system

Abstract

Provide a system. [Solution] means for receiving brain waves as input; means for pre-processing the received electroencephalogram data; means for extracting features from the preprocessed data; a means for recognizing a user's intention using the extracted features; means for generating a natural-sounding output message based on the recognized intent; means for feeding back the generated output message to a user; A system including:

Description

[Technical Field]

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

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the 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] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] Conventional communication tools that use EEG as an input source have limited the output messages to short, fixed phrases such as "Yes," "No," "I want to go to the bathroom," and "Can I change my position?" This has prevented users from fully expressing complex emotions and needs, making it difficult to achieve natural, in-depth communication. This has significantly limited the quality of communication, especially for users with limited communication skills. The present invention aims to solve these problems. [Means for solving the problem]

[0005] The present invention provides a system including means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for recognizing a user's intention using the extracted features, means for generating a natural-sounding output message using a generative AI based on the recognized intention, and means for feeding back the generated output message to the user, thereby enabling users to communicate more naturally and in-depth than just using simple templated phrases.

[0006] "Electroencephalograms" are electrical signals generated in association with neural activity in the brain.

[0007] "Means for receiving" refers to the function of a device or software for capturing and incorporating data sent from outside.

[0008] "Preprocessing means" refers to the function of a device or software that performs initial processing to remove noise contained in data and extract useful information.

[0009] "Means for extracting features" refers to the function of a device or software that performs processing to extract essential information from data.

[0010] "Means for recognizing intentions" refers to the function of a device or software for determining the user's intentions or requests based on the extracted features.

[0011] "Generative AI" refers to a system that uses artificial intelligence technology to generate natural-sounding sentences based on input data.

[0012] "Output message" refers to text or audio information generated by the system and communicated to the user.

[0013] "Feedback means" refers to a display device or audio device for conveying the generated information to the user. [Brief explanation of the drawings]

[0014] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0015] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0016] First, the terms used in the following description will be explained.

[0017] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0018] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0019] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0020] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0021] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0022] [First embodiment]

[0023] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0024] 1, a 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.

[0025] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0026] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0027] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the 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.

[0028] 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 of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0029] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0030] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0031] 2, in the data processing device 12, a specific process 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" according to the technology of the present 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 process 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.

[0032] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0033] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0034] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0035] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intention, and generates a natural output message based on that intention and provides feedback to the user. This system is composed of a user, a terminal, and a server, and functions as follows.

[0036] System components and operation overview

[0037] User

[0038] The user wears an EEG sensor and can use it in a natural environment. The EEG generated by the user's thoughts is detected by the sensor in real time.

[0039] Terminal

[0040] The device receives real-time EEG data from an EEG sensor worn by the user. The received EEG data is pre-processed at regular intervals to remove noise.

[0041] server

[0042] The server performs the following steps:

[0043] Data preprocessing: Upon receiving the EEG data sent from the device, the server performs noise removal and filtering to improve the quality of the data.

[0044] Feature extraction: Extract essential features from the preprocessed data. These features are used for subsequent intent recognition.

[0045] Intention recognition: A machine learning model is used to analyze the extracted features and recognize the user's intention. The recognized intentions are basic expressions such as "yes," "no," "I want to go to the toilet," and "I want to change my position."

[0046] Message generation: The generative AI generates natural output messages based on the recognized intent. For example, in response to the intent "I want to change my position," the system generates the message "I'm not feeling well, so please change your position."

[0047] Message transmission: The generated output message is sent from the server to the terminal.

[0048] feedback

[0049] The device may display the received message to the user or provide audio feedback, for example by visually displaying the message on a display or by using speech synthesis to audibly convey the message.

[0050] Specific examples

[0051] Example 1: If the user thinks "Yes"

[0052] 1. Brainwave data collection: When the user thinks "yes," the brainwave sensor detects this and sends the data to the terminal.

[0053] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[0054] 3. Feature extraction: The server extracts features from the preprocessed data.

[0055] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "yes" from the features.

[0056] 5. Message generation: The generative AI generates natural messages such as "That's right."

[0057] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0058] Example 2: When the user thinks "I want to go to the toilet"

[0059] 1. Brainwave data collection: When the user thinks, "I want to go to the toilet," the brainwave sensor detects this and sends the data to the terminal.

[0060] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[0061] 3. Feature extraction: The server extracts features from the preprocessed data.

[0062] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "I want to go to the toilet" from the features.

[0063] 5. Message generation: The generative AI generates a natural message such as, "Sorry, but I need to go to the bathroom."

[0064] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0065] In this way, the system enables users to generate more natural and specific messages based on EEG data, rather than simple formulaic phrases, allowing for richer and more natural communication.

[0066] The processing flow will be explained below.

[0067] Step 1:

[0068] The user wears the EEG sensor and begins using it in a natural environment. The EEG sensor detects the user's brain waves in real time and converts them into digital signals.

[0069] Step 2:

[0070] The device receives digital signals from the EEG sensor. The received EEG data is stored in a buffer and passed on to the next process at regular intervals.

[0071] Step 3:

[0072] The device performs preprocessing on the received EEG data, which includes noise reduction using a bandpass filter. This filter emphasizes only specific frequency bands and removes unnecessary noise.

[0073] Step 4:

[0074] The device converts the preprocessed EEG data into features, for example, by extracting features in the time and frequency domains, thereby extracting essential information from the data.

[0075] Step 5:

[0076] The device sends the feature data to the server, which is used as important data for recognizing the user's intention.

[0077] Step 6:

[0078] The server receives the feature data and inputs it into a machine learning model. This model is pre-trained to recognize various intents (e.g., "yes," "no," "I want to go to the toilet") and recognizes the most appropriate intent from the input data.

[0079] Step 7:

[0080] The server uses generative AI to generate natural-sounding output messages based on the recognized intent. For example, for the intent "I want to go to the toilet," it generates the message "Sorry, but I'm starting to feel like I need to go to the toilet."

[0081] Step 8:

[0082] The server then sends the generated output message to the device, which also includes metadata such as the user ID and a timestamp.

[0083] Step 9:

[0084] The device then feeds back the received message to the user, for example by displaying the message visually on a display screen or by using voice synthesis technology to communicate the message aloud.

[0085] Step 10:

[0086] The user checks the feedback message from the terminal and decides on the next action, thereby realizing natural communication between the user and the external system.

[0087] Through the above steps, natural and intimate communication based on the user's brain wave data is realized.

[0088] Example 1

[0089] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0090] Conventional systems using EEG have had difficulty accurately recognizing the user's intentions and providing feedback in natural language. Furthermore, insufficient noise removal and data preprocessing can reduce the accuracy of intention recognition. This has led to the issue of preventing smooth communication between the user and the computer.

[0091] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0092] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for recognizing a user's intention using the recognized features, means for generating a natural-sounding output message based on the recognized intention using a generative AI model, and means for feeding back the generated output message to the user. This makes it possible to recognize the user's intention with high accuracy and feed back the user with a natural-sounding message.

[0093] "Electroencephalograms" are signals that refer to electrical activity that occurs in the brain when a user thinks or feels something.

[0094] "Electroencephalogram data" refers to data indicating the measurement results of electroencephalograms detected by an electroencephalogram sensor.

[0095] The "receiving means" is a hardware and software device for acquiring the electroencephalogram data sent from the electroencephalogram sensor.

[0096] The "pre-processing means" is a processing device that performs noise removal and filtering on the received electroencephalogram data to improve the quality of the data.

[0097] The "means for extracting features" refers to means for extracting essential features from preprocessed electroencephalogram data and using them for subsequent analysis.

[0098] The "means for recognizing the user's intention" is a means for analyzing the extracted features and determining what the user is thinking and what intentions they have.

[0099] A "generative AI model" is an artificial intelligence model that generates natural-looking language expressions based on specific input.

[0100] An "output message" is a natural language expression created by a generative AI model based on the recognized user intent.

[0101] The "feedback means" refers to a means for visually or audibly conveying the generated output message to the user.

[0102] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intention, generates a natural output message based on the intention, and provides feedback to the user. This system is composed of a user, a terminal, and a server.

[0103] User

[0104] The user wears an EEG sensor, which can be, for example, a general EEG sensor device (e.g., an electroencephalogram (EEG) measurement device). When the user thinks about an intention, such as "yes" or "I want to go to the toilet," EEG signals are generated based on that intention, and the EEG sensor detects the data in real time.

[0105] Terminal

[0106] The device receives EEG data from an EEG sensor worn by the user. For example, a small computer (e.g., a small, high-performance processing unit) is used. The device performs noise removal and preprocessing on the received EEG data. Preprocessing includes removing high-frequency noise and filtering. This uses scientific computing libraries (e.g., numerical calculation libraries and signal processing libraries). After preprocessing, the EEG data is sent to the server at regular intervals.

[0107] server

[0108] The server performs the following steps:

[0109] 1. Data preprocessing: The server receives the preprocessed EEG data sent from the device and performs further advanced noise removal and filtering, thereby improving the quality of the data.

[0110] 2. Feature extraction: Extract essential features from the preprocessed data. This feature extraction uses a library for large-scale data analysis (e.g., a machine learning library).

[0111] 3. Intention Recognition: The server uses a machine learning model (e.g., a machine learning algorithm) to recognize the user's intention from the extracted features. The recognized intentions are basic expressions such as "yes," "no," "I want to go to the toilet," and "I want to change my position."

[0112] 4. Message generation: The server uses a generative AI model (e.g., an artificial intelligence language generation model) to generate a natural-sounding output message based on the recognized intent. For example, for the intent "I want to change my position," the server generates the message "I'm not feeling well, so please change my position."

[0113] 5. Message transmission: The generated output message is sent from the server to the terminal.

[0114] feedback

[0115] The device then provides feedback to the user about the received message. Feedback can be provided using a visual display or a speech synthesis engine (e.g., a text-to-speech system). For example, the generated message can be displayed on an LCD display or spoken to the user.

[0116] Specific examples

[0117] Example 1: If the user thinks "Yes"

[0118] 1. Brainwave data collection: When the user thinks "yes," the brainwave sensor detects this and sends the data to the terminal.

[0119] 2. Preprocessing: Noise removal is performed on the device and the data is sent to the server.

[0120] 3. Feature extraction: The server extracts features from the preprocessed data.

[0121] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "yes" from the features.

[0122] 5. Message generation: The generative AI generates natural messages such as "That's right."

[0123] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0124] Example prompt sentence:

[0125] Your users are thinking "yes." Express that intent in natural language.

[0126] Example 2: When the user thinks "I want to go to the toilet"

[0127] 1. Brainwave data collection: When the user thinks, "I want to go to the toilet," the brainwave sensor detects this and sends the data to the terminal.

[0128] 2. Preprocessing: Noise removal is performed on the device and the data is sent to the server.

[0129] 3. Feature extraction: The server extracts features from the preprocessed data.

[0130] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "I want to go to the toilet" from the features.

[0131] 5. Message generation: The generative AI generates a natural message such as, "Sorry, but I need to go to the bathroom."

[0132] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0133] Example prompt sentence:

[0134] The user wants to "go to the bathroom." Express that intent in natural language.

[0135] In this way, the system enables users to generate more natural and specific messages based on EEG data, rather than simple formulaic phrases, allowing for more in-depth and natural communication.

[0136] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0137] Step 1:

[0138] The user wears an EEG sensor. When the user thinks about an intention, such as "yes" or "I want to go to the toilet," brain waves based on that intention are generated. The EEG sensor detects this brain wave data in real time and transmits it to the device. The input is the user's brain waves, and the output is brain wave data.

[0139] Step 2:

[0140] The device receives EEG data sent from the EEG sensor. It performs noise removal and preprocessing on the received data. For preprocessing, the numpy and scipy libraries are used to remove high-frequency noise and improve data quality. This process yields EEG data with noise removed. The input is raw data from the EEG sensor, and the output is preprocessed EEG data.

[0141] Step 3:

[0142] The device sends preprocessed EEG data to the server at regular intervals. It transfers data to the server using HTTP requests. The input is the preprocessed EEG data, and the output is a data packet sent to the server.

[0143] Step 4:

[0144] The server receives the preprocessed EEG data sent from the device. It then performs advanced noise removal and smoothing on the received data, further improving the accuracy of the data. The input is the preprocessed EEG data sent from the device, and the output is highly preprocessed EEG data.

[0145] Step 5:

[0146] The server extracts features from the highly preprocessed EEG data. This feature extraction uses machine learning libraries such as TENSORFLOW (registered trademark) and PyTorch. The extracted features are then used for intent recognition. The input is the highly preprocessed EEG data, and the output is the extracted features.

[0147] Step 6:

[0148] The server recognizes the user's intent based on the features. It uses a machine learning model (e.g., SVM or neural network) to analyze the features and determine the user's intent, such as "Yes" or "I want to go to the toilet." The input is the extracted features, and the output is the recognized user intent.

[0149] Step 7:

[0150] The server uses a generative AI model (e.g., OpenAI (registered trademark) generative model) to generate a natural-sounding output message based on the recognized user intent. For example, for the intent "I want to go to the toilet," the message "Sorry, but I'm starting to feel like I need to go to the toilet" is generated. The input is the recognized user intent, and the output is the generated output message.

[0151] Step 8:

[0152] The server sends the generated message to the terminal. It transfers the message to the terminal using an HTTP request. The input is the generated output message, and the output is the data packet sent to the terminal.

[0153] Step 9:

[0154] The terminal feeds back the received message to the user. The terminal displays the message on a display or speaks it aloud using a speech synthesis engine. For example, the terminal may display a message on an LCD display or speak "Sorry, but I need to go to the toilet." The input is the output message sent from the server, and the output is visual or spoken feedback to the user.

[0155] (Application example 1)

[0156] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0157] Conventional factory robots require manual input when operating, which reduces work efficiency. Furthermore, errors or delays that occur when workers operate robots can impair the productivity of the entire factory. To solve this problem, a more intuitive and faster way to operate robots is needed.

[0158] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0159] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for analyzing a user's intention and generating instructions for operating a machine based on the intention, and means for transmitting the generated operation instructions to the machine, thereby enabling a user to intuitively and quickly operate a robot using only their electroencephalograms.

[0160] 1. "EEG" refers to electrical signals generated by a user's brain activity.

[0161] 2. An "EEG sensor" is a device that is worn on the user's scalp to detect brain waves.

[0162] 3. "Preprocessing" refers to the process of removing noise from the received EEG data and improving the quality of the signal.

[0163] 4. "Features" are data that represent essential information extracted from preprocessed EEG data.

[0164] 5. "Intention recognition" is the process of using a machine learning model to identify the user's purpose and intention from extracted features.

[0165] 6. "Output message" is a naturally-expressed message generated based on the recognized user intent.

[0166] 7. "Feedback" is the means by which generated output messages are communicated to the user.

[0167] 8. "Instructions to operate a machine" are instructions to guide a machine, such as a factory robot, to perform a specific operation based on the user's intentions.

[0168] 9. "Real-time" means that EEG data is processed and analyzed in real time.

[0169] 10. A "machine learning model" is a model that uses statistical methods and algorithms to learn from data and automatically recognize patterns.

[0170] 11. "Server" refers to a central management system that performs processes such as preprocessing of EEG data, feature extraction, intention recognition, and message generation.

[0171] 12. A "factory robot" is a mechanical device used to automate work in a factory.

[0172] 13. "Generative AI" is an artificial intelligence system that generates natural-looking output messages based on the user's intent.

[0173] This invention is a system consisting of an EEG sensor worn by the user, a terminal that processes EEG data, and a server that performs data analysis and intention recognition. The user wears the EEG sensor, which detects the user's EEG in real time. This EEG data is transmitted to the terminal.

[0174] Processing by the terminal

[0175] The device receives the EEG data in real time and performs preprocessing such as noise removal and filtering, which ensures reliable data is sent to the server.

[0176] Server processing

[0177] The server receives the preprocessed EEG data sent from the device and extracts features, which are used in the subsequent intent recognition process.The server then analyzes the features using a machine learning model to recognize the user's intent.

[0178] Processing Recognized Intent

[0179] Based on the recognized intent, the generative AI model generates a natural-sounding output message. For example, if the intent is "I want to change my position," the model generates a natural-sounding message: "I'm not feeling well, so please change my position." This output message is then sent to the device.

[0180] User Feedback

[0181] The terminal feeds back the received output message to the user either visually on a display or audibly using speech synthesis.

[0182] Factory robot operation

[0183] Instructions generated based on the user's intentions are transmitted to the factory robot, enabling intuitive and rapid robot control through brain waves.

[0184] In an actual use case, if a user wants to move the robot to the right, the EEG sensor detects this and the pre-processed data is analyzed by the server. The server recognizes the intention of "move right" and generates an instruction based on that intention. Finally, the message "move right" is generated and sent to the robot.

