and chat data for one week are collected from employee B's PC and securely transferred to the cloud. The collected data is analyzed using natural language processing technology and an emotion engine. If negative tones are detected frequently and emotions such as "anxiety" are identified from the voice data, the stress score becomes 80 (the threshold is 70). In this case, an alert is generated and a suggestion such as "Take a break" is made. The alert is sent to the administrator and employee B, and employee B can submit feedback.

[1302] Example prompt: "Collect one week's worth of emails, chats, and meeting recordings from employee B and analyze them using natural language processing and a sentiment engine. If the results show a high proportion of negative tones, generate specific care suggestions and notify the employee."

[1303] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1304] Step 1:

[1305] Terminal: Employees log in to the system using a work PC or mobile device.

[1306] Input: Employee login information (User ID, password)

[1307] Operation: The system authenticates login information and prepares for data collection.

[1308] Output: Login success message, instruction to start data collection script.

[1309] Step 2:

[1310] Terminal: After logging in, a script will run that automatically collects communication data such as emails, chats, meeting recordings, and voice memos.

[1311] Input: Employee account information after login authentication

[1312] Operation: Email data is collected from the subject and body text of employee mailboxes. Chat data is retrieved using an API. Meeting recordings and voice memos are recorded via the device's microphone and stored in temporary files.

[1313] Output: Text data (emails and chats), audio data (meeting recordings and voice memos), temporary files

[1314] Step 3:

[1315] Terminal: The collected data is saved to a temporary file and then encrypted using the AES-256 encryption algorithm.

[1316] Input: Raw data stored in a temporary file

[1317] Operation: Encrypts data using the AES-256 encryption algorithm. Sends the encrypted data to the server using the HTTPS protocol.

[1318] Output: Encrypted data, HTTPS request

[1319] Step 4:

[1320] Server: Receives data transferred from terminals and stores it in a secure database.

[1321] Input: Encrypted data included in the HTTPS request

[1322] Operation: Saves data to a database. After saving, preprocessing is performed on the data, such as noise reduction and formatting standardization.

[1323] Output: Preprocessed data

[1324] Step 5:

[1325] Server: Applies natural language processing (NLP) algorithms to analyze text and audio data.

[1326] Input: Preprocessed text data and audio data

[1327] Operation: For NLP analysis, libraries such as spaCy or NLTK are used to tokenize text data and tag parts of speech. Audio data is converted to text using the Google Cloud Speech-to-Text API and integrated with the text data for analysis.

[1328] Output: Analyzed text data, text data converted from audio data

[1329] Step 6:

[1330] Server: Uses an emotion engine to recognize emotions from text and audio data.

[1331] Input: Analyzed text data, text data converted from audio data

[1332] Operation: The emotion engine uses generative AI models such as BERT or GPT-3 to perform emotion classification tasks. It assigns emotion labels such as positive, negative, and neutral to the data.

[1333] Output: Data with emotion labels

[1334] Step 7:

[1335] Server: Calculates emotion scores and stress scores based on the results of NLP and the emotion engine.

[1336] Input: Data with emotion labels

[1337] Operation: The scoring system uses weighted averages or rule-based scoring systems to calculate scores. A higher sentiment score is assigned when positive language usage is frequent, and a higher stress score is assigned when negative language usage is frequent.

[1338] Output: Recorded sentiment score and stress score

[1339] Step 8:

[1340] Server: Compares the emotion score and stress score to a threshold, and generates an alert if the stress score exceeds the threshold.

[1341] Input: Calculated sentiment score and stress score, threshold setting

[1342] Operation: Determines whether the stress score exceeds a threshold and generates an alert if it does.

[1343] Output: Alert Information

[1344] Step 9:

[1345] Server: Based on alerts recorded in the database, it notifies administrators and employees themselves.

[1346] Input: Alert Information

[1347] Operation: Notifies administrators and employees via email or a dedicated application. Sends messages such as, "Your stress level is high. We recommend you take a break."

[1348] Output: Notification message

[1349] Step 10:

[1350] Device: Notifications will be displayed on employees' PCs and mobile devices.

[1351] Input: Notification message

[1352] Action: A notification message is displayed. A pop-up appears saying something like, "Your stress level is high. We recommend you take a 10-minute break."

[1353] Output: Displayed notification message

[1354] Step 11:

[1355] User: The employee chooses whether to accept the proposed care action.

[1356] Input: Displayed notification message

[1357] Operation: Employees can choose to "accept" or "reject" a notification, and the system receives feedback on their choice.

[1358] Output: Acceptance or rejection feedback information

[1359] Step 12:

[1360] Server: Collects feedback from employees and administrators to improve the system's NLP algorithms, sentiment engine, and threshold settings.

[1361] Input: Feedback information

[1362] Operation: Feedback is recorded in a database and used to adjust algorithms, emotion engines, and threshold settings for future analyses.

[1363] Output: Improved NLP algorithm, emotion engine, threshold settings

[1364] Step 13:

[1365] Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[1366] Input: Continuously collected communication data

[1367] Operation: Performs weekly data analysis. Provides administrators with monthly reports summarizing the data.

[1368] Output: Weekly data analysis report, monthly report

[1369] (Application Example 2)

[1370] 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."

[1371] The mental health of factory workers is susceptible to the effects of harsh working conditions and stress, making appropriate monitoring and care essential. However, current systems lack the ability to grasp workers' stress and emotions in real time and provide appropriate care suggestions. In particular, there is a need to quickly detect changes in emotions and provide workers with effective improvement suggestions. Therefore, there is a demand for a system that efficiently and accurately monitors the mental health of employees working on-site and provides appropriate care suggestions in real time.

[1372] 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.

[1373] In this invention, the server includes means for monitoring the mental health status of employees, means for analyzing employee communication data using natural language processing technology, means for calculating an emotion score and a stress score based on the analysis results, means for notifying managers and the employee themselves when the stress score exceeds a set threshold, means for providing specific mental care suggestions, means for collecting voice and text communications of factory workers, means for securely transferring the collected data, means for notifying workers of alerts using factory robots, head-mounted displays, or smart glasses, and means for collecting feedback from managers and workers and improving the system's algorithms and threshold settings. This makes it possible to accurately monitor the mental health of factory workers in real time and to quickly provide appropriate care suggestions.

[1374] "Employee mental health status" refers to the psychological and emotional health of workers.

[1375] "Monitoring methods" refer to technologies and devices used to monitor the status of employees by collecting and analyzing data in real time or periodically.

[1376] "Natural language processing technology" is a technology that enables computers to analyze, understand, and generate human language.

[1377] "Communication data" refers to data in the form of emails, chats, meeting recordings, voice memos, work reports, conversations, etc., that workers exchange in their daily work.

[1378] "Emotion score" refers to a numerical value that quantitatively represents the emotional state of a worker, calculated using natural language processing and emotion recognition technologies.

[1379] A "stress score" refers to a numerical value that evaluates the stress level of workers based on their emotional scores.

[1380] A "threshold" refers to a value that serves as a benchmark for indicating that an emotional score or stress score has reached a specific state.

[1381] "Means of notification" refers to technologies and devices used to transmit information to relevant parties using alerts or messages.

[1382] "Mental care suggestions" refer to specific advice and action proposals aimed at improving the mental health of workers.

[1383] "Means of collecting voice and text communications" refers to technologies and devices for acquiring voice and text data from workers.

[1384] "Secure transfer" refers to encrypting data and transmitting it using a secure protocol.

[1385] "Factory robots" refer to mechanical devices designed to automatically perform tasks and operations within a factory.

[1386] A "head-mounted display (HMD)" refers to a device that displays visual information when worn on the head.

[1387] "Smart glasses" refer to glasses-type wearable devices that are capable of displaying visual information and collecting data.