[0185] An example prompt is:

[0186] The user wears an EEG sensor and consciously moves the robot to the right. EEG data from the sensor is collected in real time and pre-filtered to remove noise. The data obtained through feature extraction is analyzed by a machine learning model to recognize the user's intention to move the robot to the right. The message "Move right" is then generated and given as voice feedback.

[0187] In this way, the system utilizes the user's brainwaves to enable intuitive and efficient operation of factory robots.

[0188] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0189] Step 1:

[0190] The user wears the EEG sensor and starts collecting EEG data. The EEG sensor detects the user's EEG activity in real time. The input is the user's EEG signal, and the output is raw EEG data.

[0191] Step 2:

[0192] The terminal preprocesses the raw EEG data received in real time. This preprocessing includes noise removal and filtering. Specifically, it removes noise components using a high-frequency filter to improve signal quality. The input is raw EEG data, and the output is preprocessed EEG data.

[0193] Step 3:

[0194] The terminal sends the preprocessed EEG data to the server. The input is the preprocessed EEG data, and the output is the data sent to the server.

[0195] Step 4:

[0196] The server extracts features from the received preprocessed EEG data. Specifically, it analyzes the statistical characteristics of the signal and the frequency components using Fourier transform. The input is the preprocessed EEG data, and the output is feature data.

[0197] Step 5:

[0198] The server inputs the feature data into a machine learning model to recognize the user's intention. This process uses a model that has learned the correspondence between past EEG data and intention. The input is the feature data, and the output is the user's intention.

[0199] Step 6:

[0200] The server generates a natural-sounding output message using a generative AI model based on the recognized intent. For example, for the intent "Move the robot to the right," the server generates the message "Move to the right." The input is the user's intent, and the output is the output message.

[0201] Step 7:

[0202] The server generates and sends output messages to the terminal. The input is the output message, and the output is the message sent to the terminal.

[0203] Step 8:

[0204] The device then feeds back the received output message to the user. Specifically, the device displays the message on a display or transmits it audibly using speech synthesis. The input is the output message, and the output is visual or audio feedback.

[0205] Step 9:

[0206] Operation instructions generated based on the user's intentions are sent to the factory robot. The terminal or server sends specific operation instructions to the robot. The input is the operation instruction, and the output is the robot's operation.

[0207] This series of steps enables users to intuitively and quickly operate factory robots through their brain waves.

[0208] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0209] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intentions and emotions, generates natural output messages based on the intentions and emotions, and provides feedback to the user. This system is composed of a user, a terminal, and a server, and functions as follows.

[0210] System components and operation overview

[0211] User

[0212] The user wears an EEG sensor and can use it in a natural environment. The EEG generated by the user's thoughts is detected by the sensor in real time.

[0213] Terminal

[0214] The device receives real-time EEG data from an EEG sensor worn by the user. The received EEG data is pre-processed at regular intervals to remove noise.

[0215] server

[0216] The server performs the following steps:

[0217] Data preprocessing: Upon receiving the EEG data sent from the device, the server performs noise removal and filtering to improve the quality of the data.

[0218] Feature extraction: Extract essential features from the preprocessed data. These features are used for intent and emotion recognition.

[0219] Intention recognition: A machine learning model is used to analyze the extracted features and recognize the user's intention. The recognized intentions are basic expressions such as "yes," "no," "I want to go to the toilet," and "I want to change my position."

[0220] Emotion Recognition: The emotion engine is used to recognize the user's emotions from the extracted features, such as pleasant / unpleasant, joy, sadness, and anger.

[0221] Message generation: The generative AI generates natural-sounding output messages based on the recognized intent and emotion. For example, if the intent is "I want to go to the toilet" and the emotion is "impatience," the message generated is "Sorry, but I'm in a hurry to go to the toilet."

[0222] Message transmission: The generated output message is sent from the server to the terminal.

[0223] feedback

[0224] The device may display the received message to the user or provide audio feedback, for example by visually displaying the message on a display or by using speech synthesis to audibly convey the message.

[0225] Specific examples

[0226] Example 1: If the user thinks "Yes"

[0227] 1. Brainwave data collection: When the user thinks "yes," the brainwave sensor detects this and sends the data to the terminal.

[0228] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[0229] 3. Feature extraction: The server extracts features from the preprocessed data.

[0230] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "yes" from the features.

[0231] 5. Emotion Recognition: The server uses an emotion engine to recognize the user's emotions, such as "relief."

[0232] 6. Message generation: The generative AI generates natural messages such as "That's right, don't worry."

[0233] 7. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0234] Example 2: When the user thinks "I want to go to the toilet"

[0235] 1. Brainwave data collection: When the user thinks, "I want to go to the toilet," the brainwave sensor detects this and sends the data to the terminal.

[0236] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[0237] 3. Feature extraction: The server extracts features from the preprocessed data.

[0238] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "I want to go to the toilet" from the features.

[0239] 5. Emotion Recognition: The server uses an emotion engine to recognize the user's emotions, such as "impatience."

[0240] 6. Message generation: The generative AI generates a natural message such as, "Sorry, but I need to go to the bathroom quickly."

[0241] 7. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0242] In this way, the system enables users to communicate more effectively and naturally by generating more natural and specific messages based on EEG data and emotions, rather than simple formulaic phrases.

[0243] The processing flow will be explained below.

[0244] Step 1:

[0245] The user wears the EEG sensor and begins using it in a natural environment. The EEG sensor detects the user's brain waves in real time and converts them into digital signals.

[0246] Step 2:

[0247] The device receives digital signals from the EEG sensor. The received EEG data is stored in a buffer and passed on to the next process at regular intervals.

[0248] Step 3:

[0249] The device performs preprocessing on the received EEG data, which includes noise reduction using a bandpass filter. This filter emphasizes only specific frequency bands and removes unnecessary noise.

[0250] Step 4:

[0251] The device converts the preprocessed EEG data into features. Feature extraction uses techniques such as FFT (Fast Fourier Transform) to extract important features in the time and frequency domains. These features are used to recognize intentions and emotions.

[0252] Step 5:

[0253] The device transmits the feature data to the server, which is used as important data for recognizing the user's intentions and emotions.

[0254] Step 6:

[0255] The server receives the feature data and inputs it into a machine learning model. This model is pre-trained to recognize various intents (e.g., "yes," "no," "I want to go to the toilet") and recognizes the most appropriate intent from the input data.

[0256] Step 7:

[0257] At the same time, the server uses an emotion engine to recognize the user's emotions from the feature data. The emotion engine determines pleasant / unpleasant feelings, such as joy, sadness, and anger, in real time.

[0258] Step 8:

[0259] The server uses generative AI to generate natural output messages based on the recognized intent and emotion. For example, if the intent is "I want to go to the toilet" and the emotion is "impatience," the server generates a specific message such as "I'm sorry, but I feel like I need to go to the toilet quickly."

[0260] Step 9:

[0261] The server then sends the generated output message to the device, optionally including metadata such as a timestamp and user ID.

[0262] Step 10:

[0263] The device may then visually display the received message to the user or provide audio feedback, for example by displaying the message on a display screen or by using speech synthesis to communicate the message aloud.

[0264] Step 11:

[0265] The user checks the feedback message from the device and decides on the next action, thereby realizing natural and intimate communication between the user and the external system.

[0266] Example 2

[0267] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0268] Conventional intention recognition systems using EEG have difficulty accurately grasping a user's intention. There is a need for a system that can simultaneously recognize not only a user's intention but also their emotions, and generate more natural and appropriate output messages.

[0269] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0270] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for recognizing a user's intention using the extracted features, means for recognizing the user's emotion using the extracted features, means for generating a natural output message based on the recognized intention and emotion, and means for feeding back the generated output message to the user, thereby enabling the generation of a more natural and appropriate output message based on the user's intention and emotion.

[0271] "Electroencephalograms" are weak electrical potential changes resulting from electrical activity in the brain.

[0272] "Preprocessing" refers to processes such as data shaping, noise removal, and filtering that are carried out before data analysis.

[0273] A "feature" is significant information extracted from data for use in analysis by a machine learning model.

[0274] "Intention" refers to the specific intention or purpose of what the user wants or thinks.

[0275] "Emotion" refers to the sensations and psychological states that a user is experiencing.

[0276] An "output message" is a reply or feedback message to the user that is generated based on the recognized intent and emotion.

[0277] "Feedback" refers to the provision of information or responses from the system to the user.

[0278] A "machine learning model" is an algorithm that learns from data and uses the learning results to make predictions and classify new data.

[0279] A "generative AI model" is an artificial intelligence algorithm that generates natural-looking sentences based on the user's intentions and emotions, a type of generative artificial intelligence.

[0280] "Filtering" is the process of removing unwanted components from a data signal.

[0281] The present invention is a system that receives brain waves as input and recognizes a user's intentions and emotions. To implement this system, a user, a terminal, and a server work together. Each step uses specific hardware and software.

[0282] User

[0283] The user wears an EEG sensor and expresses their intentions and emotions naturally. This EEG sensor detects the electrical activity of the brain in real time and transmits the data to a terminal. A commercially available electroencephalograph can be used as the EEG sensor.

[0284] Terminal

[0285] The device receives EEG data in real time from an EEG sensor worn by the user. The received EEG data undergoes preprocessing such as noise removal, and the clean data is sent to the server. Specific software that can be used for the device's preprocessing includes MATLAB (registered trademark) and Python libraries (NumPy, SciPy).

[0286] server

[0287] The server receives the preprocessed EEG data sent from the device and performs the following data processing and calculations:

[0288] 1. Feature extraction:

[0289] The server extracts features from the preprocessed data, specifically, calculates the power spectrum for each frequency band using a Fourier transform, and analyzes the time-domain data using a machine learning model.

[0290] 2. Intention Recognition:

[0291] The user's intent is recognized from the extracted features using a machine learning model (e.g., random forest, support vector machine (SVM), deep learning, etc.). Examples of intent include "yes," "no," "I want to go to the toilet," and "I want to change my position."

[0292] 3. Emotion recognition:

[0293] The emotion engine is used to recognize the user's emotions from the extracted features. Examples of emotions include pleasant / unpleasant, joy, sadness, and anger. The emotion engine uses Microsoft® Azure® Emotion API and IBM Watson® Emotion Recognition.

[0294] 4. Message Creation:

[0295] A generative AI model (e.g., OpenAI's GPT-3 (registered trademark) or Google's BERT (registered trademark)) is used to generate natural-sounding output messages based on the recognized intent and emotion. By inputting a prompt sentence into the generative AI model, an appropriate output message is generated.

[0296] 5. Sending a message:

[0297] The generated output message is sent from the server to the terminal.

[0298] feedback

[0299] The device displays or provides audio feedback to the user based on the output messages received from the server. Visual feedback is displayed on the display, and audio feedback is provided using a speech synthesis engine (e.g., Google Text-to-Speech API or Amazon Polly).

[0300] Specific examples

[0301] Example 1: If the user thinks "Yes"

[0302] 1. The user thinks "yes."

[0303] 2. The EEG sensor detects brain waves and transmits them to the device.

[0304] 3. The device performs noise removal and sends the data to the server.

[0305] 4. The server extracts features from the preprocessed data.

[0306] 5. The machine learning model recognizes the intent "yes."

[0307] 6. The emotion engine recognizes the emotion of "relief."

[0308] 7. The generative AI model generates a message such as "That's right, don't worry."

[0309] 8. The terminal displays or audibly announces the generated message.

[0310] Prompt Sentence Examples

[0311] If the user thinks "yes": "Receive EEG data indicating that the user is thinking 'yes', recognize that intention, and generate a reassuring message."

[0312] When the user thinks, "I want to go to the toilet": "Receive EEG data when the user is thinking, 'I want to go to the toilet,' recognize that intention, and generate a message that includes a sense of urgency."

[0313] As a result, the present invention enables the generation of more natural and appropriate output messages based on the user's electroencephalogram data and emotions, providing a more effective means of communication.

[0314] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0315] Step 1:

[0316] The user wears an EEG sensor

[0317] The user wears an EEG sensor, which detects the user's brain waves in real time and converts them into digital signals, collecting brain wave data generated when the user thinks about something.

[0318] Input: User's brainwaves

[0319] Output: EEG data converted into digital signals

[0320] Specific behavior:

[0321] The user thinks, "I want to go to the toilet."

[0322] The EEG sensor detects brain waves and converts them into digital signals.

[0323] The converted digital signal is transmitted to the terminal.

[0324] Step 2:

[0325] The terminal receives and preprocesses the EEG data.

[0326] The device receives the digital signal sent from the EEG sensor and performs preprocessing such as filtering and noise removal to create clean data.

[0327] Input: Converted digital signal

[0328] Output: Preprocessed and clean EEG data

[0329] Specific behavior:

[0330] The device begins receiving brainwave data.

[0331] A band pass filter is applied to remove low and high frequency noise.

[0332] The preprocessed data is sent to the server.

[0333] Step 3:

[0334] The server extracts features

[0335] The server receives the preprocessed data sent from the device, calculates the power spectrum for each frequency band using Frue, and extracts important features from the received data.

[0336] Input: Preprocessed EEG data

[0337] Output: Extracted features

[0338] Specific behavior:

[0339] The server performs a Fourier transform on the pre-processed data.

[0340] Calculate the power spectral density of each frequency band.

[0341] Specific features are extracted and passed to the next processing step.

[0342] Step 4:

[0343] The server recognizes the user's intent

[0344] The server inputs the extracted features into a machine learning model to recognize the user's intent. The recognized intent indicates the meaning of thoughts and actions, such as "yes," "no," or "I want to go to the toilet."

[0345] Input: extracted features

[0346] Output: Recognized user intent

[0347] Specific behavior:

[0348] Input the features into a machine learning model (e.g., random forest or deep learning model).

[0349] The model performs analysis and recognizes the intention of "I want to go to the toilet."

[0350] The recognized intent is passed to the next processing step.

[0351] Step 5:

[0352] The server recognizes the user's emotions

[0353] The server uses the same features to input them into an emotion engine to recognize the user's emotions. The recognized emotions indicate the user's psychological state. Examples include "pleasant / unpleasant," "joy," "sadness," and "anger."

[0354] Input: extracted features, recognized intent

[0355] Output: Recognized user emotion

[0356] Specific behavior:

[0357] Input the features into a sentiment analysis model (e.g., Microsoft Azure Emotion API).

[0358] The model performs analysis and recognizes the emotion of "impatience."

[0359] The recognized emotion is passed on to the next processing step.

[0360] Step 6:

[0361] The server generates an output message

[0362] The server utilizes a generative AI model based on the recognized user's intention and emotion to generate natural-sounding output messages that reflect the user's intention and appropriately convey their emotions.

[0363] Input: Perceived Intent, Perceived Emotion

[0364] Output: The generated output message

[0365] Specific behavior:

[0366] The recognized intent "I want to go to the toilet" and emotion "impatience" are input as prompts into the generative AI model.

[0367] The generative AI model generates the message, "Sorry, but I need to go to the bathroom quickly."

[0368] Send the generated message to the terminal.

[0369] Step 7:

[0370] The device provides feedback to the user

[0371] The terminal provides the output messages received from the server to the user, visually on a display and using a speech synthesis engine for voice feedback.

[0372] Input: The generated output message

[0373] Output: Feedback provided to the user

[0374] Specific behavior:

[0375] The terminal receives the generated message.

[0376] The display will say, "Sorry, but I need to go to the toilet quickly."

[0377] Or use a text-to-speech engine to play the message aloud.

[0378] As described above, each processing step functions in cooperation with one another, and the present invention realizes the generation of a more natural and appropriate output message based on the user's electroencephalogram data and emotions.

[0379] (Application example 2)

[0380] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0381] In manufacturing sites such as factories, it is difficult to understand the intentions and emotions of workers in real time and respond appropriately accordingly. Rapid response is particularly important when workers become fatigued, stressed, or need help, but current systems cannot adequately meet these requirements. Solving this issue directly leads to improved safety and efficiency in the work environment.

[0382] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0383] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for recognizing a user's intention and emotion using the extracted features, means for generating a natural-sounding output message based on the recognized intention and emotion, means for feeding back the generated output message to the user, and means for adjusting the operation of the machine based on the generated output message. This makes it possible to understand the intention and emotion of the worker in real time and respond quickly and appropriately.

[0384] "Electroencephalograms" are electrical signals emitted by the human brain that reflect states such as thoughts and emotions.

[0385] "Means for receiving" refers to a device or process for receiving signals or data input from the outside.

[0386] A "preprocessing means" is a device or process that removes noise or filters the received data to make it easier to analyze.

[0387] A "means for extracting features" is a device or process that extracts essential information from raw data for use in subsequent analysis.

[0388] The "means for recognizing intentions" is a device or process for identifying the user's actions and wishes from the extracted features.

[0389] The "means for recognizing emotions" is a device or process for identifying the emotional state of the user from the extracted features.

[0390] A "means for generating an output message" is a device or process that creates an appropriate message for the user based on the perceived intent or emotion.

[0391] A "means for providing feedback to the user" is a device or process that visually or audibly conveys the generated output message to the user.

[0392] A "means for regulating the operation of a machine" is a device or process for controlling or modifying the operation of a machine based on the generated output messages.

[0393] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intentions and emotions, generates natural output messages based on the intentions and emotions, and then adjusts the operation of a machine based on the generated messages. This system is mainly composed of a user, a terminal, and a server.

[0394] User

[0395] The user wears an EEG sensor and can use it in a natural environment. The EEG generated by the user's thoughts is detected by the sensor in real time.

[0396] Terminal

[0397] The device receives real-time EEG data from an EEG sensor worn by the user. The received EEG data is pre-processed at regular intervals to remove noise.

[0398] server

[0399] The server performs the following steps:

[0400] 1. Data preprocessing: Upon receiving the EEG data sent from the device, the server performs preprocessing to improve the quality of the data by removing noise and filtering.

[0401] 2. Feature extraction: Extract essential features from the preprocessed data. These features are used for intent and emotion recognition.

[0402] 3. Intent and emotion recognition: The server uses a machine learning model (e.g., TensorFlow, PyTorch) and an emotion engine to analyze the extracted features and recognize the user's intent and emotion. Recognized intents include "I need help" and "I need a break," and emotions include "fatigue" and "stress."

[0403] 4. Message Generation: A generative AI (e.g., GPT-4®) generates a natural-sounding output message based on the perceived intent and sentiment. For example, a message such as "The worker is tired, so please slow down."

[0404] 5. Feedback: The server sends the generated output message to the terminal, which then feeds it back to the user and any associated machines, for example visually displaying it on a display or audibly using speech synthesis to communicate the message.

[0405] 6. Adjusting machine operation: Based on the generated output message, the terminal controls or changes the machine's operation, for example, slowing down the work pace or asking other workers for help.

[0406] Specific examples

[0407] Example 1: When a worker thinks they need help

[0408] 1. Brainwave data collection: When a worker thinks, "I need help," the brainwave sensor detects this and sends the data to the terminal.

[0409] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[0410] 3. Feature extraction: The server extracts features from the preprocessed data.

[0411] 4. Intent and emotion recognition: The server uses a machine learning model to recognize the intent of "need help" from the features, and an emotion engine to recognize the emotion of "tired."

[0412] 5. Message generation: The generation AI generates a natural message such as, "The worker needs help and seems tired. Please provide assistance immediately."

[0413] 6. Feedback: The terminal provides visual or audio feedback of the generated message and adjusts the machine's operation.

[0414] Example prompt sentence:

[0415] The intent of "I need help" and the emotion of "fatigue" have been recognized. Please respond.

[0416] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0417] Step 1:

[0418] The user wears an EEG sensor and works in a natural environment. The EEG sensor detects the brain waves generated by the user's thoughts in real time. The input is the user's brain waves, and the output is the brain wave data sent from the EEG sensor to the terminal.

[0419] Step 2:

[0420] The device acquires EEG data received from the EEG sensor in real time. The acquired EEG data is preprocessed at regular intervals to remove noise. Specifically, noise filtering and noise removal algorithms (e.g., bandpass filters) are applied. The input is raw EEG data, and the output is preprocessed EEG data.

[0421] Step 3:

[0422] The device sends the preprocessed EEG data to a server, which receives the preprocessed EEG data and performs additional noise removal and filtering to improve the quality of the data. The input is the preprocessed EEG data, and the output is the quality-improved EEG data.

[0423] Step 4:

[0424] The server extracts essential features from the improved EEG data. Specifically, it uses machine learning algorithms (e.g., PCA, ICA) to extract features from the EEG data. The input is the improved EEG data, and the output is feature data.

[0425] Step 5:

[0426] The server uses the extracted features to recognize the user's intention and emotion. It uses a machine learning model (e.g., TensorFlow, PyTorch) to recognize intents such as "I need help" and "I need a break" from the features, and an emotion engine to recognize emotions such as "fatigue" and "stress." The input is feature data, and the output is recognized intent and emotion data.

[0427] Step 6:

[0428] The server generates a natural-sounding output message using a generative AI model (e.g., GPT-4) based on the recognized intent and emotion. For example, a message such as "A worker needs help and appears tired. Please provide assistance immediately" is generated. The input is the recognized intent and emotion data, and the output is the generated output message.

[0429] Step 7:

[0430] The server sends the generated output message to the terminal. The terminal provides visual or audio feedback of the received message to the user. Visual feedback uses a display, and audio feedback uses speech synthesis technology. The input is the generated output message, and the output is the feedback to the user.

[0431] Step 8:

[0432] The terminal adjusts the machine's operation based on the generated output message. For example, it issues instructions to adjust the work pace or to ask other workers for help. The input is the generated output message, and the output is an instruction to control the machine's operation.

[0433] 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 a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the 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.

[0434] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0435] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0436] [Second embodiment]

[0437] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0438] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0439] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0440] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0441] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0442] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0443] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0444] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0445] The specific processing program 56 is an example of a "program" according to the technology of the present 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.

[0446] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0447] In the smart glasses 214, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0448] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. 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."

[0449] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intention, and generates a natural output message based on that intention and provides feedback to the user. This system is composed of a user, a terminal, and a server, and functions as follows.

[0450] System components and operation overview

[0451] User

[0452] The user wears an EEG sensor and can use it in a natural environment. The EEG generated by the user's thoughts is detected by the sensor in real time.

[0453] Terminal

[0454] The device receives real-time EEG data from an EEG sensor worn by the user. The received EEG data is pre-processed at regular intervals to remove noise.

[0455] server

[0456] The server performs the following steps:

[0457] Data preprocessing: Upon receiving the EEG data sent from the device, the server performs noise removal and filtering to improve the quality of the data.

[0458] Feature extraction: Extract essential features from the preprocessed data. These features are used for subsequent intent recognition.

[0459] Intention recognition: A machine learning model is used to analyze the extracted features and recognize the user's intention. The recognized intentions are basic expressions such as "yes," "no," "I want to go to the toilet," and "I want to change my position."

[0460] Message generation: The generative AI generates natural output messages based on the recognized intent. For example, in response to the intent "I want to change my position," the system generates the message "I'm not feeling well, so please change your position."

[0461] Message transmission: The generated output message is sent from the server to the terminal.

[0462] feedback

[0463] The device may display the received message to the user or provide audio feedback, for example by visually displaying the message on a display or by using speech synthesis to audibly convey the message.

[0464] Specific examples

[0465] Example 1: If the user thinks "Yes"

[0466] 1. Brainwave data collection: When the user thinks "yes," the brainwave sensor detects this and sends the data to the terminal.

[0467] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[0468] 3. Feature extraction: The server extracts features from the preprocessed data.

[0469] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "yes" from the features.

[0470] 5. Message generation: The generative AI generates natural messages such as "That's right."

[0471] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0472] Example 2: When the user thinks "I want to go to the toilet"

[0473] 1. Brainwave data collection: When the user thinks, "I want to go to the toilet," the brainwave sensor detects this and sends the data to the terminal.

[0474] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[0475] 3. Feature extraction: The server extracts features from the preprocessed data.

[0476] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "I want to go to the toilet" from the features.

[0477] 5. Message generation: The generative AI generates a natural message such as, "Sorry, but I need to go to the bathroom."

[0478] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0479] In this way, the system enables users to generate more natural and specific messages based on EEG data, rather than simple formulaic phrases, allowing for richer and more natural communication.

[0480] The processing flow will be explained below.

[0481] Step 1:

[0482] The user wears the EEG sensor and begins using it in a natural environment. The EEG sensor detects the user's brain waves in real time and converts them into digital signals.

[0483] Step 2:

[0484] The device receives digital signals from the EEG sensor. The received EEG data is stored in a buffer and passed on to the next process at regular intervals.

[0485] Step 3:

[0486] The device performs preprocessing on the received EEG data, which includes noise reduction using a bandpass filter. This filter emphasizes only specific frequency bands and removes unnecessary noise.

[0487] Step 4:

[0488] The device converts the preprocessed EEG data into features, for example, by extracting features in the time and frequency domains, thereby extracting essential information from the data.

[0489] Step 5:

[0490] The device sends the feature data to the server, which is used as important data for recognizing the user's intention.

[0491] Step 6:

[0492] The server receives the feature data and inputs it into a machine learning model. This model is pre-trained to recognize various intents (e.g., "yes," "no," "I want to go to the toilet") and recognizes the most appropriate intent from the input data.

[0493] Step 7:

[0494] The server uses generative AI to generate natural-sounding output messages based on the recognized intent. For example, for the intent "I want to go to the toilet," it generates the message "Sorry, but I'm starting to feel like I need to go to the toilet."

[0495] Step 8:

[0496] The server then sends the generated output message to the device, which also includes metadata such as the user ID and a timestamp.

[0497] Step 9:

[0498] The device then feeds back the received message to the user, for example by displaying the message visually on a display screen or by using voice synthesis technology to communicate the message aloud.

[0499] Step 10:

[0500] The user checks the feedback message from the terminal and decides on the next action, thereby realizing natural communication between the user and the external system.

[0501] Through the above steps, natural and intimate communication based on the user's brain wave data is realized.

[0502] Example 1

[0503] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0504] Conventional systems using EEG have had difficulty accurately recognizing the user's intentions and providing feedback in natural language. Furthermore, insufficient noise removal and data preprocessing can reduce the accuracy of intention recognition. This has led to the issue of preventing smooth communication between the user and the computer.

[0505] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0506] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for recognizing a user's intention using the recognized features, means for generating a natural-sounding output message based on the recognized intention using a generative AI model, and means for feeding back the generated output message to the user. This makes it possible to recognize the user's intention with high accuracy and feed back the user with a natural-sounding message.

[0507] "Electroencephalograms" are signals that refer to electrical activity that occurs in the brain when a user thinks or feels something.

[0508] "Electroencephalogram data" refers to data indicating the measurement results of electroencephalograms detected by an electroencephalogram sensor.

[0509] The "receiving means" is a hardware and software device for acquiring the electroencephalogram data sent from the electroencephalogram sensor.

[0510] The "pre-processing means" is a processing device that performs noise removal and filtering on the received electroencephalogram data to improve the quality of the data.

[0511] The "means for extracting features" refers to means for extracting essential features from preprocessed electroencephalogram data and using them for subsequent analysis.

[0512] The "means for recognizing the user's intention" is a means for analyzing the extracted features and determining what the user is thinking and what intentions they have.

[0513] A "generative AI model" is an artificial intelligence model that generates natural-looking language expressions based on specific input.

[0514] An "output message" is a natural language expression created by a generative AI model based on the recognized user intent.

[0515] The "feedback means" refers to a means for visually or audibly conveying the generated output message to the user.

[0516] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intention, generates a natural output message based on the intention, and provides feedback to the user. This system is composed of a user, a terminal, and a server.

[0517] User

[0518] The user wears an EEG sensor, which can be, for example, a general EEG sensor device (e.g., an electroencephalogram (EEG) measurement device). When the user thinks about an intention, such as "yes" or "I want to go to the toilet," EEG signals are generated based on that intention, and the EEG sensor detects the data in real time.

[0519] Terminal

[0520] The device receives EEG data from an EEG sensor worn by the user. For example, a small computer (e.g., a small, high-performance processing unit) is used. The device performs noise removal and preprocessing on the received EEG data. Preprocessing includes removing high-frequency noise and filtering. This uses scientific computing libraries (e.g., numerical calculation libraries and signal processing libraries). After preprocessing, the EEG data is sent to the server at regular intervals.

[0521] server

[0522] The server performs the following steps:

[0523] 1. Data preprocessing: The server receives the preprocessed EEG data sent from the device and performs further advanced noise removal and filtering, thereby improving the quality of the data.

[0524] 2. Feature extraction: Extract essential features from the preprocessed data. This feature extraction uses a library for large-scale data analysis (e.g., a machine learning library).

[0525] 3. Intention Recognition: The server uses a machine learning model (e.g., a machine learning algorithm) to recognize the user's intention from the extracted features. The recognized intentions are basic expressions such as "yes," "no," "I want to go to the toilet," and "I want to change my position."

[0526] 4. Message generation: The server uses a generative AI model (e.g., an artificial intelligence language generation model) to generate a natural-sounding output message based on the recognized intent. For example, for the intent "I want to change my position," the server generates the message "I'm not feeling well, so please change my position."

[0527] 5. Message transmission: The generated output message is sent from the server to the terminal.

[0528] feedback

[0529] The device then provides feedback to the user about the received message. Feedback can be provided using a visual display or a speech synthesis engine (e.g., a text-to-speech system). For example, the generated message can be displayed on an LCD display or spoken to the user.

[0530] Specific examples

[0531] Example 1: If the user thinks "Yes"

[0532] 1. Brainwave data collection: When the user thinks "yes," the brainwave sensor detects this and sends the data to the terminal.

[0533] 2. Preprocessing: Noise removal is performed on the device and the data is sent to the server.

[0534] 3. Feature extraction: The server extracts features from the preprocessed data.

[0535] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "yes" from the features.

[0536] 5. Message generation: The generative AI generates natural messages such as "That's right."

[0537] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0538] Example prompt sentence:

[0539] Your users are thinking "yes." Express that intent in natural language.

[0540] Example 2: When the user thinks "I want to go to the toilet"

[0541] 1. Brainwave data collection: When the user thinks, "I want to go to the toilet," the brainwave sensor detects this and sends the data to the terminal.

[0542] 2. Preprocessing: Noise removal is performed on the device and the data is sent to the server.

[0543] 3. Feature extraction: The server extracts features from the preprocessed data.

[0544] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "I want to go to the toilet" from the features.

[0545] 5. Message generation: The generative AI generates a natural message such as, "Sorry, but I need to go to the bathroom."

[0546] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0547] Example prompt sentence:

[0548] The user wants to "go to the bathroom." Express that intent in natural language.

[0549] In this way, the system enables users to generate more natural and specific messages based on EEG data, rather than simple formulaic phrases, allowing for more in-depth and natural communication.

[0550] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0551] Step 1:

[0552] The user wears an EEG sensor. When the user thinks about an intention, such as "yes" or "I want to go to the toilet," brain waves based on that intention are generated. The EEG sensor detects this brain wave data in real time and transmits it to the device. The input is the user's brain waves, and the output is brain wave data.

[0553] Step 2:

[0554] The device receives EEG data sent from the EEG sensor. It performs noise removal and preprocessing on the received data. For preprocessing, the numpy and scipy libraries are used to remove high-frequency noise and improve data quality. This process yields EEG data with noise removed. The input is raw data from the EEG sensor, and the output is preprocessed EEG data.

[0555] Step 3:

[0556] The device sends preprocessed EEG data to the server at regular intervals. It transfers data to the server using HTTP requests. The input is the preprocessed EEG data, and the output is a data packet sent to the server.

[0557] Step 4:

[0558] The server receives the preprocessed EEG data sent from the device. It then performs advanced noise removal and smoothing on the received data, further improving the accuracy of the data. The input is the preprocessed EEG data sent from the device, and the output is highly preprocessed EEG data.

[0559] Step 5:

[0560] The server extracts features from the highly preprocessed EEG data. This feature extraction uses machine learning libraries such as TensorFlow and PyTorch. The extracted features are then used for intent recognition. The input is the highly preprocessed EEG data, and the output is the extracted features.

[0561] Step 6:

[0562] The server recognizes the user's intent based on the features. It uses a machine learning model (e.g., SVM or neural network) to analyze the features and determine the user's intent, such as "Yes" or "I want to go to the toilet." The input is the extracted features, and the output is the recognized user intent.

[0563] Step 7:

[0564] The server uses a generative AI model (e.g., OpenAI's generative model) to generate a natural-sounding output message based on the recognized user intent. For example, for the intent "I want to go to the toilet," the message "Sorry, but I'm starting to feel like I need to go to the toilet" is generated. The input is the recognized user intent, and the output is the generated output message.

[0565] Step 8:

[0566] The server sends the generated message to the terminal. It transfers the message to the terminal using an HTTP request. The input is the generated output message, and the output is the data packet sent to the terminal.

[0567] Step 9:

[0568] The terminal feeds back the received message to the user. The terminal displays the message on a display or speaks it aloud using a speech synthesis engine. For example, the terminal may display a message on an LCD display or speak "Sorry, but I need to go to the toilet." The input is the output message sent from the server, and the output is visual or spoken feedback to the user.