[1388] An "alert" refers to a warning message that is sent when a specific situation or condition occurs.

[1389] "Means of collecting feedback" refers to technologies and equipment used to gather opinions and evaluations from workers and managers.

[1390] "Means of improving system algorithms and threshold settings" refers to techniques and processes for optimizing system performance and settings based on feedback and data.

[1391] This invention relates to a system for monitoring the mental health of factory workers in real time and detecting stress levels. In particular, by combining it with an emotion engine, it is possible to accurately grasp the emotional state of workers and propose appropriate care. The following hardware and software are required to implement the system of this invention.

[1392] 1. Hardware

[1393] Factory robot: A robotic device that performs tasks within a factory.

[1394] Head-mounted displays (HMDs): For example, Microsoft HoloLens can be used to display visual information to workers and deliver notifications.

[1395] Smart glasses: For example, wearable devices that use Google Glass to collect voice and text data.

[1396] 2. Software

[1397] TensorFlow: Data analysis using natural language processing (NLP) technology.

[1398] IBM Watson: Used as an emotion engine to calculate emotion scores from collected data.

[1399] AWS: Manage databases and perform secure data transfer in the cloud.

[1400] Python: Creating programs for data preprocessing, analysis, and alert generation.

[1401] The system operates as follows:

[1402] 1. Data Collection

[1403] The system collects voice and text data from the terminal through the worker's HMD or smart glasses. For example, conversations and work reports made by workers during their work are collected as data.

[1404] 2. Data Transfer

[1405] The collected data is stored as a temporary file, encrypted, and securely transmitted to the AWS cloud using the HTTPS protocol.

[1406] 3. Data Analysis

[1407] Data is preprocessed on the cloud by a server, and text data is analyzed using NLP analysis with TensorFlow. In addition, IBM Watson's emotion engine analyzes audio data to calculate emotion scores and stress scores. For example, if many negative emotions such as "tired" or "irritated" are detected during work, the stress score will be high.

[1408] 4. Alert generation and notification

[1409] The server generates alerts based on emotion and stress scores, and sends alert messages containing appropriate care suggestions to the worker's HMD or smart glasses if the threshold is exceeded. For example, if the stress score exceeds the threshold, a message such as "We recommend you take a break" will be displayed.

[1410] 5. Gathering feedback and making improvements

[1411] The server collects feedback from workers and administrators, and uses it to improve the system's algorithms and threshold settings. This feedback allows the system to continuously improve its performance.

[1412] As a concrete example, consider a case where worker A at a factory wears an HMD while working. Daily voice and text data collected from this worker is transferred to the AWS cloud and analyzed by TensorFlow and IBM Watson. If the emotion score indicates "fatigue" or "dissatisfaction," and the stress score exceeds a threshold, an alert is displayed on the HMD. Subsequently, feedback from worker A is sent to the server, and the system is further optimized.

[1413] Examples of prompts to input into a generative AI model are as follows:

[1414] "Develop a system to monitor the mental health of factory workers and provide real-time care suggestions based on their stress levels. This system will use factory robots, HMDs, and smart glasses to collect and analyze employee voice and text data. It should calculate employee emotion and stress scores using natural language processing algorithms and an emotion recognition engine, and then provide appropriate care suggestions."

[1415] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1416] Step 1:

[1417] The terminal collects voice and text data from the worker's head-mounted display (HMD) or smart glasses. The input at this stage is voice and text communication data generated during the worker's daily work. The output is the collected data, which is stored as a temporary file.

[1418] Step 2:

[1419] The device encrypts the collected temporary file data and sends it to the AWS cloud using a secure protocol (e.g., HTTPS). The input at this stage is the unencrypted data stored in the temporary file, and the output is a notification that the transfer to the secure cloud is complete.

[1420] Step 3:

[1421] The server decrypts the data received in the cloud and performs preprocessing (noise reduction, format standardization). The input is encrypted collected data, and the output is preprocessed data. Specifically, background noise is removed from the audio data, and the format of the text data is standardized.

[1422] Step 4:

[1423] The server applies natural language processing (NLP) to preprocessed data to analyze the text data. The software used is TensorFlow; the input is preprocessed text data, and the output is the result of the text analysis. Specifically, it extracts emotional tone and the frequency of positive and negative language use.

[1424] Step 5:

[1425] The server uses an emotion engine (IBM Watson) to analyze voice data and calculate an emotion score. The input is pre-processed voice data, and the output is the emotion score. Specifically, it analyzes the intonation and tone of the voice to determine the emotional state.

[1426] Step 6:

[1427] The server calculates a stress score based on NLP analysis results and emotion engine results. The input is text analysis results and emotion score, and the output is the stress score. Specifically, it integrates the emotional tone of text and voice to assess the employee's stress level.

[1428] Step 7:

[1429] The server checks whether the calculated stress score exceeds a set threshold. The input is the stress score, and the output is a flag that instructs the server to generate an alert if the threshold is exceeded.

[1430] Step 8:

[1431] The server generates an alert when a threshold is exceeded and sends an alert message, including specific care suggestions, to the worker's HMD or smart glasses. The input is the result of the threshold being exceeded, and the output is the alert notification to the worker. Specifically, a message such as "We recommend you take a break" might be displayed.

[1432] Step 9:

[1433] The server collects feedback from workers and managers through HMDs and smart glasses, and uses this feedback to improve the system's algorithms and threshold settings. The input is feedback data, and the output is the improved algorithms and settings. Specifically, it analyzes the feedback and adjusts the algorithm parameters.

[1434] The following example prompts illustrate the specific actions required to perform each step:

[1435] "Develop a system to monitor the mental health of factory workers and provide real-time care suggestions based on their stress levels. This system will use factory robots, HMDs, and smart glasses to collect and analyze employee voice and text data. It should calculate employee emotion and stress scores using natural language processing algorithms and an emotion recognition engine, and then provide appropriate care suggestions."

[1436] 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.

[1437] 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.

[1438] 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.

[1439] [Fourth Embodiment]

[1440] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1441] 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.

[1442] 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).

[1443] 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.

[1444] 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.

[1445] 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).

[1446] 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.

[1447] 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.

[1448] 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.

[1449] 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.

[1450] 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.

[1451] 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.

[1452] 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".

[1453] The system of this invention monitors the mental health of employees working remotely in real time and detects alerts for stress levels. Specific embodiments are described below.

[1454] Data collection

[1455] 1. Terminal: Employees log in to the system using a work-issued PC or mobile device. Based on the login information, the system prepares to collect data.

[1456] 2. Terminal: A script starts running to collect communication data such as employee emails, chats, meeting recordings, and voice memos. The collected data is saved to a temporary file.

[1457] 3. Terminal: Communication data stored in temporary files is encrypted and transferred to the server using a secure protocol.

[1458] Data Analysis

[1459] 1. Server: Receives the transferred data and stores it in a secure database. The stored data is pre-processed (e.g., noise reduction, formatting standardization).

[1460] 2. Server: Applies natural language processing algorithms to perform sentiment analysis on text and audio data. The analysis includes emotional tone, frequency of positive and negative language use, and specific stress signs (e.g., frequent use of expressions like "tired" or "irritated").

[1461] 3. Server: Calculates emotional and stress scores to assess the mental health status of each employee.

[1462] Alert generation

[1463] 1. Server: Compares the emotional score and stress score to a threshold and generates an alert if the stress score exceeds the threshold. The alert includes specific care suggestions or suggestions for taking a break.

[1464] 2. Server: Records generated alerts in the logging system for future analysis and feedback.

[1465] Notifications and suggestions

[1466] 1. Server: Sends alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[1467] 2. Devices: Notifications are displayed on employees' PCs and mobile devices, along with a message such as, "Your stress level is high; we recommend you take a break."