[0569] (Application example 1)

[0570] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0571] Conventional factory robots require manual input when operating, which reduces work efficiency. Furthermore, errors or delays that occur when workers operate robots can impair the productivity of the entire factory. To solve this problem, a more intuitive and faster way to operate robots is needed.

[0572] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0573] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for analyzing a user's intention and generating instructions for operating a machine based on the intention, and means for transmitting the generated operation instructions to the machine, thereby enabling a user to intuitively and quickly operate a robot using only their electroencephalograms.

[0574] 1. "EEG" refers to electrical signals generated by a user's brain activity.

[0575] 2. An "EEG sensor" is a device that is worn on the user's scalp to detect brain waves.

[0576] 3. "Preprocessing" refers to the process of removing noise from the received EEG data and improving the quality of the signal.

[0577] 4. "Features" are data that represent essential information extracted from preprocessed EEG data.

[0578] 5. "Intention recognition" is the process of using a machine learning model to identify the user's purpose and intention from extracted features.

[0579] 6. "Output message" is a naturally-expressed message generated based on the recognized user intent.

[0580] 7. "Feedback" is the means by which generated output messages are communicated to the user.

[0581] 8. "Instructions to operate a machine" are instructions to guide a machine, such as a factory robot, to perform a specific operation based on the user's intentions.

[0582] 9. "Real-time" means that EEG data is processed and analyzed in real time.

[0583] 10. A "machine learning model" is a model that uses statistical methods and algorithms to learn from data and automatically recognize patterns.

[0584] 11. "Server" refers to a central management system that performs processes such as preprocessing of EEG data, feature extraction, intention recognition, and message generation.

[0585] 12. A "factory robot" is a mechanical device used to automate work in a factory.

[0586] 13. "Generative AI" is an artificial intelligence system that generates natural-looking output messages based on the user's intent.

[0587] This invention is a system consisting of an EEG sensor worn by the user, a terminal that processes EEG data, and a server that performs data analysis and intention recognition. The user wears the EEG sensor, which detects the user's EEG in real time. This EEG data is transmitted to the terminal.

[0588] Processing by the terminal

[0589] The device receives the EEG data in real time and performs preprocessing such as noise removal and filtering, which ensures reliable data is sent to the server.

[0590] Server processing

[0591] The server receives the preprocessed EEG data sent from the device and extracts features, which are used in the subsequent intent recognition process.The server then analyzes the features using a machine learning model to recognize the user's intent.

[0592] Processing Recognized Intent

[0593] Based on the recognized intent, the generative AI model generates a natural-sounding output message. For example, if the intent is "I want to change my position," the model generates a natural-sounding message: "I'm not feeling well, so please change my position." This output message is then sent to the device.

[0594] User Feedback

[0595] The terminal feeds back the received output message to the user either visually on a display or audibly using speech synthesis.

[0596] Factory robot operation

[0597] Instructions generated based on the user's intentions are transmitted to the factory robot, enabling intuitive and rapid robot control through brain waves.

[0598] In an actual use case, if a user wants to move the robot to the right, the EEG sensor detects this and the pre-processed data is analyzed by the server. The server recognizes the intention of "move right" and generates an instruction based on that intention. Finally, the message "move right" is generated and sent to the robot.

[0599] An example prompt is:

[0600] The user wears an EEG sensor and consciously moves the robot to the right. EEG data from the sensor is collected in real time and pre-filtered to remove noise. The data obtained through feature extraction is analyzed by a machine learning model to recognize the user's intention to move the robot to the right. The message "Move right" is then generated and given as voice feedback.

[0601] In this way, the system utilizes the user's brainwaves to enable intuitive and efficient operation of factory robots.

[0602] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0603] Step 1:

[0604] The user wears the EEG sensor and starts collecting EEG data. The EEG sensor detects the user's EEG activity in real time. The input is the user's EEG signal, and the output is raw EEG data.

[0605] Step 2:

[0606] The terminal preprocesses the raw EEG data received in real time. This preprocessing includes noise removal and filtering. Specifically, it removes noise components using a high-frequency filter to improve signal quality. The input is raw EEG data, and the output is preprocessed EEG data.

[0607] Step 3:

[0608] The terminal sends the preprocessed EEG data to the server. The input is the preprocessed EEG data, and the output is the data sent to the server.

[0609] Step 4:

[0610] The server extracts features from the received preprocessed EEG data. Specifically, it analyzes the statistical characteristics of the signal and the frequency components using Fourier transform. The input is the preprocessed EEG data, and the output is feature data.

[0611] Step 5:

[0612] The server inputs the feature data into a machine learning model to recognize the user's intention. This process uses a model that has learned the correspondence between past EEG data and intention. The input is the feature data, and the output is the user's intention.

[0613] Step 6:

[0614] The server generates a natural-sounding output message using a generative AI model based on the recognized intent. For example, for the intent "Move the robot to the right," the server generates the message "Move to the right." The input is the user's intent, and the output is the output message.

[0615] Step 7:

[0616] The server generates and sends output messages to the terminal. The input is the output message, and the output is the message sent to the terminal.

[0617] Step 8:

[0618] The device then feeds back the received output message to the user. Specifically, the device displays the message on a display or transmits it audibly using speech synthesis. The input is the output message, and the output is visual or audio feedback.

[0619] Step 9:

[0620] Operation instructions generated based on the user's intentions are sent to the factory robot. The terminal or server sends specific operation instructions to the robot. The input is the operation instruction, and the output is the robot's operation.

[0621] This series of steps enables users to intuitively and quickly operate factory robots through their brain waves.

[0622] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0623] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intentions and emotions, generates natural output messages based on the intentions and emotions, and provides feedback to the user. This system is composed of a user, a terminal, and a server, and functions as follows.

[0624] System components and operation overview

[0625] User

[0626] The user wears an EEG sensor and can use it in a natural environment. The EEG generated by the user's thoughts is detected by the sensor in real time.

[0627] Terminal

[0628] The device receives real-time EEG data from an EEG sensor worn by the user. The received EEG data is pre-processed at regular intervals to remove noise.

[0629] server

[0630] The server performs the following steps:

[0631] Data preprocessing: Upon receiving the EEG data sent from the device, the server performs noise removal and filtering to improve the quality of the data.

[0632] Feature extraction: Extract essential features from the preprocessed data. These features are used for intent and emotion recognition.

[0633] Intention recognition: A machine learning model is used to analyze the extracted features and recognize the user's intention. The recognized intentions are basic expressions such as "yes," "no," "I want to go to the toilet," and "I want to change my position."

[0634] Emotion Recognition: The emotion engine is used to recognize the user's emotions from the extracted features, such as pleasant / unpleasant, joy, sadness, and anger.

[0635] Message generation: The generative AI generates natural-sounding output messages based on the recognized intent and emotion. For example, if the intent is "I want to go to the toilet" and the emotion is "impatience," the message generated is "Sorry, but I'm in a hurry to go to the toilet."

[0636] Message transmission: The generated output message is sent from the server to the terminal.

[0637] feedback

[0638] The device may display the received message to the user or provide audio feedback, for example by visually displaying the message on a display or by using speech synthesis to audibly convey the message.

[0639] Specific examples

[0640] Example 1: If the user thinks "Yes"

[0641] 1. Brainwave data collection: When the user thinks "yes," the brainwave sensor detects this and sends the data to the terminal.

[0642] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[0643] 3. Feature extraction: The server extracts features from the preprocessed data.

[0644] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "yes" from the features.

[0645] 5. Emotion Recognition: The server uses an emotion engine to recognize the user's emotions, such as "relief."

[0646] 6. Message generation: The generative AI generates natural messages such as "That's right, don't worry."

[0647] 7. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0648] Example 2: When the user thinks "I want to go to the toilet"

[0649] 1. Brainwave data collection: When the user thinks, "I want to go to the toilet," the brainwave sensor detects this and sends the data to the terminal.

[0650] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[0651] 3. Feature extraction: The server extracts features from the preprocessed data.

[0652] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "I want to go to the toilet" from the features.

[0653] 5. Emotion Recognition: The server uses an emotion engine to recognize the user's emotions, such as "impatience."

[0654] 6. Message generation: The generative AI generates a natural message such as, "Sorry, but I need to go to the bathroom quickly."

[0655] 7. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0656] In this way, the system enables users to communicate more effectively and naturally by generating more natural and specific messages based on EEG data and emotions, rather than simple formulaic phrases.

[0657] The processing flow will be explained below.

[0658] Step 1:

[0659] The user wears the EEG sensor and begins using it in a natural environment. The EEG sensor detects the user's brain waves in real time and converts them into digital signals.

[0660] Step 2:

[0661] The device receives digital signals from the EEG sensor. The received EEG data is stored in a buffer and passed on to the next process at regular intervals.

[0662] Step 3:

[0663] The device performs preprocessing on the received EEG data, which includes noise reduction using a bandpass filter. This filter emphasizes only specific frequency bands and removes unnecessary noise.

[0664] Step 4:

[0665] The device converts the preprocessed EEG data into features. Feature extraction uses techniques such as FFT (Fast Fourier Transform) to extract important features in the time and frequency domains. These features are used to recognize intentions and emotions.

[0666] Step 5:

[0667] The device transmits the feature data to the server, which is used as important data for recognizing the user's intentions and emotions.

[0668] Step 6:

[0669] The server receives the feature data and inputs it into a machine learning model. This model is pre-trained to recognize various intents (e.g., "yes," "no," "I want to go to the toilet") and recognizes the most appropriate intent from the input data.

[0670] Step 7:

[0671] At the same time, the server uses an emotion engine to recognize the user's emotions from the feature data. The emotion engine determines pleasant / unpleasant feelings, such as joy, sadness, and anger, in real time.

[0672] Step 8:

[0673] The server uses generative AI to generate natural output messages based on the recognized intent and emotion. For example, if the intent is "I want to go to the toilet" and the emotion is "impatience," the server generates a specific message such as "I'm sorry, but I feel like I need to go to the toilet quickly."

[0674] Step 9:

[0675] The server then sends the generated output message to the device, optionally including metadata such as a timestamp and user ID.

[0676] Step 10:

[0677] The device may then visually display the received message to the user or provide audio feedback, for example by displaying the message on a display screen or by using speech synthesis to communicate the message aloud.

[0678] Step 11:

[0679] The user checks the feedback message from the device and decides on the next action, thereby realizing natural and intimate communication between the user and the external system.

[0680] Example 2

[0681] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0682] Conventional intention recognition systems using EEG have difficulty accurately grasping a user's intention. There is a need for a system that can simultaneously recognize not only a user's intention but also their emotions, and generate more natural and appropriate output messages.

[0683] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0684] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for recognizing a user's intention using the extracted features, means for recognizing the user's emotion using the extracted features, means for generating a natural output message based on the recognized intention and emotion, and means for feeding back the generated output message to the user, thereby enabling the generation of a more natural and appropriate output message based on the user's intention and emotion.

[0685] "Electroencephalograms" are weak electrical potential changes resulting from electrical activity in the brain.

[0686] "Preprocessing" refers to processes such as data shaping, noise removal, and filtering that are carried out before data analysis.

[0687] A "feature" is significant information extracted from data for use in analysis by a machine learning model.

[0688] "Intention" refers to the specific intention or purpose of what the user wants or thinks.

[0689] "Emotion" refers to the sensations and psychological states that a user is experiencing.

[0690] An "output message" is a reply or feedback message to the user that is generated based on the recognized intent and emotion.

[0691] "Feedback" refers to the provision of information or responses from the system to the user.

[0692] A "machine learning model" is an algorithm that learns from data and uses the learning results to make predictions and classify new data.

[0693] A "generative AI model" is an artificial intelligence algorithm that generates natural-looking sentences based on the user's intentions and emotions, a type of generative artificial intelligence.

[0694] "Filtering" is the process of removing unwanted components from a data signal.

[0695] The present invention is a system that receives brain waves as input and recognizes a user's intentions and emotions. To implement this system, a user, a terminal, and a server work together. Each step uses specific hardware and software.

[0696] User

[0697] The user wears an EEG sensor and expresses their intentions and emotions naturally. This EEG sensor detects the electrical activity of the brain in real time and transmits the data to a terminal. A commercially available electroencephalograph can be used as the EEG sensor.

[0698] Terminal

[0699] The device receives EEG data in real time from an EEG sensor worn by the user. The received EEG data undergoes preprocessing such as noise removal, and the clean data is sent to the server. Specific software that can be used for the device's preprocessing includes MATLAB and Python libraries (NumPy, SciPy).

[0700] server

[0701] The server receives the preprocessed EEG data sent from the device and performs the following data processing and calculations:

[0702] 1. Feature extraction:

[0703] The server extracts features from the preprocessed data, specifically, calculates the power spectrum for each frequency band using a Fourier transform, and analyzes the time-domain data using a machine learning model.

[0704] 2. Intention Recognition:

[0705] The user's intent is recognized from the extracted features using a machine learning model (e.g., random forest, support vector machine (SVM), deep learning, etc.). Examples of intent include "yes," "no," "I want to go to the toilet," and "I want to change my position."

[0706] 3. Emotion recognition:

[0707] The emotion engine is used to recognize the user's emotions from the extracted features. Examples of emotions include pleasant / unpleasant, joy, sadness, and anger. The emotion engine uses Microsoft Azure Emotion API and IBM Watson Emotion Recognition.

[0708] 4. Message Creation:

[0709] A generative AI model (e.g., OpenAI's GPT-3 or Google's BERT) is used to generate natural-sounding output messages based on the recognized intent and emotion. By inputting a prompt sentence into the generative AI model, an appropriate output message is generated.

[0710] 5. Sending a message:

[0711] The generated output message is sent from the server to the terminal.

[0712] feedback

[0713] The device displays or provides audio feedback to the user based on the output messages received from the server. Visual feedback is displayed on the display, and audio feedback is provided using a speech synthesis engine (e.g., Google Text-to-Speech API or Amazon Polly).

[0714] Specific examples

[0715] Example 1: If the user thinks "Yes"

[0716] 1. The user thinks "yes."

[0717] 2. The EEG sensor detects brain waves and transmits them to the device.

[0718] 3. The device performs noise removal and sends the data to the server.

[0719] 4. The server extracts features from the preprocessed data.

[0720] 5. The machine learning model recognizes the intent "yes."

[0721] 6. The emotion engine recognizes the emotion of "relief."

[0722] 7. The generative AI model generates a message such as "That's right, don't worry."

[0723] 8. The terminal displays or audibly announces the generated message.

[0724] Prompt Sentence Examples

[0725] If the user thinks "yes": "Receive EEG data indicating that the user is thinking 'yes', recognize that intention, and generate a reassuring message."

[0726] When the user thinks, "I want to go to the toilet": "Receive EEG data when the user is thinking, 'I want to go to the toilet,' recognize that intention, and generate a message that includes a sense of urgency."

[0727] As a result, the present invention enables the generation of more natural and appropriate output messages based on the user's electroencephalogram data and emotions, providing a more effective means of communication.

[0728] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0729] Step 1:

[0730] The user wears an EEG sensor

[0731] The user wears an EEG sensor, which detects the user's brain waves in real time and converts them into digital signals, collecting brain wave data generated when the user thinks about something.

[0732] Input: User's brainwaves

[0733] Output: EEG data converted into digital signals

[0734] Specific behavior:

[0735] The user thinks, "I want to go to the toilet."

[0736] The EEG sensor detects brain waves and converts them into digital signals.

[0737] The converted digital signal is transmitted to the terminal.

[0738] Step 2:

[0739] The terminal receives and preprocesses the EEG data.

[0740] The device receives the digital signal sent from the EEG sensor and performs preprocessing such as filtering and noise removal to create clean data.

[0741] Input: Converted digital signal

[0742] Output: Preprocessed and clean EEG data

[0743] Specific behavior:

[0744] The device begins receiving brainwave data.

[0745] A band pass filter is applied to remove low and high frequency noise.

[0746] The preprocessed data is sent to the server.

[0747] Step 3:

[0748] The server extracts features

[0749] The server receives the preprocessed data sent from the device, calculates the power spectrum for each frequency band using Frue, and extracts important features from the received data.

[0750] Input: Preprocessed EEG data

[0751] Output: Extracted features

[0752] Specific behavior:

[0753] The server performs a Fourier transform on the pre-processed data.

[0754] Calculate the power spectral density of each frequency band.

[0755] Specific features are extracted and passed to the next processing step.

[0756] Step 4:

[0757] The server recognizes the user's intent

[0758] The server inputs the extracted features into a machine learning model to recognize the user's intent. The recognized intent indicates the meaning of thoughts and actions, such as "yes," "no," or "I want to go to the toilet."

[0759] Input: extracted features

[0760] Output: Recognized user intent

[0761] Specific behavior:

[0762] Input the features into a machine learning model (e.g., random forest or deep learning model).