[1468] 3. User: Employees can choose to accept or reject the proposed mental health care action. They submit their decision to the system as feedback.

[1469] Feedback and continuous monitoring

[1470] 1. Server: Collects the results of the proposal's implementation and feedback, and records them in the system log.

[1471] 2. Server: Based on the collected feedback, the server will make timely improvements to the natural language processing algorithm and threshold settings.

[1472] 3. Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[1473] Specific example

[1474] This section describes a specific example of monitoring employee A's mental health.

[1475] 1. Device: Collect one week's worth of email and chat data from employee A's PC and securely transfer it to the cloud.

[1476] 2. Server: The server analyzes the data using natural language processing technology and determines that "negative tones" account for 50% of the responses. The stress score is calculated to be 75 (threshold is 70).

[1477] 3. Server: Generates an alert such as "Take a break" because the stress score exceeds the threshold.

[1478] 4. Server: Send an alert to the administrator and employee A, including specific care suggestions such as "Consult a mental health professional."

[1479] 5. Terminal: A notification will appear on employee A's PC, allowing employee A to submit feedback.

[1480] This system allows for proper management of employees' mental health and provision of necessary care, even in a remote work environment.

[1481] The following describes the processing flow.

[1482] Step 1:

[1483] Terminal: An employee logs into the system using a work-issued PC or mobile device. The system verifies the login information and begins preparing to collect data.

[1484] Step 2:

[1485] Terminal: A script runs that collects communication data such as employee emails, chats, meeting recordings, and voice memos. This data is stored in temporary files.

[1486] Step 3:

[1487] Terminal: Data stored in temporary files is encrypted and sent to the server using a secure protocol (e.g., HTTPS).

[1488] Step 4:

[1489] Server: Receives the transferred data and stores it in a secure database. The stored data is preprocessed for analysis (e.g., noise reduction, formatting standardization).

[1490] Step 5:

[1491] Server: Applies natural language processing (NLP) algorithms to analyze collected text and audio data. The analysis extracts emotional tones, the frequency of positive and negative language use, and specific stress signs.

[1492] Step 6:

[1493] Server: Based on the NLP results, it calculates emotional and stress scores to assess the mental health status of each employee.

[1494] Step 7:

[1495] Server: Compares the emotion score and stress score to pre-set thresholds and generates an alert if the stress score exceeds the threshold.

[1496] Step 8:

[1497] Server: Alerts include specific care suggestions and break suggestions. These alerts are recorded in the logging system.

[1498] Step 9:

[1499] Server: Sends generated alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[1500] Step 10:

[1501] Device: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a break" may be displayed.

[1502] Step 11:

[1503] User: Employees choose whether to accept the proposed mental health care action. The choice is sent to the system as feedback.

[1504] Step 12:

[1505] Server: Collects feedback from employees and administrators and uses it to make timely improvements to the system's natural language processing algorithms and threshold settings.

[1506] Step 13:

[1507] Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[1508] (Example 1)

[1509] 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".

[1510] With the widespread adoption of remote work, managing employees' mental health, in addition to monitoring work progress, has become a crucial issue. However, traditional methods make it difficult to grasp employees' stress levels and emotional states in real time and provide appropriate countermeasures, often resulting in mental health problems being left unaddressed. This leads to problems such as decreased work efficiency and increased employee turnover, making an effective mental health monitoring system necessary.

[1511] 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.

[1512] In this invention, the server includes means for monitoring the mental health status of employees in real time, means for collecting and analyzing employee communication data using natural language processing technology, and means for calculating and evaluating emotion scores and stress scores based on the analysis results. This makes it possible to effectively monitor the mental health of employees and provide timely and appropriate alerts and care suggestions.

[1513] "Mental health" refers to an individual's psychological, emotional, and social well-being. Mental health is a crucial factor that influences one's ability to cope with everyday stress, manage work and relationships, and more.

[1514] "Real-time monitoring" is a process of monitoring and evaluating data and conditions almost instantly. This process allows for a rapid response when problems occur.

[1515] "Natural language processing technology" refers to the technology that enables computers to understand and process human language. This includes text analysis and sentiment assessment.

[1516] "Communication data" refers to information such as emails, chats, meeting recordings, and voice memos exchanged by employees in the course of their work. This information is important data for evaluating the mental state of employees.

[1517] An "emotion score" is a numerical representation of an employee's emotional state, based on data analyzed using natural language processing technology. The score is used to evaluate emotions such as positive and negative.

[1518] A "stress score" is a numerical representation of an employee's stress level, based on data analyzed using natural language processing technology. This score is used to assess the mental burden on employees.

[1519] A "threshold" is a set score limit that triggers a specific alert or action. If the stress score exceeds this value, the system will take a specific action.

[1520] "Alert generation" is the process by which a system issues a warning or notification when set conditions or thresholds are met. This allows relevant parties to take immediate action.

[1521] "Notifying administrators and employees" refers to the process of sending analysis results and alert information to administrators' and employees' devices. Notifications are sent via methods such as email or applications.

[1522] "Mental health suggestions" refer to specific advice and action plans provided to improve employees' mental health. Examples include suggestions for breaks or consultations with mental health professionals.

[1523] "Collecting feedback" is the process by which employees submit to the system the results and opinions they have had regarding the mental health care suggestions they have received. The system then uses this feedback to make improvements.

[1524] "Improving analysis algorithms and threshold settings" refers to using collected feedback data to promptly update the system's analysis methods and the conditions for triggering actions to the latest and most optimal state.

[1525] The system of this invention monitors the mental health of employees working remotely in real time and detects alerts for stress levels. Specific embodiments are described below.

[1526] System Overview

[1527] This system uses natural language processing technology to collect and analyze employee communication data in order to monitor employees' mental health status in real time, and to calculate and evaluate emotion scores and stress scores. The server generates alerts based on stress scores that exceed a threshold and notifies both administrators and the employees themselves. It also provides specific mental care suggestions and collects employee feedback to improve the analysis algorithms and threshold settings.

[1528] Data collection

[1529] When an employee logs into the system using their work PC or mobile device, they enter their login information (e.g., user ID, password), and the system prepares to collect data. A script automatically starts on the terminal, collecting communication data such as emails, chats, meeting recordings, and voice memos, and saving it to a temporary file. The collected data is encrypted using AES (Advanced Encryption Standard) or similar and transferred to a secure server using the SSL / TLS protocol.

[1530] Data Analysis

[1531] The server receives the transferred data and stores it in a secure database such as AWS RDS. The stored data undergoes a preprocessing process (e.g., noise reduction, formatting standardization) and then sentiment analysis is performed using natural language processing algorithms (e.g., GPT-3, Python's NLTK library). This analysis calculates each employee's sentiment score (0-100) and stress score (1-10). This allows for an assessment of each employee's mental health status.

[1532] Alert generation and notification

[1533] The server compares the emotion score and stress score to a threshold (e.g., stress score of 70 or higher) and generates an alert if the threshold is exceeded. The generated alert is recorded in a logging system (e.g., MongoDB) and notified to the administrator and the employee. The notification includes specific mental health care suggestions (e.g., "Take a break," "Consult a mental health professional") via email (e.g., Amazon SES) or a dedicated application.

[1534] Employee feedback and system improvements

[1535] Employees, as users, choose to accept or reject proposed mental health care actions and provide feedback to the system. The server collects this feedback information and records it in the system log (e.g., Redis). Based on the collected feedback, the analysis algorithms and threshold settings are improved.

[1536] Continuous monitoring

[1537] The server periodically collects, analyzes, and generates alerts to continuously monitor employees' mental health. This makes it possible to properly manage employee mental health and provide necessary care, even in a remote work environment.