[0763] The model performs analysis and recognizes the intention of "I want to go to the toilet."

[0764] The recognized intent is passed to the next processing step.

[0765] Step 5:

[0766] The server recognizes the user's emotions

[0767] The server uses the same features to input them into an emotion engine to recognize the user's emotions. The recognized emotions indicate the user's psychological state. Examples include "pleasant / unpleasant," "joy," "sadness," and "anger."

[0768] Input: extracted features, recognized intent

[0769] Output: Recognized user emotion

[0770] Specific behavior:

[0771] Input the features into a sentiment analysis model (e.g., Microsoft Azure Emotion API).

[0772] The model performs analysis and recognizes the emotion of "impatience."

[0773] The recognized emotion is passed on to the next processing step.

[0774] Step 6:

[0775] The server generates an output message

[0776] The server utilizes a generative AI model based on the recognized user's intention and emotion to generate natural-sounding output messages that reflect the user's intention and appropriately convey their emotions.

[0777] Input: Perceived Intent, Perceived Emotion

[0778] Output: The generated output message

[0779] Specific behavior:

[0780] The recognized intent "I want to go to the toilet" and emotion "impatience" are input as prompts into the generative AI model.

[0781] The generative AI model generates the message, "Sorry, but I need to go to the bathroom quickly."

[0782] Send the generated message to the terminal.

[0783] Step 7:

[0784] The device provides feedback to the user

[0785] The terminal provides the output messages received from the server to the user, visually on a display and using a speech synthesis engine for voice feedback.

[0786] Input: The generated output message

[0787] Output: Feedback provided to the user

[0788] Specific behavior:

[0789] The terminal receives the generated message.

[0790] The display will say, "Sorry, but I need to go to the toilet quickly."

[0791] Or use a text-to-speech engine to play the message aloud.

[0792] As described above, each processing step functions in cooperation with one another, and the present invention realizes the generation of a more natural and appropriate output message based on the user's electroencephalogram data and emotions.

[0793] (Application example 2)

[0794] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0795] In manufacturing sites such as factories, it is difficult to understand the intentions and emotions of workers in real time and respond appropriately accordingly. Rapid response is particularly important when workers become fatigued, stressed, or need help, but current systems cannot adequately meet these requirements. Solving this issue directly leads to improved safety and efficiency in the work environment.

[0796] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0797] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for recognizing a user's intention and emotion using the extracted features, means for generating a natural-sounding output message based on the recognized intention and emotion, means for feeding back the generated output message to the user, and means for adjusting the operation of the machine based on the generated output message. This makes it possible to understand the intention and emotion of the worker in real time and respond quickly and appropriately.

[0798] "Electroencephalograms" are electrical signals emitted by the human brain that reflect states such as thoughts and emotions.

[0799] "Means for receiving" refers to a device or process for receiving signals or data input from the outside.

[0800] A "preprocessing means" is a device or process that removes noise or filters the received data to make it easier to analyze.

[0801] A "means for extracting features" is a device or process that extracts essential information from raw data for use in subsequent analysis.

[0802] The "means for recognizing intentions" is a device or process for identifying the user's actions and wishes from the extracted features.

[0803] The "means for recognizing emotions" is a device or process for identifying the emotional state of the user from the extracted features.

[0804] A "means for generating an output message" is a device or process that creates an appropriate message for the user based on the perceived intent or emotion.

[0805] A "means for providing feedback to the user" is a device or process that visually or audibly conveys the generated output message to the user.

[0806] A "means for regulating the operation of a machine" is a device or process for controlling or modifying the operation of a machine based on the generated output messages.

[0807] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intentions and emotions, generates natural output messages based on the intentions and emotions, and then adjusts the operation of a machine based on the generated messages. This system is mainly composed of a user, a terminal, and a server.

[0808] User

[0809] The user wears an EEG sensor and can use it in a natural environment. The EEG generated by the user's thoughts is detected by the sensor in real time.

[0810] Terminal

[0811] The device receives real-time EEG data from an EEG sensor worn by the user. The received EEG data is pre-processed at regular intervals to remove noise.

[0812] server

[0813] The server performs the following steps:

[0814] 1. Data preprocessing: Upon receiving the EEG data sent from the device, the server performs preprocessing to improve the quality of the data by removing noise and filtering.

[0815] 2. Feature extraction: Extract essential features from the preprocessed data. These features are used for intent and emotion recognition.

[0816] 3. Intent and emotion recognition: The server uses a machine learning model (e.g., TensorFlow, PyTorch) and an emotion engine to analyze the extracted features and recognize the user's intent and emotion. Recognized intents include "I need help" and "I need a break," and emotions include "fatigue" and "stress."

[0817] 4. Message generation: Generative AI (e.g., GPT-4) generates natural-sounding output messages based on the perceived intent and sentiment, such as "The worker is tired, so slow down."

[0818] 5. Feedback: The server sends the generated output message to the terminal, which then feeds it back to the user and any associated machines, for example visually displaying it on a display or audibly using speech synthesis to communicate the message.

[0819] 6. Adjusting machine operation: Based on the generated output message, the terminal controls or changes the machine's operation, for example, slowing down the work pace or asking other workers for help.

[0820] Specific examples

[0821] Example 1: When a worker thinks they need help

[0822] 1. Brainwave data collection: When a worker thinks, "I need help," the brainwave sensor detects this and sends the data to the terminal.

[0823] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[0824] 3. Feature extraction: The server extracts features from the preprocessed data.

[0825] 4. Intent and emotion recognition: The server uses a machine learning model to recognize the intent of "need help" from the features, and an emotion engine to recognize the emotion of "tired."

[0826] 5. Message generation: The generation AI generates a natural message such as, "The worker needs help and seems tired. Please provide assistance immediately."

[0827] 6. Feedback: The terminal provides visual or audio feedback of the generated message and adjusts the machine's operation.

[0828] Example prompt sentence:

[0829] The intent of "I need help" and the emotion of "fatigue" have been recognized. Please respond.

[0830] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0831] Step 1:

[0832] The user wears an EEG sensor and works in a natural environment. The EEG sensor detects the brain waves generated by the user's thoughts in real time. The input is the user's brain waves, and the output is the brain wave data sent from the EEG sensor to the terminal.

[0833] Step 2:

[0834] The device acquires EEG data received from the EEG sensor in real time. The acquired EEG data is preprocessed at regular intervals to remove noise. Specifically, noise filtering and noise removal algorithms (e.g., bandpass filters) are applied. The input is raw EEG data, and the output is preprocessed EEG data.

[0835] Step 3:

[0836] The device sends the preprocessed EEG data to a server, which receives the preprocessed EEG data and performs additional noise removal and filtering to improve the quality of the data. The input is the preprocessed EEG data, and the output is the quality-improved EEG data.

[0837] Step 4:

[0838] The server extracts essential features from the improved EEG data. Specifically, it uses machine learning algorithms (e.g., PCA, ICA) to extract features from the EEG data. The input is the improved EEG data, and the output is feature data.

[0839] Step 5:

[0840] The server uses the extracted features to recognize the user's intention and emotion. It uses a machine learning model (e.g., TensorFlow, PyTorch) to recognize intents such as "I need help" and "I need a break" from the features, and an emotion engine to recognize emotions such as "fatigue" and "stress." The input is feature data, and the output is recognized intent and emotion data.

[0841] Step 6:

[0842] The server generates a natural-sounding output message using a generative AI model (e.g., GPT-4) based on the recognized intent and emotion. For example, a message such as "A worker needs help and appears tired. Please provide assistance immediately" is generated. The input is the recognized intent and emotion data, and the output is the generated output message.

[0843] Step 7:

[0844] The server sends the generated output message to the terminal. The terminal provides visual or audio feedback of the received message to the user. Visual feedback uses a display, and audio feedback uses speech synthesis technology. The input is the generated output message, and the output is the feedback to the user.

[0845] Step 8:

[0846] The terminal adjusts the machine's operation based on the generated output message. For example, it issues instructions to adjust the work pace or to ask other workers for help. The input is the generated output message, and the output is an instruction to control the machine's operation.

[0847] 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 a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0848] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0849] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0850] [Third embodiment]

[0851] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0852] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0853] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0854] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0855] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0856] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0857] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0858] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0859] The specific processing program 56 is an example of a "program" according to the technology of the present 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.

[0860] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0861] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0862] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. 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."

[0863] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intention, and generates a natural output message based on that intention and provides feedback to the user. This system is composed of a user, a terminal, and a server, and functions as follows.

[0864] System components and operation overview

[0865] User

[0866] The user wears an EEG sensor and can use it in a natural environment. The EEG generated by the user's thoughts is detected by the sensor in real time.

[0867] Terminal

[0868] The device receives real-time EEG data from an EEG sensor worn by the user. The received EEG data is pre-processed at regular intervals to remove noise.

[0869] server

[0870] The server performs the following steps:

[0871] Data preprocessing: Upon receiving the EEG data sent from the device, the server performs noise removal and filtering to improve the quality of the data.

[0872] Feature extraction: Extract essential features from the preprocessed data. These features are used for subsequent intent recognition.

[0873] Intention recognition: A machine learning model is used to analyze the extracted features and recognize the user's intention. The recognized intentions are basic expressions such as "yes," "no," "I want to go to the toilet," and "I want to change my position."

[0874] Message generation: The generative AI generates natural output messages based on the recognized intent. For example, in response to the intent "I want to change my position," the system generates the message "I'm not feeling well, so please change your position."

[0875] Message transmission: The generated output message is sent from the server to the terminal.

[0876] feedback

[0877] The device may display the received message to the user or provide audio feedback, for example by visually displaying the message on a display or by using speech synthesis to audibly convey the message.

[0878] Specific examples

[0879] Example 1: If the user thinks "Yes"

[0880] 1. Brainwave data collection: When the user thinks "yes," the brainwave sensor detects this and sends the data to the terminal.

[0881] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[0882] 3. Feature extraction: The server extracts features from the preprocessed data.

[0883] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "yes" from the features.

[0884] 5. Message generation: The generative AI generates natural messages such as "That's right."

[0885] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0886] Example 2: When the user thinks "I want to go to the toilet"

[0887] 1. Brainwave data collection: When the user thinks, "I want to go to the toilet," the brainwave sensor detects this and sends the data to the terminal.

[0888] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[0889] 3. Feature extraction: The server extracts features from the preprocessed data.

[0890] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "I want to go to the toilet" from the features.

[0891] 5. Message generation: The generative AI generates a natural message such as, "Sorry, but I need to go to the bathroom."

[0892] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0893] In this way, the system enables users to generate more natural and specific messages based on EEG data, rather than simple formulaic phrases, allowing for richer and more natural communication.

[0894] The processing flow will be explained below.

[0895] Step 1:

[0896] The user wears the EEG sensor and begins using it in a natural environment. The EEG sensor detects the user's brain waves in real time and converts them into digital signals.

[0897] Step 2:

[0898] The device receives digital signals from the EEG sensor. The received EEG data is stored in a buffer and passed on to the next process at regular intervals.

[0899] Step 3:

[0900] The device performs preprocessing on the received EEG data, which includes noise reduction using a bandpass filter. This filter emphasizes only specific frequency bands and removes unnecessary noise.

[0901] Step 4:

[0902] The device converts the preprocessed EEG data into features, for example, by extracting features in the time and frequency domains, thereby extracting essential information from the data.

[0903] Step 5:

[0904] The device sends the feature data to the server, which is used as important data for recognizing the user's intention.

[0905] Step 6:

[0906] The server receives the feature data and inputs it into a machine learning model. This model is pre-trained to recognize various intents (e.g., "yes," "no," "I want to go to the toilet") and recognizes the most appropriate intent from the input data.

[0907] Step 7:

[0908] The server uses generative AI to generate natural-sounding output messages based on the recognized intent. For example, for the intent "I want to go to the toilet," it generates the message "Sorry, but I'm starting to feel like I need to go to the toilet."

[0909] Step 8:

[0910] The server then sends the generated output message to the device, which also includes metadata such as the user ID and a timestamp.

[0911] Step 9:

[0912] The device then feeds back the received message to the user, for example by displaying the message visually on a display screen or by using voice synthesis technology to communicate the message aloud.

[0913] Step 10:

[0914] The user checks the feedback message from the terminal and decides on the next action, thereby realizing natural communication between the user and the external system.

[0915] Through the above steps, natural and intimate communication based on the user's brain wave data is realized.

[0916] Example 1

[0917] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0918] Conventional systems using EEG have had difficulty accurately recognizing the user's intentions and providing feedback in natural language. Furthermore, insufficient noise removal and data preprocessing can reduce the accuracy of intention recognition. This has led to the issue of preventing smooth communication between the user and the computer.

[0919] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0920] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for recognizing a user's intention using the recognized features, means for generating a natural-sounding output message based on the recognized intention using a generative AI model, and means for feeding back the generated output message to the user. This makes it possible to recognize the user's intention with high accuracy and feed back the user with a natural-sounding message.

[0921] "Electroencephalograms" are signals that refer to electrical activity that occurs in the brain when a user thinks or feels something.

[0922] "Electroencephalogram data" refers to data indicating the measurement results of electroencephalograms detected by an electroencephalogram sensor.

[0923] The "receiving means" is a hardware and software device for acquiring the electroencephalogram data sent from the electroencephalogram sensor.

[0924] The "pre-processing means" is a processing device that performs noise removal and filtering on the received electroencephalogram data to improve the quality of the data.

[0925] The "means for extracting features" refers to means for extracting essential features from preprocessed electroencephalogram data and using them for subsequent analysis.

[0926] The "means for recognizing the user's intention" is a means for analyzing the extracted features and determining what the user is thinking and what intentions they have.

[0927] A "generative AI model" is an artificial intelligence model that generates natural-looking language expressions based on specific input.

[0928] An "output message" is a natural language expression created by a generative AI model based on the recognized user intent.

[0929] The "feedback means" refers to a means for visually or audibly conveying the generated output message to the user.

[0930] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intention, generates a natural output message based on the intention, and provides feedback to the user. This system is composed of a user, a terminal, and a server.

[0931] User

[0932] The user wears an EEG sensor, which can be, for example, a general EEG sensor device (e.g., an electroencephalogram (EEG) measurement device). When the user thinks about an intention, such as "yes" or "I want to go to the toilet," EEG signals are generated based on that intention, and the EEG sensor detects the data in real time.

[0933] Terminal

[0934] The device receives EEG data from an EEG sensor worn by the user. For example, a small computer (e.g., a small, high-performance processing unit) is used. The device performs noise removal and preprocessing on the received EEG data. Preprocessing includes removing high-frequency noise and filtering. This uses scientific computing libraries (e.g., numerical calculation libraries and signal processing libraries). After preprocessing, the EEG data is sent to the server at regular intervals.

[0935] server

[0936] The server performs the following steps:

[0937] 1. Data preprocessing: The server receives the preprocessed EEG data sent from the device and performs further advanced noise removal and filtering, thereby improving the quality of the data.

[0938] 2. Feature extraction: Extract essential features from the preprocessed data. This feature extraction uses a library for large-scale data analysis (e.g., a machine learning library).

[0939] 3. Intention Recognition: The server uses a machine learning model (e.g., a machine learning algorithm) to recognize the user's intention from the extracted features. The recognized intentions are basic expressions such as "yes," "no," "I want to go to the toilet," and "I want to change my position."

[0940] 4. Message generation: The server uses a generative AI model (e.g., an artificial intelligence language generation model) to generate a natural-sounding output message based on the recognized intent. For example, for the intent "I want to change my position," the server generates the message "I'm not feeling well, so please change my position."

[0941] 5. Message transmission: The generated output message is sent from the server to the terminal.

[0942] feedback

[0943] The device then provides feedback to the user about the received message. Feedback can be provided using a visual display or a speech synthesis engine (e.g., a text-to-speech system). For example, the generated message can be displayed on an LCD display or spoken to the user.

[0944] Specific examples

[0945] Example 1: If the user thinks "Yes"

[0946] 1. Brainwave data collection: When the user thinks "yes," the brainwave sensor detects this and sends the data to the terminal.

[0947] 2. Preprocessing: Noise removal is performed on the device and the data is sent to the server.

[0948] 3. Feature extraction: The server extracts features from the preprocessed data.

[0949] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "yes" from the features.

[0950] 5. Message generation: The generative AI generates natural messages such as "That's right."

[0951] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0952] Example prompt sentence:

[0953] Your users are thinking "yes." Express that intent in natural language.

[0954] Example 2: When the user thinks "I want to go to the toilet"

[0955] 1. Brainwave data collection: When the user thinks, "I want to go to the toilet," the brainwave sensor detects this and sends the data to the terminal.

[0956] 2. Preprocessing: Noise removal is performed on the device and the data is sent to the server.

[0957] 3. Feature extraction: The server extracts features from the preprocessed data.

[0958] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "I want to go to the toilet" from the features.

[0959] 5. Message generation: The generative AI generates a natural message such as, "Sorry, but I need to go to the bathroom."

[0960] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[0961] Example prompt sentence:

[0962] The user wants to "go to the bathroom." Express that intent in natural language.

[0963] In this way, the system enables users to generate more natural and specific messages based on EEG data, rather than simple formulaic phrases, allowing for more in-depth and natural communication.

[0964] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0965] Step 1:

[0966] The user wears an EEG sensor. When the user thinks about an intention, such as "yes" or "I want to go to the toilet," brain waves based on that intention are generated. The EEG sensor detects this brain wave data in real time and transmits it to the device. The input is the user's brain waves, and the output is brain wave data.

[0967] Step 2:

[0968] The device receives EEG data sent from the EEG sensor. It performs noise removal and preprocessing on the received data. For preprocessing, the numpy and scipy libraries are used to remove high-frequency noise and improve data quality. This process yields EEG data with noise removed. The input is raw data from the EEG sensor, and the output is preprocessed EEG data.

[0969] Step 3:

[0970] The device sends preprocessed EEG data to the server at regular intervals. It transfers data to the server using HTTP requests. The input is the preprocessed EEG data, and the output is a data packet sent to the server.

[0971] Step 4:

[0972] The server receives the preprocessed EEG data sent from the device. It then performs advanced noise removal and smoothing on the received data, further improving the accuracy of the data. The input is the preprocessed EEG data sent from the device, and the output is highly preprocessed EEG data.

[0973] Step 5:

[0974] The server extracts features from the highly preprocessed EEG data. This feature extraction uses machine learning libraries such as TensorFlow and PyTorch. The extracted features are then used for intent recognition. The input is the highly preprocessed EEG data, and the output is the extracted features.

[0975] Step 6:

[0976] The server recognizes the user's intent based on the features. It uses a machine learning model (e.g., SVM or neural network) to analyze the features and determine the user's intent, such as "Yes" or "I want to go to the toilet." The input is the extracted features, and the output is the recognized user intent.

[0977] Step 7:

[0978] The server uses a generative AI model (e.g., OpenAI's generative model) to generate a natural-sounding output message based on the recognized user intent. For example, for the intent "I want to go to the toilet," the message "Sorry, but I'm starting to feel like I need to go to the toilet" is generated. The input is the recognized user intent, and the output is the generated output message.

[0979] Step 8:

[0980] The server sends the generated message to the terminal. It transfers the message to the terminal using an HTTP request. The input is the generated output message, and the output is the data packet sent to the terminal.

[0981] Step 9:

[0982] The terminal feeds back the received message to the user. The terminal displays the message on a display or speaks it aloud using a speech synthesis engine. For example, the terminal may display a message on an LCD display or speak "Sorry, but I need to go to the toilet." The input is the output message sent from the server, and the output is visual or spoken feedback to the user.

[0983] (Application example 1)

[0984] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0985] Conventional factory robots require manual input when operating, which reduces work efficiency. Furthermore, errors or delays that occur when workers operate robots can impair the productivity of the entire factory. To solve this problem, a more intuitive and faster way to operate robots is needed.

[0986] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0987] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for analyzing a user's intention and generating instructions for operating a machine based on the intention, and means for transmitting the generated operation instructions to the machine, thereby enabling a user to intuitively and quickly operate a robot using only their electroencephalograms.

[0988] 1. "EEG" refers to electrical signals generated by a user's brain activity.

[0989] 2. An "EEG sensor" is a device that is worn on the user's scalp to detect brain waves.

[0990] 3. "Preprocessing" refers to the process of removing noise from the received EEG data and improving the quality of the signal.

[0991] 4. "Features" are data that represent essential information extracted from preprocessed EEG data.

[0992] 5. "Intention recognition" is the process of using a machine learning model to identify the user's purpose and intention from extracted features.

[0993] 6. "Output message" is a naturally-expressed message generated based on the recognized user intent.

[0994] 7. "Feedback" is the means by which generated output messages are communicated to the user.

[0995] 8. "Instructions to operate a machine" are instructions to guide a machine, such as a factory robot, to perform a specific operation based on the user's intentions.

[0996] 9. "Real-time" means that EEG data is processed and analyzed in real time.

[0997] 10. A "machine learning model" is a model that uses statistical methods and algorithms to learn from data and automatically recognize patterns.

[0998] 11. "Server" refers to a central management system that performs processes such as preprocessing of EEG data, feature extraction, intention recognition, and message generation.

[0999] 12. A "factory robot" is a mechanical device used to automate work in a factory.

[1000] 13. "Generative AI" is an artificial intelligence system that generates natural-looking output messages based on the user's intent.

[1001] This invention is a system consisting of an EEG sensor worn by the user, a terminal that processes EEG data, and a server that performs data analysis and intention recognition. The user wears the EEG sensor, which detects the user's EEG in real time. This EEG data is transmitted to the terminal.

[1002] Processing by the terminal

[1003] The device receives the EEG data in real time and performs preprocessing such as noise removal and filtering, which ensures reliable data is sent to the server.

[1004] Server processing

[1005] The server receives the preprocessed EEG data sent from the device and extracts features, which are used in the subsequent intent recognition process.The server then analyzes the features using a machine learning model to recognize the user's intent.

[1006] Processing Recognized Intent

[1007] Based on the recognized intent, the generative AI model generates a natural-sounding output message. For example, if the intent is "I want to change my position," the model generates a natural-sounding message: "I'm not feeling well, so please change my position." This output message is then sent to the device.

[1008] User Feedback

[1009] The terminal feeds back the received output message to the user either visually on a display or audibly using speech synthesis.

[1010] Factory robot operation

[1011] Instructions generated based on the user's intentions are transmitted to the factory robot, enabling intuitive and rapid robot control through brain waves.

[1012] In an actual use case, if a user wants to move the robot to the right, the EEG sensor detects this and the pre-processed data is analyzed by the server. The server recognizes the intention of "move right" and generates an instruction based on that intention. Finally, the message "move right" is generated and sent to the robot.

[1013] An example prompt is:

[1014] The user wears an EEG sensor and consciously moves the robot to the right. EEG data from the sensor is collected in real time and pre-filtered to remove noise. The data obtained through feature extraction is analyzed by a machine learning model to recognize the user's intention to move the robot to the right. The message "Move right" is then generated and given as voice feedback.

[1015] In this way, the system utilizes the user's brainwaves to enable intuitive and efficient operation of factory robots.

[1016] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1017] Step 1:

[1018] The user wears the EEG sensor and starts collecting EEG data. The EEG sensor detects the user's EEG activity in real time. The input is the user's EEG signal, and the output is raw EEG data.

[1019] Step 2:

[1020] The terminal preprocesses the raw EEG data received in real time. This preprocessing includes noise removal and filtering. Specifically, it removes noise components using a high-frequency filter to improve signal quality. The input is raw EEG data, and the output is preprocessed EEG data.

[1021] Step 3:

[1022] The terminal sends the preprocessed EEG data to the server. The input is the preprocessed EEG data, and the output is the data sent to the server.

[1023] Step 4:

[1024] The server extracts features from the received preprocessed EEG data. Specifically, it analyzes the statistical characteristics of the signal and the frequency components using Fourier transform. The input is the preprocessed EEG data, and the output is feature data.

[1025] Step 5:

[1026] The server inputs the feature data into a machine learning model to recognize the user's intention. This process uses a model that has learned the correspondence between past EEG data and intention. The input is the feature data, and the output is the user's intention.

[1027] Step 6:

[1028] The server generates a natural-sounding output message using a generative AI model based on the recognized intent. For example, for the intent "Move the robot to the right," the server generates the message "Move to the right." The input is the user's intent, and the output is the output message.

[1029] Step 7:

[1030] The server generates and sends output messages to the terminal. The input is the output message, and the output is the message sent to the terminal.

[1031] Step 8:

[1032] The device then feeds back the received output message to the user. Specifically, the device displays the message on a display or transmits it audibly using speech synthesis. The input is the output message, and the output is visual or audio feedback.

[1033] Step 9:

[1034] Operation instructions generated based on the user's intentions are sent to the factory robot. The terminal or server sends specific operation instructions to the robot. The input is the operation instruction, and the output is the robot's operation.

[1035] This series of steps enables users to intuitively and quickly operate factory robots through their brain waves.

[1036] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1037] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intentions and emotions, generates natural output messages based on the intentions and emotions, and provides feedback to the user. This system is composed of a user, a terminal, and a server, and functions as follows.

[1038] System components and operation overview

[1039] User

[1040] The user wears an EEG sensor and can use it in a natural environment. The EEG generated by the user's thoughts is detected by the sensor in real time.

[1041] Terminal

[1042] The device receives real-time EEG data from an EEG sensor worn by the user. The received EEG data is pre-processed at regular intervals to remove noise.

[1043] server

[1044] The server performs the following steps:

[1045] Data preprocessing: Upon receiving the EEG data sent from the device, the server performs noise removal and filtering to improve the quality of the data.

[1046] Feature extraction: Extract essential features from the preprocessed data. These features are used for intent and emotion recognition.

[1047] Intention recognition: A machine learning model is used to analyze the extracted features and recognize the user's intention. The recognized intentions are basic expressions such as "yes," "no," "I want to go to the toilet," and "I want to change my position."

[1048] Emotion Recognition: The emotion engine is used to recognize the user's emotions from the extracted features, such as pleasant / unpleasant, joy, sadness, and anger.

[1049] Message generation: The generative AI generates natural-sounding output messages based on the recognized intent and emotion. For example, if the intent is "I want to go to the toilet" and the emotion is "impatience," the message generated is "Sorry, but I'm in a hurry to go to the toilet."

[1050] Message transmission: The generated output message is sent from the server to the terminal.

[1051] feedback

[1052] The device may display the received message to the user or provide audio feedback, for example by visually displaying the message on a display or by using speech synthesis to audibly convey the message.

[1053] Specific examples

[1054] Example 1: If the user thinks "Yes"

[1055] 1. Brainwave data collection: When the user thinks "yes," the brainwave sensor detects this and sends the data to the terminal.

[1056] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[1057] 3. Feature extraction: The server extracts features from the preprocessed data.

[1058] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "yes" from the features.

[1059] 5. Emotion Recognition: The server uses an emotion engine to recognize the user's emotions, such as "relief."

[1060] 6. Message generation: The generative AI generates natural messages such as "That's right, don't worry."

[1061] 7. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[1062] Example 2: When the user thinks "I want to go to the toilet"

[1063] 1. Brainwave data collection: When the user thinks, "I want to go to the toilet," the brainwave sensor detects this and sends the data to the terminal.

[1064] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[1065] 3. Feature extraction: The server extracts features from the preprocessed data.

[1066] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "I want to go to the toilet" from the features.

[1067] 5. Emotion Recognition: The server uses an emotion engine to recognize the user's emotions, such as "impatience."

[1068] 6. Message generation: The generative AI generates a natural message such as, "Sorry, but I need to go to the bathroom quickly."

[1069] 7. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[1070] In this way, the system enables users to communicate more effectively and naturally by generating more natural and specific messages based on EEG data and emotions, rather than simple formulaic phrases.

[1071] The processing flow will be explained below.

[1072] Step 1:

[1073] The user wears the EEG sensor and begins using it in a natural environment. The EEG sensor detects the user's brain waves in real time and converts them into digital signals.

[1074] Step 2:

[1075] The device receives digital signals from the EEG sensor. The received EEG data is stored in a buffer and passed on to the next process at regular intervals.

[1076] Step 3:

[1077] The device performs preprocessing on the received EEG data, which includes noise reduction using a bandpass filter. This filter emphasizes only specific frequency bands and removes unnecessary noise.

[1078] Step 4:

[1079] The device converts the preprocessed EEG data into features. Feature extraction uses techniques such as FFT (Fast Fourier Transform) to extract important features in the time and frequency domains. These features are used to recognize intentions and emotions.

[1080] Step 5:

[1081] The device transmits the feature data to the server, which is used as important data for recognizing the user's intentions and emotions.

[1082] Step 6:

[1083] The server receives the feature data and inputs it into a machine learning model. This model is pre-trained to recognize various intents (e.g., "yes," "no," "I want to go to the toilet") and recognizes the most appropriate intent from the input data.

[1084] Step 7:

[1085] At the same time, the server uses an emotion engine to recognize the user's emotions from the feature data. The emotion engine determines pleasant / unpleasant feelings, such as joy, sadness, and anger, in real time.

[1086] Step 8:

[1087] The server uses generative AI to generate natural output messages based on the recognized intent and emotion. For example, if the intent is "I want to go to the toilet" and the emotion is "impatience," the server generates a specific message such as "I'm sorry, but I feel like I need to go to the toilet quickly."

[1088] Step 9:

[1089] The server then sends the generated output message to the device, optionally including metadata such as a timestamp and user ID.

[1090] Step 10:

[1091] The device may then visually display the received message to the user or provide audio feedback, for example by displaying the message on a display screen or by using speech synthesis to communicate the message aloud.

[1092] Step 11:

[1093] The user checks the feedback message from the device and decides on the next action, thereby realizing natural and intimate communication between the user and the external system.

[1094] Example 2

[1095] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1096] Conventional intention recognition systems using EEG have difficulty accurately grasping a user's intention. There is a need for a system that can simultaneously recognize not only a user's intention but also their emotions, and generate more natural and appropriate output messages.

[1097] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1098] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for recognizing a user's intention using the extracted features, means for recognizing the user's emotion using the extracted features, means for generating a natural output message based on the recognized intention and emotion, and means for feeding back the generated output message to the user, thereby enabling the generation of a more natural and appropriate output message based on the user's intention and emotion.

[1099] "Electroencephalograms" are weak electrical potential changes resulting from electrical activity in the brain.

[1100] "Preprocessing" refers to processes such as data shaping, noise removal, and filtering that are carried out before data analysis.

[1101] A "feature" is significant information extracted from data for use in analysis by a machine learning model.

[1102] "Intention" refers to the specific intention or purpose of what the user wants or thinks.

[1103] "Emotion" refers to the sensations and psychological states that a user is experiencing.

[1104] An "output message" is a reply or feedback message to the user that is generated based on the recognized intent and emotion.

[1105] "Feedback" refers to the provision of information or responses from the system to the user.

[1106] A "machine learning model" is an algorithm that learns from data and uses the learning results to make predictions and classify new data.

[1107] A "generative AI model" is an artificial intelligence algorithm that generates natural-looking sentences based on the user's intentions and emotions, a type of generative artificial intelligence.

[1108] "Filtering" is the process of removing unwanted components from a data signal.

[1109] The present invention is a system that receives brain waves as input and recognizes a user's intentions and emotions. To implement this system, a user, a terminal, and a server work together. Each step uses specific hardware and software.

[1110] User

[1111] The user wears an EEG sensor and expresses their intentions and emotions naturally. This EEG sensor detects the electrical activity of the brain in real time and transmits the data to a terminal. A commercially available electroencephalograph can be used as the EEG sensor.

[1112] Terminal

[1113] The device receives EEG data in real time from an EEG sensor worn by the user. The received EEG data undergoes preprocessing such as noise removal, and the clean data is sent to the server. Specific software that can be used for the device's preprocessing includes MATLAB and Python libraries (NumPy, SciPy).

[1114] server

[1115] The server receives the preprocessed EEG data sent from the device and performs the following data processing and calculations:

[1116] 1. Feature extraction:

[1117] The server extracts features from the preprocessed data, specifically, calculates the power spectrum for each frequency band using a Fourier transform, and analyzes the time-domain data using a machine learning model.

[1118] 2. Intention Recognition:

[1119] The user's intent is recognized from the extracted features using a machine learning model (e.g., random forest, support vector machine (SVM), deep learning, etc.). Examples of intent include "yes," "no," "I want to go to the toilet," and "I want to change my position."

[1120] 3. Emotion recognition:

[1121] The emotion engine is used to recognize the user's emotions from the extracted features. Examples of emotions include pleasant / unpleasant, joy, sadness, and anger. The emotion engine uses Microsoft Azure Emotion API and IBM Watson Emotion Recognition.

[1122] 4. Message Creation:

[1123] A generative AI model (e.g., OpenAI's GPT-3 or Google's BERT) is used to generate natural-sounding output messages based on the recognized intent and emotion. By inputting a prompt sentence into the generative AI model, an appropriate output message is generated.

[1124] 5. Sending a message:

[1125] The generated output message is sent from the server to the terminal.

[1126] feedback

[1127] The device displays or provides audio feedback to the user based on the output messages received from the server. Visual feedback is displayed on the display, and audio feedback is provided using a speech synthesis engine (e.g., Google Text-to-Speech API or Amazon Polly).