[1538] Specific example

[1539] For example, employee A collects a week's worth of email and chat data from their PC and securely transfers it to the cloud. The server then analyzes the data using natural language processing technology and determines that 50% of it has a "negative tone." If the stress score is calculated to be 75 (the threshold is 70), the system generates an alert such as "Take a break" and notifies both the administrator and employee A. It also includes specific care suggestions such as "Consult a mental health professional." Finally, employee A can choose whether to take the suggested action and provide feedback on the result.

[1540] Prompt example

[1541] "Create a program that uses employee email and chat data to monitor their mental health in real time. It should generate an alert and send a notification if the stress score exceeds a threshold. The technologies to use are the Python NLTK library and GPT-3."

[1542] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1543] Step 1:

[1544] Terminal: Employees log in to the system using a work-issued PC or mobile device.

[1545] Input: Authentication information such as user ID and password.

[1546] Data processing: Authentication processing of login information.

[1547] Output: A message indicating successful or failed login. The specific action involves entering a user ID and password and sending a request to the system's authentication server.

[1548] Step 2:

[1549] Terminal: Upon successful login, the script will automatically launch to collect communication data such as employee emails, chats, meeting recordings, and voice memos.

[1550] Input: Employee communication data (emails, chat logs, voice memos, etc.).

[1551] Data processing: Data collection and saving to temporary files.

[1552] Output: Communication data saved in a temporary file. Specifically, data is retrieved from each application (Outlook, Teams, etc.) using APIs and saved to a temporary file.

[1553] Step 3:

[1554] Terminal: Encrypts data stored in temporary files using AES or similar encryption methods and transfers it to the server using the SSL / TLS protocol.

[1555] Input: Communication data stored in a temporary file.

[1556] Data processing: Data encryption and secure transfer.

[1557] Output: An encrypted temporary file is transferred to the server. Specifically, the data is encrypted using an encryption library and uploaded to the server via HTTPS.

[1558] Step 4:

[1559] Server: Receives the transferred data and stores it in a secure database.

[1560] Input: Encrypted data that arrived on the server.

[1561] Data processing: Decrypting data and saving it to a database.

[1562] Output: Communication data stored in a secure database. Specifically, the data is decrypted and inserted into a database (e.g., AWS RDS).

[1563] Step 5:

[1564] Server: Preprocesses the stored data (noise reduction, format standardization).

[1565] Input: Communication data stored in the database.

[1566] Data processing: Data cleaning and formatting standardization.

[1567] Output: Preprocessed dataset. Specifically, data cleaning is performed using a script such as Python to remove noise and unwanted parts.

[1568] Step 6:

[1569] Server: Applies natural language processing algorithms to perform sentiment analysis.

[1570] Input: Preprocessed dataset.

[1571] Data processing: Sentiment analysis of text and audio data.

[1572] Output: Sentiment score and stress score. Specifically, the positive and negative sentiment of each document is analyzed using GPT-3 and the Python NLTK library.

[1573] Step 7:

[1574] Server: Evaluates each employee's mental health status based on the calculated emotion score and stress score.

[1575] Input: Emotion score and stress score.

[1576] Data processing: Assessment of mental health status.

[1577] Output: Evaluation result (normal, caution needed, dangerous, etc.). Specifically, the system compares the score with a threshold and classifies the state.

[1578] Step 8:

[1579] Server: Generates an alert when the stress score exceeds a threshold.

[1580] Input: Stress score for each employee.

[1581] Data processing: Comparison with thresholds and alert generation.

[1582] Output: Alert message. Specifically, if the stress score exceeds a specified threshold, an alert message such as "Take a break" is generated.

[1583] Step 9:

[1584] Server: Logs generated alerts and notifies administrators and employees themselves.

[1585] Input: Alert message.

[1586] Data processing: logging and sending notifications.

[1587] Output: Log entries and notification messages. Specifically, alert information is recorded in a logging system such as MongoDB, and administrators and employees are notified via email or a dedicated application.

[1588] Step 10:

[1589] User: Employees receive a notification and can choose whether to take the suggested mental health care action.

[1590] Input: Alert notification.

[1591] Data processing: Implementation and feedback of mental care actions.

[1592] Output: Feedback data. Specifically, an employee clicks the "Take Break" button and sends the result to the system.

[1593] Step 11:

[1594] Server: Collects feedback data and records it in the system log.

[1595] Input: Employee feedback.

[1596] Data processing: Recording feedback.

[1597] Output: Log entries. Specifically, feedback information is saved to a database such as Redis.

[1598] Step 12:

[1599] Server: Based on the collected feedback data, improve the analysis algorithm and threshold settings.

[1600] Input: Feedback data.

[1601] Data processing: Updating algorithms and settings.

[1602] Output: Improved analysis algorithm and threshold settings. Specifically, the parameters of the machine learning algorithm are readjusted based on the collected feedback.

[1603] Step 13:

[1604] Server: Regularly collects, analyzes, and generates alerts to continuously monitor employees' mental health.

[1605] Input: Communication data and feedback data from all employees.

[1606] Data processing: Iterative monitoring and alert generation process.

[1607] Output: Continuously updated mental health status and alerts. Specifically, the system periodically runs scripts to collect and analyze new data to monitor employee mental health.

[1608] (Application Example 1)

[1609] 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".

[1610] Monitoring the mental health of employees working remotely presents a challenge. In particular, there is a lack of technology to detect stress and fatigue in real time from employee voice and facial expressions, and to provide effective support. This can increase employee health risks and reduce work efficiency.

[1611] 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.

[1612] In this invention, the server includes means for monitoring the mental health status of employees, means for analyzing employee communication data using natural language processing technology, means for collecting voice and visual data using sensors in smart glasses, means for analyzing the employee's stress level from the collected visual and voice data, means for calculating an emotional score and a stress score based on the analysis results, means for notifying the administrator and the employee themselves when the stress score exceeds a set threshold, and means for providing specific mental care suggestions. This makes it possible to monitor the mental health status of employees in real time and provide appropriate care quickly.

[1613] An "employee" refers to an individual who belongs to an organization or company and engages in its operations.

[1614] "Means of monitoring mental health status" refers to techniques or methods for assessing and tracking the psychological and emotional health status of employees.

[1615] "Natural language processing technology" refers to computer technologies and algorithms used to understand and analyze human language.

[1616] "Communication data" refers to information collected in the form of emails, chats, meeting recordings, voice memos, and other similar formats.

[1617] "Smart glasses" are wearable devices that are used for data collection and analysis, and are capable of augmented reality and voice control.

[1618] "Means for collecting audio and visual data using sensors" refers to methods of acquiring audio and video information using devices such as microphones and cameras built into smart glasses.

[1619] "Methods for analyzing stress levels" refers to technologies that analyze collected audio and visual data to evaluate the stress levels of employees.

[1620] "Means for calculating emotional and stress scores" refers to techniques or methods for quantifying and evaluating indicators related to emotions and stress.

[1621] A "threshold" refers to a numerical value or condition that serves as the basis for generating an alert.

[1622] "Means of notification" refers to technologies and methods for communicating alerts and information to administrators and employees themselves.

[1623] "Mental health suggestions" refer to specific actions and advice recommended to improve employees' mental health.

[1624] A "server" refers to a computer system that handles data processing and storage.

[1625] The system of this invention monitors employees' mental health in real time, detects stress levels, and provides appropriate care. Specific embodiments are described below.

[1626] Data collection

[1627] Terminal (Smart Glasses): When an employee wears smart glasses, the device collects audio and visual data. The microphone and camera built into the smart glasses record the employee's voice and facial expressions in real time and temporarily store them in storage.