[1128] Specific examples

[1129] Example 1: If the user thinks "Yes"

[1130] 1. The user thinks "yes."

[1131] 2. The EEG sensor detects brain waves and transmits them to the device.

[1132] 3. The device performs noise removal and sends the data to the server.

[1133] 4. The server extracts features from the preprocessed data.

[1134] 5. The machine learning model recognizes the intent "yes."

[1135] 6. The emotion engine recognizes the emotion of "relief."

[1136] 7. The generative AI model generates a message such as "That's right, don't worry."

[1137] 8. The terminal displays or audibly announces the generated message.

[1138] Prompt Sentence Examples

[1139] If the user thinks "yes": "Receive EEG data indicating that the user is thinking 'yes', recognize that intention, and generate a reassuring message."

[1140] When the user thinks, "I want to go to the toilet": "Receive EEG data when the user is thinking, 'I want to go to the toilet,' recognize that intention, and generate a message that includes a sense of urgency."

[1141] As a result, the present invention enables the generation of more natural and appropriate output messages based on the user's electroencephalogram data and emotions, providing a more effective means of communication.

[1142] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1143] Step 1:

[1144] The user wears an EEG sensor

[1145] The user wears an EEG sensor, which detects the user's brain waves in real time and converts them into digital signals, collecting brain wave data generated when the user thinks about something.

[1146] Input: User's brainwaves

[1147] Output: EEG data converted into digital signals

[1148] Specific behavior:

[1149] The user thinks, "I want to go to the toilet."

[1150] The EEG sensor detects brain waves and converts them into digital signals.

[1151] The converted digital signal is transmitted to the terminal.

[1152] Step 2:

[1153] The terminal receives and preprocesses the EEG data.

[1154] The device receives the digital signal sent from the EEG sensor and performs preprocessing such as filtering and noise removal to create clean data.

[1155] Input: Converted digital signal

[1156] Output: Preprocessed and clean EEG data

[1157] Specific behavior:

[1158] The device begins receiving brainwave data.

[1159] A band pass filter is applied to remove low and high frequency noise.

[1160] The preprocessed data is sent to the server.

[1161] Step 3:

[1162] The server extracts features

[1163] The server receives the preprocessed data sent from the device, calculates the power spectrum for each frequency band using Frue, and extracts important features from the received data.

[1164] Input: Preprocessed EEG data

[1165] Output: Extracted features

[1166] Specific behavior:

[1167] The server performs a Fourier transform on the pre-processed data.

[1168] Calculate the power spectral density of each frequency band.

[1169] Specific features are extracted and passed to the next processing step.

[1170] Step 4:

[1171] The server recognizes the user's intent

[1172] The server inputs the extracted features into a machine learning model to recognize the user's intent. The recognized intent indicates the meaning of thoughts and actions, such as "yes," "no," or "I want to go to the toilet."

[1173] Input: extracted features

[1174] Output: Recognized user intent

[1175] Specific behavior:

[1176] Input the features into a machine learning model (e.g., random forest or deep learning model).

[1177] The model performs analysis and recognizes the intention of "I want to go to the toilet."

[1178] The recognized intent is passed to the next processing step.

[1179] Step 5:

[1180] The server recognizes the user's emotions

[1181] The server uses the same features to input them into an emotion engine to recognize the user's emotions. The recognized emotions indicate the user's psychological state. Examples include "pleasant / unpleasant," "joy," "sadness," and "anger."

[1182] Input: extracted features, recognized intent

[1183] Output: Recognized user emotion

[1184] Specific behavior:

[1185] Input the features into a sentiment analysis model (e.g., Microsoft Azure Emotion API).

[1186] The model performs analysis and recognizes the emotion of "impatience."

[1187] The recognized emotion is passed on to the next processing step.

[1188] Step 6:

[1189] The server generates an output message

[1190] The server utilizes a generative AI model based on the recognized user's intention and emotion to generate natural-sounding output messages that reflect the user's intention and appropriately convey their emotions.

[1191] Input: Perceived Intent, Perceived Emotion

[1192] Output: The generated output message

[1193] Specific behavior:

[1194] The recognized intent "I want to go to the toilet" and emotion "impatience" are input as prompts into the generative AI model.

[1195] The generative AI model generates the message, "Sorry, but I need to go to the bathroom quickly."

[1196] Send the generated message to the terminal.

[1197] Step 7:

[1198] The device provides feedback to the user

[1199] The terminal provides the output messages received from the server to the user, visually on a display and using a speech synthesis engine for voice feedback.

[1200] Input: The generated output message

[1201] Output: Feedback provided to the user

[1202] Specific behavior:

[1203] The terminal receives the generated message.

[1204] The display will say, "Sorry, but I need to go to the toilet quickly."

[1205] Or use a text-to-speech engine to play the message aloud.

[1206] As described above, each processing step functions in cooperation with one another, and the present invention realizes the generation of a more natural and appropriate output message based on the user's electroencephalogram data and emotions.

[1207] (Application example 2)

[1208] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1209] In manufacturing sites such as factories, it is difficult to understand the intentions and emotions of workers in real time and respond appropriately accordingly. Rapid response is particularly important when workers become fatigued, stressed, or need help, but current systems cannot adequately meet these requirements. Solving this issue directly leads to improved safety and efficiency in the work environment.

[1210] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1211] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for recognizing a user's intention and emotion using the extracted features, means for generating a natural-sounding output message based on the recognized intention and emotion, means for feeding back the generated output message to the user, and means for adjusting the operation of the machine based on the generated output message. This makes it possible to understand the intention and emotion of the worker in real time and respond quickly and appropriately.

[1212] "Electroencephalograms" are electrical signals emitted by the human brain that reflect states such as thoughts and emotions.

[1213] "Means for receiving" refers to a device or process for receiving signals or data input from the outside.

[1214] A "preprocessing means" is a device or process that removes noise or filters the received data to make it easier to analyze.

[1215] A "means for extracting features" is a device or process that extracts essential information from raw data for use in subsequent analysis.

[1216] The "means for recognizing intentions" is a device or process for identifying the user's actions and wishes from the extracted features.

[1217] The "means for recognizing emotions" is a device or process for identifying the emotional state of the user from the extracted features.

[1218] A "means for generating an output message" is a device or process that creates an appropriate message for the user based on the perceived intent or emotion.

[1219] A "means for providing feedback to the user" is a device or process that visually or audibly conveys the generated output message to the user.

[1220] A "means for regulating the operation of a machine" is a device or process for controlling or modifying the operation of a machine based on the generated output messages.

[1221] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intentions and emotions, generates natural output messages based on the intentions and emotions, and then adjusts the operation of a machine based on the generated messages. This system is mainly composed of a user, a terminal, and a server.

[1222] User

[1223] The user wears an EEG sensor and can use it in a natural environment. The EEG generated by the user's thoughts is detected by the sensor in real time.

[1224] Terminal

[1225] The device receives real-time EEG data from an EEG sensor worn by the user. The received EEG data is pre-processed at regular intervals to remove noise.

[1226] server

[1227] The server performs the following steps:

[1228] 1. Data preprocessing: Upon receiving the EEG data sent from the device, the server performs preprocessing to improve the quality of the data by removing noise and filtering.

[1229] 2. Feature extraction: Extract essential features from the preprocessed data. These features are used for intent and emotion recognition.

[1230] 3. Intent and emotion recognition: The server uses a machine learning model (e.g., TensorFlow, PyTorch) and an emotion engine to analyze the extracted features and recognize the user's intent and emotion. Recognized intents include "I need help" and "I need a break," and emotions include "fatigue" and "stress."

[1231] 4. Message generation: Generative AI (e.g., GPT-4) generates natural-sounding output messages based on the perceived intent and sentiment, such as "The worker is tired, so slow down."

[1232] 5. Feedback: The server sends the generated output message to the terminal, which then feeds it back to the user and any associated machines, for example visually displaying it on a display or audibly using speech synthesis to communicate the message.

[1233] 6. Adjusting machine operation: Based on the generated output message, the terminal controls or changes the machine's operation, for example, slowing down the work pace or asking other workers for help.

[1234] Specific examples

[1235] Example 1: When a worker thinks they need help

[1236] 1. Brainwave data collection: When a worker thinks, "I need help," the brainwave sensor detects this and sends the data to the terminal.

[1237] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[1238] 3. Feature extraction: The server extracts features from the preprocessed data.

[1239] 4. Intent and emotion recognition: The server uses a machine learning model to recognize the intent of "need help" from the features, and an emotion engine to recognize the emotion of "tired."

[1240] 5. Message generation: The generation AI generates a natural message such as, "The worker needs help and seems tired. Please provide assistance immediately."

[1241] 6. Feedback: The terminal provides visual or audio feedback of the generated message and adjusts the machine's operation.

[1242] Example prompt sentence:

[1243] The intent of "I need help" and the emotion of "fatigue" have been recognized. Please respond.

[1244] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1245] Step 1:

[1246] The user wears an EEG sensor and works in a natural environment. The EEG sensor detects the brain waves generated by the user's thoughts in real time. The input is the user's brain waves, and the output is the brain wave data sent from the EEG sensor to the terminal.

[1247] Step 2:

[1248] The device acquires EEG data received from the EEG sensor in real time. The acquired EEG data is preprocessed at regular intervals to remove noise. Specifically, noise filtering and noise removal algorithms (e.g., bandpass filters) are applied. The input is raw EEG data, and the output is preprocessed EEG data.

[1249] Step 3:

[1250] The device sends the preprocessed EEG data to a server, which receives the preprocessed EEG data and performs additional noise removal and filtering to improve the quality of the data. The input is the preprocessed EEG data, and the output is the quality-improved EEG data.

[1251] Step 4:

[1252] The server extracts essential features from the improved EEG data. Specifically, it uses machine learning algorithms (e.g., PCA, ICA) to extract features from the EEG data. The input is the improved EEG data, and the output is feature data.

[1253] Step 5:

[1254] The server uses the extracted features to recognize the user's intention and emotion. It uses a machine learning model (e.g., TensorFlow, PyTorch) to recognize intents such as "I need help" and "I need a break" from the features, and an emotion engine to recognize emotions such as "fatigue" and "stress." The input is feature data, and the output is recognized intent and emotion data.

[1255] Step 6:

[1256] The server generates a natural-sounding output message using a generative AI model (e.g., GPT-4) based on the recognized intent and emotion. For example, a message such as "A worker needs help and appears tired. Please provide assistance immediately" is generated. The input is the recognized intent and emotion data, and the output is the generated output message.

[1257] Step 7:

[1258] The server sends the generated output message to the terminal. The terminal provides visual or audio feedback of the received message to the user. Visual feedback uses a display, and audio feedback uses speech synthesis technology. The input is the generated output message, and the output is the feedback to the user.

[1259] Step 8:

[1260] The terminal adjusts the machine's operation based on the generated output message. For example, it issues instructions to adjust the work pace or to ask other workers for help. The input is the generated output message, and the output is an instruction to control the machine's operation.

[1261] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1262] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1263] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1264] [Fourth embodiment]

[1265] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1266] 7, a 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.

[1267] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1268] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1269] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1270] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1271] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1272] The control object 443 includes a display device, LEDs in the eyes, and motors for driving 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 emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1273] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1274] The specific processing program 56 is an example of a "program" according to the technology of the present 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.

[1275] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1276] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1277] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1278] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intention, and generates a natural output message based on that intention and provides feedback to the user. This system is composed of a user, a terminal, and a server, and functions as follows.

[1279] System components and operation overview

[1280] User

[1281] The user wears an EEG sensor and can use it in a natural environment. The EEG generated by the user's thoughts is detected by the sensor in real time.

[1282] Terminal

[1283] The device receives real-time EEG data from an EEG sensor worn by the user. The received EEG data is pre-processed at regular intervals to remove noise.

[1284] server

[1285] The server performs the following steps:

[1286] Data preprocessing: Upon receiving the EEG data sent from the device, the server performs noise removal and filtering to improve the quality of the data.

[1287] Feature extraction: Extract essential features from the preprocessed data. These features are used for subsequent intent recognition.

[1288] Intention recognition: A machine learning model is used to analyze the extracted features and recognize the user's intention. The recognized intentions are basic expressions such as "yes," "no," "I want to go to the toilet," and "I want to change my position."

[1289] Message generation: The generative AI generates natural output messages based on the recognized intent. For example, in response to the intent "I want to change my position," the system generates the message "I'm not feeling well, so please change your position."

[1290] Message transmission: The generated output message is sent from the server to the terminal.

[1291] feedback

[1292] The device may display the received message to the user or provide audio feedback, for example by visually displaying the message on a display or by using speech synthesis to audibly convey the message.

[1293] Specific examples

[1294] Example 1: If the user thinks "Yes"

[1295] 1. Brainwave data collection: When the user thinks "yes," the brainwave sensor detects this and sends the data to the terminal.

[1296] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[1297] 3. Feature extraction: The server extracts features from the preprocessed data.

[1298] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "yes" from the features.

[1299] 5. Message generation: The generative AI generates natural messages such as "That's right."

[1300] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[1301] Example 2: When the user thinks "I want to go to the toilet"

[1302] 1. Brainwave data collection: When the user thinks, "I want to go to the toilet," the brainwave sensor detects this and sends the data to the terminal.

[1303] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[1304] 3. Feature extraction: The server extracts features from the preprocessed data.

[1305] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "I want to go to the toilet" from the features.

[1306] 5. Message generation: The generative AI generates a natural message such as, "Sorry, but I need to go to the bathroom."

[1307] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[1308] In this way, the system enables users to generate more natural and specific messages based on EEG data, rather than simple formulaic phrases, allowing for richer and more natural communication.

[1309] The processing flow will be explained below.

[1310] Step 1:

[1311] The user wears the EEG sensor and begins using it in a natural environment. The EEG sensor detects the user's brain waves in real time and converts them into digital signals.

[1312] Step 2:

[1313] The device receives digital signals from the EEG sensor. The received EEG data is stored in a buffer and passed on to the next process at regular intervals.

[1314] Step 3:

[1315] The device performs preprocessing on the received EEG data, which includes noise reduction using a bandpass filter. This filter emphasizes only specific frequency bands and removes unnecessary noise.

[1316] Step 4:

[1317] The device converts the preprocessed EEG data into features, for example, by extracting features in the time and frequency domains, thereby extracting essential information from the data.

[1318] Step 5:

[1319] The device sends the feature data to the server, which is used as important data for recognizing the user's intention.

[1320] Step 6:

[1321] The server receives the feature data and inputs it into a machine learning model. This model is pre-trained to recognize various intents (e.g., "yes," "no," "I want to go to the toilet") and recognizes the most appropriate intent from the input data.

[1322] Step 7:

[1323] The server uses generative AI to generate natural-sounding output messages based on the recognized intent. For example, for the intent "I want to go to the toilet," it generates the message "Sorry, but I'm starting to feel like I need to go to the toilet."

[1324] Step 8:

[1325] The server then sends the generated output message to the device, which also includes metadata such as the user ID and a timestamp.

[1326] Step 9:

[1327] The device then feeds back the received message to the user, for example by displaying the message visually on a display screen or by using voice synthesis technology to communicate the message aloud.

[1328] Step 10:

[1329] The user checks the message fed back from the terminal and decides on the next action, thereby realizing natural communication between the user and the external system.

[1330] Through the above steps, natural and intimate communication based on the user's brain wave data is realized.

[1331] Example 1

[1332] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1333] Conventional systems using EEG have had difficulty accurately recognizing the user's intentions and providing feedback in natural language. Furthermore, insufficient noise removal and data preprocessing can reduce the accuracy of intention recognition. This has led to the issue of preventing smooth communication between the user and the computer.

[1334] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1335] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for recognizing a user's intention using the recognized features, means for generating a natural-sounding output message based on the recognized intention using a generative AI model, and means for feeding back the generated output message to the user. This makes it possible to recognize the user's intention with high accuracy and feed back the user with a natural-sounding message.

[1336] "Electroencephalograms" are signals that refer to electrical activity that occurs in the brain when a user thinks or feels something.

[1337] "Electroencephalogram data" refers to data indicating the measurement results of electroencephalograms detected by an electroencephalogram sensor.

[1338] The "receiving means" is a hardware and software device for acquiring the electroencephalogram data sent from the electroencephalogram sensor.

[1339] The "pre-processing means" is a processing device that performs noise removal and filtering on the received electroencephalogram data to improve the quality of the data.

[1340] The "means for extracting features" refers to means for extracting essential features from preprocessed electroencephalogram data and using them for subsequent analysis.

[1341] The "means for recognizing the user's intention" is a means for analyzing the extracted features and determining what the user is thinking and what intentions they have.

[1342] A "generative AI model" is an artificial intelligence model that generates natural-looking language expressions based on specific input.