[1628] Data transfer and encryption

[1629] Device (smart glasses): The collected data is encrypted and transferred to the server using a secure protocol (e.g., TLS).

[1630] Data Analysis

[1631] Server: The server receives the transferred data and stores it in a secure database (e.g., PostgreSQL). The data is then preprocessed, and audio data is converted to text using the Google Speech-to-Text API. For visual data, sentiment analysis is performed using an emotion recognition API (e.g., Amazon Comprehend).

[1632] Stress level analysis and score calculation

[1633] Server: Analyzes employee stress levels from collected audio and visual data. Natural language processing techniques and speech analysis libraries are used for the analysis. Based on the analysis results, emotion scores and stress scores are calculated.

[1634] Alert generation

[1635] Server: Generates an alert when the stress score exceeds a set threshold. The alert includes specific mental health care suggestions such as "Take a break" or "Consult a mental health professional."

[1636] notification

[1637] Server: Generated alerts are notified to both the administrator and the employee themselves. The notification is displayed on the smart glasses' screen, visually indicating the content of the alert.

[1638] Feedback and continuous monitoring

[1639] User: Employees can choose to accept or reject proposed care actions. The choice is sent to the system as feedback. This allows the system to continuously monitor the employee's mental health status and improve the analysis algorithms and threshold settings as needed.

[1640] Specific example: Stress detection during meetings

[1641] While employee B is participating in an important meeting, smart glasses analyze B's voice and facial expressions in real time. If the analysis determines that B is experiencing high stress levels, the smart glasses display a message saying, "Take a deep breath and relax." After the meeting, B is suggested to take a break, and if necessary, recommended to consult a mental health professional.

[1642] Prompt example:

[1643] Prompt text to input to the generative AI model:

[1644] "Please describe in detail the design of a smart glasses application that monitors employee stress levels in real time and suggests necessary care. Specifically, explain how voice and visual data will be collected, analyzed, and alerts will be generated. Also, describe in detail how employee feedback will be provided to the system regarding whether they accept the suggestions."

[1645] While the embodiments for carrying out the present invention have been described in detail, the invention is not limited thereto, and various modifications and improvements are possible.

[1646] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1647] Step 1: Data Collection

[1648] The device (smart glasses) collects audio and visual data when worn by an employee. Input is data from the microphone and camera built into the smart glasses, and output is raw audio and visual data temporarily stored in storage. Specifically, audio data is captured from the microphone every second, and visual data is periodically captured by the camera as snapshots and videos.

[1649] Step 2: Data encryption and transfer

[1650] The device (smart glasses) encrypts the collected audio and visual data and transfers it to the server via a secure protocol. The input is unencrypted collected data, and the output is encrypted data packets. Specifically, secure protocols such as TLS are used to protect the data from hacking and unauthorized access.

[1651] Step 3: Convert speech to text

[1652] The server receives the transferred data and stores it in a database. It then converts the audio data to text using the Google Speech-to-Text API. The input is encrypted audio data, and the output is text data. Specifically, an audio file is sent via an API call and returned as text.

[1653] Step 4: Emotion Analysis

[1654] The server uses text and visual data to perform sentiment analysis with an emotion recognition API (e.g., Amazon Comprehend). The input is text and image data, and the output is sentiment scores. Specifically, text data is sent to the sentiment recognition API, which then returns positive, negative, and neutral sentiment scores. Visual data is analyzed using a face recognition algorithm, and sentiment scores are similarly generated.

[1655] Step 5: Stress level analysis and score calculation

[1656] The server analyzes employee stress levels from collected visual and audio data. Inputs are emotion scores and features from audio data, while output is a stress score. Specifically, natural language processing techniques and speech analysis libraries are used to comprehensively evaluate stress indicators (e.g., "irritation," "fatigue") extracted from text and audio.

[1657] Step 6: Generate Alerts

[1658] The server generates an alert when the stress score exceeds a set threshold. The input is the stress score, and the output is an alert notification and specific mental health care suggestions. Specifically, when the score exceeds the set threshold, an alert message is generated, including suggestions such as "Take a break" or "Consult a mental health professional."

[1659] Step 7: Notification

[1660] The server notifies administrators and employees of the generated alerts. The input is the alert message, and the output is the notification displayed on the smart glasses' screen. Specifically, notification data is sent to the smart glasses via a network protocol, and the notification is sent to the administrator via email or chat application.

[1661] Step 8: Gathering Feedback

[1662] The user (employee) chooses whether to accept or reject the proposed care action. The input is the user's response to the proposal, and the output is feedback data sent to the system. Specifically, the option to "accept" or "reject" is displayed through the smart glasses interface, and the selection result is fed back to the server.

[1663] Step 9: Continuous Monitoring

[1664] The server continuously improves the algorithm and updates threshold settings based on the collected feedback. The input is feedback data, and the output is the updated algorithm parameters. Specifically, the feedback is used to optimize the machine learning model parameters, which are then reflected in the next analysis.

[1665] 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.

[1666] This invention relates to a system for monitoring the mental health of employees working remotely in real time and detecting their stress levels. In particular, by combining it with an emotion engine, it becomes possible to more accurately understand the emotional state of employees and propose appropriate care.

[1667] Data collection

[1668] 1. Terminal: Employees log in to the system using a work-issued PC or mobile device. The system verifies the login information and prepares to collect data.

[1669] 2. Terminal: A script is executed to collect communication data such as employee emails, chats, meeting recordings, and voice memos. This data is stored in temporary files.

[1670] 3. Terminal: Data stored in temporary files is encrypted and sent to the server using a secure protocol (e.g., HTTPS).

[1671] Data analysis and emotion recognition

[1672] 1. Server: Receives the transferred data and stores it in a secure database. The stored data is preprocessed for analysis (e.g., noise reduction, formatting standardization).

[1673] 2. Server: Applies natural language processing (NLP) algorithms to analyze the collected text and audio data. From the analysis results, it extracts emotional tone, frequency of positive and negative language use, and specific stress signs.

[1674] 3. Server: Furthermore, an emotion engine is used to recognize employees' emotions from communication data. Specifically, the emotion engine analyzes voice and text data to recognize emotions in real time.

[1675] Calculation of emotional score and stress score

[1676] 1. Server: Based on the results of NLP and the emotion engine, it calculates emotion scores and stress scores to assess the mental health status of each employee.

[1677] Alert generation

[1678] 1. Server: Compares the emotion score and stress score with pre-set thresholds and generates an alert if the stress score exceeds the threshold.

[1679] 2. Server: Alerts include specific care suggestions and break suggestions. These alerts are recorded in the logging system.

[1680] Notifications and suggestions

[1681] 1. Server: Sends generated alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[1682] 2. Devices: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a break" may be displayed.

[1683] 3. User: Employees choose whether to accept the proposed mental health care action. The choice is sent to the system as feedback.

[1684] Feedback and continuous monitoring

[1685] 1. Server: Collects feedback from employees and administrators and uses it to make timely improvements to the system's natural language processing algorithms, sentiment engine, and threshold settings.

[1686] 2. Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[1687] Specific example

[1688] This section describes a specific example of monitoring employee B's mental health.

[1689] 1. Device: Collect one week's worth of email and chat data from employee B's PC and securely transfer it to the cloud.

[1690] 2. Server: The server analyzes the collected data using natural language processing technology and an emotion engine. For example, "negative tones" are detected frequently, and emotions such as "anxiety" are identified from the audio data. The stress score becomes 80 (threshold is 70).

[1691] 3. Server: Generate an alert because the stress score exceeds the threshold, and include a suggestion such as "Take a break."

[1692] 4. Server: Send an alert to the administrator and employee B, including specific care suggestions such as "Consult a mental health professional."

[1693] 5. Terminal: A notification will appear on employee B's PC, allowing employee B to submit feedback.

[1694] This system allows for accurate and real-time monitoring of employees' emotional states, even in a remote work environment, and enables the provision of appropriate care.

[1695] The following describes the processing flow.

[1696] Step 1:

[1697] Terminal: An employee logs into the system using a work-issued PC or mobile device. Based on the login information, the system begins preparing to collect data.

[1698] Step 2:

[1699] Terminal: A script automatically runs to collect communication data such as employee emails, chats, meeting recordings, and voice memos. This data is stored in temporary files.

[1700] Step 3:

[1701] Terminal: Encrypts data stored in temporary files and sends it to the server using a secure protocol (e.g., HTTPS).

[1702] Step 4:

[1703] Server: Receives data and stores it in a secure database. Stored data is preprocessed for analysis (e.g., noise reduction, formatting standardization).

[1704] Step 5:

[1705] Server: Applies natural language processing (NLP) algorithms to analyze collected text and audio data. The analysis results include emotional tone, positive and negative expressions, and specific stress signs.

[1706] Step 6:

[1707] Server: Uses an emotion engine to perform real-time emotion recognition from the same communication data. For example, emotions such as "anxiety" and "anger" are detected from voice data, while emotions such as "sadness" and "stress" are identified from text data.

[1708] Step 7:

[1709] Server: Calculates employee emotional scores and stress scores based on the analysis results of NLP and the emotion engine.

[1710] Step 8:

[1711] Server: Compares the emotion score and stress score to the set threshold. If the stress score exceeds the threshold, an alert is generated.

[1712] Step 9:

[1713] Server: The generated alerts include specific care suggestions (e.g., suggestions for taking a break or advice on deep breathing) and links to consult with professionals. These alerts are recorded in the logging system.

[1714] Step 10:

[1715] Server: Sends alerts to administrators and employees themselves. Notifications are sent via email or a dedicated application.

[1716] Step 11:

[1717] Device: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a break" may be displayed.

[1718] Step 12:

[1719] User: Employees can choose whether or not to accept the proposed mental health care action. The choice is sent to the system as feedback.

[1720] Step 13:

[1721] Server: Receives feedback and uses it to make timely improvements to natural language processing algorithms, sentiment engines, and threshold settings.

[1722] Step 14:

[1723] Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[1724] Specific example

[1725] This section describes a specific example of monitoring employee C's mental health.

[1726] Step 1:

[1727] Terminal: Collect one week's worth of email and chat data from employee C's PC and securely transfer it to the cloud.

[1728] Step 2:

[1729] Server: Analyzes data using natural language processing technology and an emotion engine. The analysis reveals a high proportion of negative tones, and emotions such as anxiety are detected from the audio data. A stress score of 85 (threshold is 70) was calculated.

[1730] Step 3:

[1731] Server: Generates an alert such as "Take a break" because the stress score exceeds the threshold. This alert also includes specific suggestions and links to consult with experts.

[1732] Step 4:

[1733] Server: Sends alerts to administrator and employee C.

[1734] Step 5:

[1735] Terminal: A notification appears on employee C's PC, and when employee C submits feedback, the system receives it and improves the algorithm.

[1736] This system allows for accurate and real-time monitoring of employees' emotional states, even in a remote work environment, and enables the provision of appropriate care.

[1737] (Example 2)

[1738] 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".

[1739] In a remote work environment, it is difficult to monitor employees' mental health in real time and accurately grasp their stress levels and emotional fluctuations. Furthermore, there is a lack of means to quickly provide appropriate care suggestions when employees are experiencing excessive stress. In such circumstances, it is difficult to prevent the deterioration of employees' mental health, potentially leading to decreased productivity and increased employee turnover.

[1740] 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.

[1741] In this invention, the server includes means for monitoring the mental health status of employees, means for analyzing employee communication data using natural language processing technology, and means for recognizing employee emotions from text and voice data using an emotion engine. This makes it possible to accurately grasp the mental health status of employees in real time and to quickly provide appropriate care suggestions.

[1742] "Mental health status" refers to an employee's mental and emotional well-being. This includes stress levels and emotional stability.

[1743] "Monitoring methods" refer to technologies and devices for continuously observing and recording employees' mental health status over time.

[1744] "Natural language processing technology" refers to artificial intelligence technology used to understand, analyze, and generate human language. This includes analyzing text data and classifying emotions.

[1745] "Communication data" refers to data obtained from communication methods used by employees during work, such as email, text chat, audio conference recordings, and voice memos.

[1746] An "emotion engine" refers to algorithms and technologies used to analyze and recognize emotions from text and audio data. This often includes generative AI models.

[1747] An "emotion score" is an indicator that numerically represents an employee's emotional state, calculated from the results of analysis by the emotion engine.

[1748] A "stress score" is an indicator that quantifies how much stress an employee is experiencing. It is calculated based on the results of an emotional analysis.

[1749] A "threshold" refers to a pre-set baseline value for emotion and stress scores. An alert is generated when this value is exceeded.

[1750] "Notification methods" refer to technologies and devices used to communicate alerts and care suggestions to administrators and employees themselves. This includes email and dedicated applications.

[1751] "Specific mental health support suggestions" refer to concrete actions and resources to improve employees' mental health. Examples include suggestions for breaks or links to mental health professionals.

[1752] This invention is a system that monitors the mental health of employees working remotely in real time and detects their stress levels. In particular, by combining it with an emotion engine, it is possible to more accurately understand the emotional state of employees and propose appropriate care.

[1753] Data collection

[1754] 1. Terminal: Employees log in to the system using a work PC or mobile device. The system verifies the login information and prepares to collect data.

[1755] 2. Device: After logging in, a script is automatically executed to collect communication data such as emails, chats, meeting recordings, and voice memos. For example, email data is collected from the mailbox by gathering the subject and body text, and chat data is retrieved using an API. Meeting recordings and voice memos are recorded via the device's microphone and saved as temporary files.

[1756] 3. Terminal: The collected data is stored in a temporary file and then encrypted using the AES-256 encryption algorithm. The encrypted data is then sent to the server using the secure HTTPS protocol.

[1757] Data analysis and emotion recognition

[1758] 1. Server: Receives data transferred from terminals and stores it in a secure database. After storage, the data undergoes preprocessing such as noise reduction and formatting standardization.

[1759] 2. Server: Applies natural language processing (NLP) algorithms to analyze text and audio data. For NLP analysis, libraries such as spaCy or NLTK are used. Audio data is converted to text using the Google Cloud Speech-to-Text API and integrated with the text data for analysis.

[1760] 3. Server: Recognizes emotions from text and audio data using an emotion engine. The emotion engine uses generative AI models such as BERT or GPT-3 to perform emotion classification tasks.

[1761] Calculation of emotional score and stress score

[1762] 1. Server: Based on the results of NLP and the emotion engine, it calculates an emotion score and a stress score. A weighted average or rule-based scoring system is used to calculate the scores. A higher emotion score is assigned when there is a high frequency of positive language use, and a higher stress score is assigned when there is a high frequency of negative language use.

[1763] Alert generation

[1764] 1. Server: Compares the emotion score and stress score to a threshold, and generates an alert if the stress score exceeds the threshold. For example, if the threshold is 70 and the stress score is 80, an alert will be generated.

[1765] 2. Server: Alerts include specific care suggestions. These suggestions may include consultation with a psychological counselor, use of relaxation apps, or suggestions for taking a break. These alerts are recorded in the logging system and stored as a dataset that can be used for analysis.

[1766] Notifications and suggestions

[1767] 1. Server: Notifies administrators and employees of generated alerts. Notifications are primarily made via email or a dedicated application (e.g., a company-specific mental health care app).

[1768] 2. Devices: Notifications are displayed on employees' PCs and mobile devices. For example, a message such as "Your stress level is high. We recommend you take a 10-minute break" may pop up.

[1769] 3. User: Employees choose whether to accept the proposed care action. The choice is sent to the system as feedback and recorded in the database. This will be reflected in future analyses.

[1770] Feedback and continuous monitoring

[1771] 1. Server: Collects feedback from employees and administrators to improve the system's NLP algorithms, emotion engine, and threshold settings. Feedback is used to improve the accuracy of the emotion recognition model and readjust thresholds.

[1772] 2. Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely. For example, it performs weekly data analysis and provides administrators with monthly reports.

[1773] Specific examples and prompt statements

[1774] As a concrete example of monitoring employee B's mental health, the following operations are performed: Email and chat data for one week are collected from employee B's PC and securely transferred to the cloud. The collected data is analyzed using natural language processing technology and an emotion engine. If negative tones are detected frequently and emotions such as "anxiety" are identified from the voice data, the stress score becomes 80 (the threshold is 70). In this case, an alert is generated and a suggestion such as "Take a break" is made. The alert is sent to the administrator and employee B, and employee B can submit feedback.

[1775] Example prompt: "Collect one week's worth of emails, chats, and meeting recordings from employee B and analyze them using natural language processing and a sentiment engine. If the results show a high proportion of negative tones, generate specific care suggestions and notify the employee."

[1776] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1777] Step 1:

[1778] Terminal: Employees log in to the system using a work PC or mobile device.

[1779] Input: Employee login information (User ID, password)

[1780] Operation: The system authenticates login information and prepares for data collection.

[1781] Output: Login success message, instruction to start data collection script.

[1782] Step 2:

[1783] Terminal: After logging in, a script will run that automatically collects communication data such as emails, chats, meeting recordings, and voice memos.

[1784] Input: Employee account information after login authentication

[1785] Operation: Email data is collected from the subject and body text of employee mailboxes. Chat data is retrieved using an API. Meeting recordings and voice memos are recorded via the device's microphone and stored in temporary files.

[1786] Output: Text data (emails and chats), audio data (meeting recordings and voice memos), temporary files

[1787] Step 3:

[1788] Terminal: The collected data is saved to a temporary file and then encrypted using the AES-256 encryption algorithm.

[1789] Input: Raw data stored in a temporary file

[1790] Operation: Encrypts data using the AES-256 encryption algorithm. Sends the encrypted data to the server using the HTTPS protocol.

[1791] Output: Encrypted data, HTTPS request

[1792] Step 4:

[1793] Server: Receives data transferred from terminals and stores it in a secure database.

[1794] Input: Encrypted data included in the HTTPS request

[1795] Operation: Saves data to a database. After saving, preprocessing is performed on the data, such as noise reduction and formatting standardization.

[1796] Output: Preprocessed data

[1797] Step 5:

[1798] Server: Applies natural language processing (NLP) algorithms to analyze text and audio data.

[1799] Input: Preprocessed text data and audio data

[1800] Operation: For NLP analysis, libraries such as spaCy or NLTK are used to tokenize text data and tag parts of speech. Audio data is converted to text using the Google Cloud Speech-to-Text API and integrated with the text data for analysis.

[1801] Output: Analyzed text data, text data converted from audio data

[1802] Step 6:

[1803] Server: Uses an emotion engine to recognize emotions from text and audio data.

[1804] Input: Analyzed text data, text data converted from audio data

[1805] Operation: The emotion engine uses generative AI models such as BERT or GPT-3 to perform emotion classification tasks. It assigns emotion labels such as positive, negative, and neutral to the data.

[1806] Output: Data with emotion labels

[1807] Step 7:

[1808] Server: Calculates emotion scores and stress scores based on the results of NLP and the emotion engine.

[1809] Input: Data with emotion labels

[1810] Operation: The scoring system uses weighted averages or rule-based scoring systems to calculate scores. A higher sentiment score is assigned when positive language usage is frequent, and a higher stress score is assigned when negative language usage is frequent.

[1811] Output: Recorded sentiment score and stress score

[1812] Step 8:

[1813] Server: Compares the emotion score and stress score to a threshold, and generates an alert if the stress score exceeds the threshold.

[1814] Input: Calculated sentiment score and stress score, threshold setting

[1815] Operation: Determines whether the stress score exceeds a threshold and generates an alert if it does.

[1816] Output: Alert Information

[1817] Step 9:

[1818] Server: Based on alerts recorded in the database, it notifies administrators and employees themselves.

[1819] Input: Alert Information

[1820] Operation: Notifies administrators and employees via email or a dedicated application. Sends messages such as, "Your stress level is high. We recommend you take a break."

[1821] Output: Notification message

[1822] Step 10:

[1823] Device: Notifications will be displayed on employees' PCs and mobile devices.

[1824] Input: Notification message

[1825] Action: A notification message is displayed. A pop-up appears saying something like, "Your stress level is high. We recommend you take a 10-minute break."

[1826] Output: Displayed notification message

[1827] Step 11:

[1828] User: The employee chooses whether to accept the proposed care action.

[1829] Input: Displayed notification message

[1830] Operation: Employees can choose to "accept" or "reject" a notification, and the system receives feedback on their choice.

[1831] Output: Acceptance or rejection feedback information

[1832] Step 12:

[1833] Server: Collects feedback from employees and administrators to improve the system's NLP algorithms, sentiment engine, and threshold settings.

[1834] Input: Feedback information

[1835] Operation: Feedback is recorded in a database and used to adjust algorithms, emotion engines, and threshold settings for future analyses.

[1836] Output: Improved NLP algorithm, emotion engine, threshold settings

[1837] Step 13:

[1838] Server: Regularly collects, analyzes, and generates alerts to continuously monitor the mental health of employees working remotely.

[1839] Input: Continuously collected communication data

[1840] Operation: Performs weekly data analysis. Provides administrators with monthly reports summarizing the data.

[1841] Output: Weekly data analysis report, monthly report

[1842] (Application Example 2)

[1843] 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".

[1844] The mental health of factory workers is susceptible to the effects of harsh working conditions and stress, making appropriate monitoring and care essential. However, current systems lack the ability to grasp workers' stress and emotions in real time and provide appropriate care suggestions. In particular, there is a need to quickly detect changes in emotions and provide workers with effective improvement suggestions. Therefore, there is a demand for a system that efficiently and accurately monitors the mental health of employees working on-site and provides appropriate care suggestions in real time.

[1845] 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.

[1846] In this invention, the server includes means for monitoring the mental health status of employees, means for analyzing employee communication data using natural language processing technology, means for calculating an emotion score and a stress score based on the analysis results, means for notifying managers and the employee themselves when the stress score exceeds a set threshold, means for providing specific mental care suggestions, means for collecting voice and text communications of factory workers, means for securely transferring the collected data, means for notifying workers of alerts using factory robots, head-mounted displays, or smart glasses, and means for collecting feedback from managers and workers and improving the system's algorithms and threshold settings. This makes it possible to accurately monitor the mental health of factory workers in real time and to quickly provide appropriate care suggestions.

[1847] "Employee mental health status" refers to the psychological and emotional health of workers.

[1848] "Monitoring methods" refer to technologies and devices used to monitor the status of employees by collecting and analyzing data in real time or periodically.

[1849] "Natural language processing technology" is a technology that enables computers to analyze, understand, and generate human language.

[1850] "Communication data" refers to data in the form of emails, chats, meeting recordings, voice memos, work reports, conversations, etc., that workers exchange in their daily work.

[1851] "Emotion score" refers to a numerical value that quantitatively represents the emotional state of a worker, calculated using natural language processing and emotion recognition technologies.

[1852] A "stress score" refers to a numerical value that evaluates the stress level of workers based on their emotional scores.

[1853] A "threshold" refers to a value that serves as a benchmark for indicating that an emotional score or stress score has reached a specific state.

[1854] "Means of notification" refers to technologies and devices used to transmit information to relevant parties using alerts or messages.

[1855] "Mental care suggestions" refer to specific advice and action proposals aimed at improving the mental health of workers.

[1856] "Means of collecting voice and text communications" refers to technologies and devices for acquiring voice and text data from workers.

[1857] "Secure transfer" refers to encrypting data and transmitting it using a secure protocol.

[1858] "Factory robots" refer to mechanical devices designed to automatically perform tasks and operations within a factory.

[1859] A "head-mounted display (HMD)" refers to a device that displays visual information when worn on the head.

[1860] "Smart glasses" refer to glasses-type wearable devices that are capable of displaying visual information and collecting data.

[1861] An "alert" refers to a warning message that is sent when a specific situation or condition occurs.

[1862] "Means of collecting feedback" refers to technologies and equipment used to gather opinions and evaluations from workers and managers.

[1863] "Means of improving system algorithms and threshold settings" refers to techniques and processes for optimizing system performance and settings based on feedback and data.

[1864] This invention relates to a system for monitoring the mental health of factory workers in real time and detecting stress levels. In particular, by combining it with an emotion engine, it is possible to accurately grasp the emotional state of workers and propose appropriate care. The following hardware and software are required to implement the system of this invention.

[1865] 1. Hardware

[1866] Factory robot: A robotic device that performs tasks within a factory.

[1867] Head-mounted displays (HMDs): For example, Microsoft HoloLens can be used to display visual information to workers and deliver notifications.

[1868] Smart glasses: For example, wearable devices that use Google Glass to collect voice and text data.

[1869] 2. Software

[1870] TensorFlow: Data analysis using natural language processing (NLP) technology.

[1871] IBM Watson: Used as an emotion engine to calculate emotion scores from collected data.

[1872] AWS: Manage databases and perform secure data transfer in the cloud.

[1873] Python: Creating programs for data preprocessing, analysis, and alert generation.

[1874] The system operates as follows:

[1875] 1. Data Collection

[1876] The system collects voice and text data from the terminal through the worker's HMD or smart glasses. For example, conversations and work reports made by workers during their work are collected as data.

[1877] 2. Data Transfer

[1878] The collected data is stored as a temporary file, encrypted, and securely transmitted to the AWS cloud using the HTTPS protocol.

[1879] 3. Data Analysis

[1880] Data is preprocessed on the cloud by a server, and text data is analyzed using NLP analysis with TensorFlow. In addition, IBM Watson's emotion engine analyzes audio data to calculate emotion scores and stress scores. For example, if many negative emotions such as "tired" or "irritated" are detected during work, the stress score will be high.

[1881] 4. Alert generation and notification

[1882] The server generates alerts based on emotion and stress scores, and sends alert messages containing appropriate care suggestions to the worker's HMD or smart glasses if the threshold is exceeded. For example, if the stress score exceeds the threshold, a message such as "We recommend you take a break" will be displayed.

[1883] 5. Gathering feedback and making improvements

[1884] The server collects feedback from workers and administrators, and uses it to improve the system's algorithms and threshold settings. This feedback allows the system to continuously improve its performance.

[1885] As a concrete example, consider a case where worker A at a factory wears an HMD while working. Daily voice and text data collected from this worker is transferred to the AWS cloud and analyzed by TensorFlow and IBM Watson. If the emotion score indicates "fatigue" or "dissatisfaction," and the stress score exceeds a threshold, an alert is displayed on the HMD. Subsequently, feedback from worker A is sent to the server, and the system is further optimized.

[1886] Examples of prompts to input into a generative AI model are as follows:

[1887] "Develop a system to monitor the mental health of factory workers and provide real-time care suggestions based on their stress levels. This system will use factory robots, HMDs, and smart glasses to collect and analyze employee voice and text data. It should calculate employee emotion and stress scores using natural language processing algorithms and an emotion recognition engine, and then provide appropriate care suggestions."

[1888] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1889] Step 1:

[1890] The terminal collects voice and text data from the worker's head-mounted display (HMD) or smart glasses. The input at this stage is voice and text communication data generated during the worker's daily work. The output is the collected data, which is stored as a temporary file.

[1891] Step 2:

[1892] The device encrypts the collected temporary file data and sends it to the AWS cloud using a secure protocol (e.g., HTTPS). The input at this stage is the unencrypted data stored in the temporary file, and the output is a notification that the transfer to the secure cloud is complete.

[1893] Step 3:

[1894] The server decrypts the data received in the cloud and performs preprocessing (noise reduction, format standardization). The input is encrypted collected data, and the output is preprocessed data. Specifically, background noise is removed from the audio data, and the format of the text data is standardized.

[1895] Step 4:

[1896] The server applies natural language processing (NLP) to preprocessed data to analyze the text data. The software used is TensorFlow; the input is preprocessed text data, and the output is the result of the text analysis. Specifically, it extracts emotional tone and the frequency of positive and negative language use.

[1897] Step 5:

[1898] The server uses an emotion engine (IBM Watson) to analyze voice data and calculate an emotion score. The input is pre-processed voice data, and the output is the emotion score. Specifically, it analyzes the intonation and tone of the voice to determine the emotional state.

[1899] Step 6:

[1900] The server calculates a stress score based on NLP analysis results and emotion engine results. The input is text analysis results and emotion score, and the output is the stress score. Specifically, it integrates the emotional tone of text and voice to assess the employee's stress level.

[1901] Step 7:

[1902] The server checks whether the calculated stress score exceeds a set threshold. The input is the stress score, and the output is a flag that instructs the server to generate an alert if the threshold is exceeded.

[1903] Step 8:

[1904] The server generates an alert when a threshold is exceeded and sends an alert message, including specific care suggestions, to the worker's HMD or smart glasses. The input is the result of the threshold being exceeded, and the output is the alert notification to the worker. Specifically, a message such as "We recommend you take a break" might be displayed.

[1905] Step 9:

[1906] The server collects feedback from workers and managers through HMDs and smart glasses, and uses this feedback to improve the system's algorithms and threshold settings. The input is feedback data, and the output is the improved algorithms and settings. Specifically, it analyzes the feedback and adjusts the algorithm parameters.

[1907] The following example prompts illustrate the specific actions required to perform each step:

[1908] "Develop a system to monitor the mental health of factory workers and provide real-time care suggestions based on their stress levels. This system will use factory robots, HMDs, and smart glasses to collect and analyze employee voice and text data. It should calculate employee emotion and stress scores using natural language processing algorithms and an emotion recognition engine, and then provide appropriate care suggestions."

[1909] 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.

[1910] 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.

[1911] 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 robot 414.

[1912] 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.

[1913] 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.

[1914] 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.

[1915] 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.

[1916] 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.

[1917] 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."

[1918] 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.

[1919] 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.

[1920] 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.

[1921] 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.

[1922] 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.

[1923] 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.

[1924] 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 per...

Claims

1. Means for monitoring the mental health status of employees, A method for analyzing employee communication data using natural language processing technology, A means for calculating an emotional score and a stress score based on the analysis results, A means of notifying the administrator and the employee themselves when the stress score exceeds a set threshold, Means of providing specific mental care suggestions, A system that includes this.

2. The system according to claim 1, wherein the communication data is at least one of email, chat, meeting recording, or voice memo.

3. The system according to claim 1, wherein the notification includes a suggestion to take a break and a link to consult a mental health professional.

Images