[1343] An "output message" is a natural language expression created by a generative AI model based on the recognized user intent.

[1344] The "feedback means" refers to a means for visually or audibly conveying the generated output message to the user.

[1345] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intention, generates a natural output message based on the intention, and provides feedback to the user. This system is composed of a user, a terminal, and a server.

[1346] User

[1347] The user wears an EEG sensor, which can be, for example, a general EEG sensor device (e.g., an electroencephalogram (EEG) measurement device). When the user thinks about an intention, such as "yes" or "I want to go to the toilet," EEG signals are generated based on that intention, and the EEG sensor detects the data in real time.

[1348] Terminal

[1349] The device receives EEG data from an EEG sensor worn by the user. For example, a small computer (e.g., a small, high-performance processing unit) is used. The device performs noise removal and preprocessing on the received EEG data. Preprocessing includes removing high-frequency noise and filtering. This uses scientific computing libraries (e.g., numerical calculation libraries and signal processing libraries). After preprocessing, the EEG data is sent to the server at regular intervals.

[1350] server

[1351] The server performs the following steps:

[1352] 1. Data preprocessing: The server receives the preprocessed EEG data sent from the device and performs further advanced noise removal and filtering, thereby improving the quality of the data.

[1353] 2. Feature extraction: Extract essential features from the preprocessed data. This feature extraction uses a library for large-scale data analysis (e.g., a machine learning library).

[1354] 3. Intention Recognition: The server uses a machine learning model (e.g., a machine learning algorithm) to recognize the user's intention from the extracted features. The recognized intentions are basic expressions such as "yes," "no," "I want to go to the toilet," and "I want to change my position."

[1355] 4. Message generation: The server uses a generative AI model (e.g., an artificial intelligence language generation model) to generate a natural-sounding output message based on the recognized intent. For example, for the intent "I want to change my position," the server generates the message "I'm not feeling well, so please change my position."

[1356] 5. Message transmission: The generated output message is sent from the server to the terminal.

[1357] feedback

[1358] The device then provides feedback to the user about the received message. Feedback can be provided using a visual display or a speech synthesis engine (e.g., a text-to-speech system). For example, the generated message can be displayed on an LCD display or spoken to the user.

[1359] Specific examples

[1360] Example 1: If the user thinks "Yes"

[1361] 1. Brainwave data collection: When the user thinks "yes," the brainwave sensor detects this and sends the data to the terminal.

[1362] 2. Preprocessing: Noise removal is performed on the device and the data is sent to the server.

[1363] 3. Feature extraction: The server extracts features from the preprocessed data.

[1364] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "yes" from the features.

[1365] 5. Message generation: The generative AI generates natural messages such as "That's right."

[1366] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[1367] Example prompt sentence:

[1368] Your users are thinking "yes." Express that intent in natural language.

[1369] Example 2: When the user thinks "I want to go to the toilet"

[1370] 1. Brainwave data collection: When the user thinks, "I want to go to the toilet," the brainwave sensor detects this and sends the data to the terminal.

[1371] 2. Preprocessing: Noise removal is performed on the device and the data is sent to the server.

[1372] 3. Feature extraction: The server extracts features from the preprocessed data.

[1373] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "I want to go to the toilet" from the features.

[1374] 5. Message generation: The generative AI generates a natural message such as, "Sorry, but I need to go to the bathroom."

[1375] 6. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[1376] Example prompt sentence:

[1377] The user wants to "go to the bathroom." Express that intent in natural language.

[1378] In this way, the system enables users to generate more natural and specific messages based on EEG data, rather than simple formulaic phrases, allowing for more in-depth and natural communication.

[1379] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1380] Step 1:

[1381] The user wears an EEG sensor. When the user thinks about an intention, such as "yes" or "I want to go to the toilet," brain waves based on that intention are generated. The EEG sensor detects this brain wave data in real time and transmits it to the device. The input is the user's brain waves, and the output is brain wave data.

[1382] Step 2:

[1383] The device receives EEG data sent from the EEG sensor. It performs noise removal and preprocessing on the received data. For preprocessing, the numpy and scipy libraries are used to remove high-frequency noise and improve data quality. This process yields EEG data with noise removed. The input is raw data from the EEG sensor, and the output is preprocessed EEG data.

[1384] Step 3:

[1385] The device sends preprocessed EEG data to the server at regular intervals. It transfers data to the server using HTTP requests. The input is the preprocessed EEG data, and the output is a data packet sent to the server.

[1386] Step 4:

[1387] The server receives the preprocessed EEG data sent from the device. It then performs advanced noise removal and smoothing on the received data, further improving the accuracy of the data. The input is the preprocessed EEG data sent from the device, and the output is highly preprocessed EEG data.

[1388] Step 5:

[1389] The server extracts features from the highly preprocessed EEG data. This feature extraction uses machine learning libraries such as TensorFlow and PyTorch. The extracted features are then used for intent recognition. The input is the highly preprocessed EEG data, and the output is the extracted features.

[1390] Step 6:

[1391] The server recognizes the user's intent based on the features. It uses a machine learning model (e.g., SVM or neural network) to analyze the features and determine the user's intent, such as "Yes" or "I want to go to the toilet." The input is the extracted features, and the output is the recognized user intent.

[1392] Step 7:

[1393] The server uses a generative AI model (e.g., OpenAI's generative model) to generate a natural-sounding output message based on the recognized user intent. For example, for the intent "I want to go to the toilet," the message "Sorry, but I'm starting to feel like I need to go to the toilet" is generated. The input is the recognized user intent, and the output is the generated output message.

[1394] Step 8:

[1395] The server sends the generated message to the terminal. It transfers the message to the terminal using an HTTP request. The input is the generated output message, and the output is the data packet sent to the terminal.

[1396] Step 9:

[1397] The terminal feeds back the received message to the user. The terminal displays the message on a display or speaks it aloud using a speech synthesis engine. For example, the terminal may display a message on an LCD display or speak "Sorry, but I need to go to the toilet." The input is the output message sent from the server, and the output is visual or spoken feedback to the user.

[1398] (Application example 1)

[1399] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1400] Conventional factory robots require manual input when operating, which reduces work efficiency. Furthermore, errors or delays that occur when workers operate robots can impair the productivity of the entire factory. To solve this problem, a more intuitive and faster way to operate robots is needed.

[1401] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1402] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for analyzing a user's intention and generating instructions for operating a machine based on the intention, and means for transmitting the generated operation instructions to the machine, thereby enabling a user to intuitively and quickly operate a robot using only their electroencephalograms.

[1403] 1. "EEG" refers to electrical signals generated by a user's brain activity.

[1404] 2. An "EEG sensor" is a device that is worn on the user's scalp to detect brain waves.

[1405] 3. "Preprocessing" refers to the process of removing noise from the received EEG data and improving the quality of the signal.

[1406] 4. "Features" are data that represent essential information extracted from preprocessed EEG data.

[1407] 5. "Intention recognition" is the process of using a machine learning model to identify the user's purpose and intention from extracted features.

[1408] 6. "Output message" is a naturally-expressed message generated based on the recognized user intent.

[1409] 7. "Feedback" is the means by which generated output messages are communicated to the user.

[1410] 8. "Instructions to operate a machine" are instructions to guide a machine, such as a factory robot, to perform a specific operation based on the user's intentions.

[1411] 9. "Real-time" means that EEG data is processed and analyzed in real time.

[1412] 10. A "machine learning model" is a model that uses statistical methods and algorithms to learn from data and automatically recognize patterns.

[1413] 11. "Server" refers to a central management system that performs processes such as preprocessing of EEG data, feature extraction, intention recognition, and message generation.

[1414] 12. A "factory robot" is a mechanical device used to automate work in a factory.

[1415] 13. "Generative AI" is an artificial intelligence system that generates natural-looking output messages based on the user's intent.

[1416] This invention is a system consisting of an EEG sensor worn by the user, a terminal that processes EEG data, and a server that performs data analysis and intention recognition. The user wears the EEG sensor, which detects the user's EEG in real time. This EEG data is transmitted to the terminal.

[1417] Processing by the terminal

[1418] The device receives the EEG data in real time and performs preprocessing such as noise removal and filtering, which ensures reliable data is sent to the server.

[1419] Server processing

[1420] The server receives the preprocessed EEG data sent from the device and extracts features, which are used in the subsequent intent recognition process.The server then analyzes the features using a machine learning model to recognize the user's intent.

[1421] Processing Recognized Intent

[1422] Based on the recognized intent, the generative AI model generates a natural-sounding output message. For example, if the intent is "I want to change my position," the model generates a natural-sounding message: "I'm not feeling well, so please change my position." This output message is then sent to the device.

[1423] User Feedback

[1424] The terminal feeds back the received output message to the user either visually on a display or audibly using speech synthesis.

[1425] Factory robot operation

[1426] Instructions generated based on the user's intentions are transmitted to the factory robot, enabling intuitive and rapid robot control through brain waves.

[1427] In an actual use case, if a user wants to move the robot to the right, the EEG sensor detects this and the pre-processed data is analyzed by the server. The server recognizes the intention of "move right" and generates an instruction based on that intention. Finally, the message "move right" is generated and sent to the robot.

[1428] An example prompt is:

[1429] The user wears an EEG sensor and consciously moves the robot to the right. EEG data from the sensor is collected in real time and pre-filtered to remove noise. The data obtained through feature extraction is analyzed by a machine learning model to recognize the user's intention to move the robot to the right. The message "Move right" is then generated and given as voice feedback.

[1430] In this way, the system utilizes the user's brainwaves to enable intuitive and efficient operation of factory robots.

[1431] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1432] Step 1:

[1433] The user wears the EEG sensor and starts collecting EEG data. The EEG sensor detects the user's EEG activity in real time. The input is the user's EEG signal, and the output is raw EEG data.

[1434] Step 2:

[1435] The terminal preprocesses the raw EEG data received in real time. This preprocessing includes noise removal and filtering. Specifically, it removes noise components using a high-frequency filter to improve signal quality. The input is raw EEG data, and the output is preprocessed EEG data.

[1436] Step 3:

[1437] The terminal sends the preprocessed EEG data to the server. The input is the preprocessed EEG data, and the output is the data sent to the server.

[1438] Step 4:

[1439] The server extracts features from the received preprocessed EEG data. Specifically, it analyzes the statistical characteristics of the signal and the frequency components using Fourier transform. The input is the preprocessed EEG data, and the output is feature data.

[1440] Step 5:

[1441] The server inputs the feature data into a machine learning model to recognize the user's intention. This process uses a model that has learned the correspondence between past EEG data and intention. The input is the feature data, and the output is the user's intention.

[1442] Step 6:

[1443] The server generates a natural-sounding output message using a generative AI model based on the recognized intent. For example, for the intent "Move the robot to the right," the server generates the message "Move to the right." The input is the user's intent, and the output is the output message.

[1444] Step 7:

[1445] The server generates and sends output messages to the terminal. The input is the output message, and the output is the message sent to the terminal.

[1446] Step 8:

[1447] The device then feeds back the received output message to the user. Specifically, the device displays the message on a display or transmits it audibly using speech synthesis. The input is the output message, and the output is visual or audio feedback.

[1448] Step 9:

[1449] Operation instructions generated based on the user's intentions are sent to the factory robot. The terminal or server sends specific operation instructions to the robot. The input is the operation instruction, and the output is the robot's operation.

[1450] This series of steps enables users to intuitively and quickly operate factory robots through their brain waves.

[1451] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1452] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intentions and emotions, generates natural output messages based on the intentions and emotions, and provides feedback to the user. This system is composed of a user, a terminal, and a server, and functions as follows.

[1453] System components and operation overview

[1454] User

[1455] The user wears an EEG sensor and can use it in a natural environment. The EEG generated by the user's thoughts is detected by the sensor in real time.

[1456] Terminal

[1457] The device receives real-time EEG data from an EEG sensor worn by the user. The received EEG data is pre-processed at regular intervals to remove noise.

[1458] server

[1459] The server performs the following steps:

[1460] Data preprocessing: Upon receiving the EEG data sent from the device, the server performs noise removal and filtering to improve the quality of the data.

[1461] Feature extraction: Extract essential features from the preprocessed data. These features are used for intent and emotion recognition.

[1462] Intention recognition: A machine learning model is used to analyze the extracted features and recognize the user's intention. The recognized intentions are basic expressions such as "yes," "no," "I want to go to the toilet," and "I want to change my position."

[1463] Emotion Recognition: The emotion engine is used to recognize the user's emotions from the extracted features, such as pleasant / unpleasant, joy, sadness, and anger.

[1464] Message generation: The generative AI generates natural-sounding output messages based on the recognized intent and emotion. For example, if the intent is "I want to go to the toilet" and the emotion is "impatience," the message generated is "Sorry, but I'm in a hurry to go to the toilet."

[1465] Message transmission: The generated output message is sent from the server to the terminal.

[1466] feedback

[1467] The device may display the received message to the user or provide audio feedback, for example by visually displaying the message on a display or by using speech synthesis to audibly convey the message.

[1468] Specific examples

[1469] Example 1: If the user thinks "Yes"

[1470] 1. Brainwave data collection: When the user thinks "yes," the brainwave sensor detects this and sends the data to the terminal.

[1471] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[1472] 3. Feature extraction: The server extracts features from the preprocessed data.

[1473] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "yes" from the features.

[1474] 5. Emotion Recognition: The server uses an emotion engine to recognize the user's emotions, such as "relief."

[1475] 6. Message generation: The generative AI generates natural messages such as "That's right, don't worry."

[1476] 7. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[1477] Example 2: When the user thinks "I want to go to the toilet"

[1478] 1. Brainwave data collection: When the user thinks, "I want to go to the toilet," the brainwave sensor detects this and sends the data to the terminal.

[1479] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[1480] 3. Feature extraction: The server extracts features from the preprocessed data.

[1481] 4. Intent recognition: The server uses a machine learning model to recognize the intent of "I want to go to the toilet" from the features.

[1482] 5. Emotion Recognition: The server uses an emotion engine to recognize the user's emotions, such as "impatience."

[1483] 6. Message generation: The generative AI generates a natural message such as, "Sorry, but I need to go to the bathroom quickly."

[1484] 7. Feedback: The terminal displays or audibly feeds back the generated message to the user.

[1485] In this way, the system enables users to communicate more effectively and naturally by generating more natural and specific messages based on EEG data and emotions, rather than simple formulaic phrases.

[1486] The processing flow will be explained below.

[1487] Step 1:

[1488] The user wears the EEG sensor and begins using it in a natural environment. The EEG sensor detects the user's brain waves in real time and converts them into digital signals.

[1489] Step 2:

[1490] The device receives digital signals from the EEG sensor. The received EEG data is stored in a buffer and passed on to the next process at regular intervals.

[1491] Step 3:

[1492] The device performs preprocessing on the received EEG data, which includes noise reduction using a bandpass filter. This filter emphasizes only specific frequency bands and removes unnecessary noise.

[1493] Step 4:

[1494] The device converts the preprocessed EEG data into features. Feature extraction uses techniques such as FFT (Fast Fourier Transform) to extract important features in the time and frequency domains. These features are used to recognize intentions and emotions.

[1495] Step 5:

[1496] The device transmits the feature data to the server, which is used as important data for recognizing the user's intentions and emotions.

[1497] Step 6:

[1498] The server receives the feature data and inputs it into a machine learning model. This model is pre-trained to recognize various intents (e.g., "yes," "no," "I want to go to the toilet") and recognizes the most appropriate intent from the input data.

[1499] Step 7:

[1500] At the same time, the server uses an emotion engine to recognize the user's emotions from the feature data. The emotion engine determines pleasant / unpleasant feelings, such as joy, sadness, and anger, in real time.

[1501] Step 8:

[1502] The server uses generative AI to generate natural output messages based on the recognized intent and emotion. For example, if the intent is "I want to go to the toilet" and the emotion is "impatience," the server generates a specific message such as "I'm sorry, but I feel like I need to go to the toilet quickly."

[1503] Step 9:

[1504] The server then sends the generated output message to the device, optionally including metadata such as a timestamp and user ID.

[1505] Step 10:

[1506] The device may then visually display the received message to the user or provide audio feedback, for example by displaying the message on a display screen or by using speech synthesis to communicate the message aloud.

[1507] Step 11:

[1508] The user checks the feedback message from the device and decides on the next action, thereby realizing natural and intimate communication between the user and the external system.

[1509] Example 2

[1510] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1511] Conventional intention recognition systems using EEG have difficulty accurately grasping a user's intention. There is a need for a system that can simultaneously recognize not only a user's intention but also their emotions, and generate more natural and appropriate output messages.

[1512] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1513] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for recognizing a user's intention using the extracted features, means for recognizing the user's emotion using the extracted features, means for generating a natural output message based on the recognized intention and emotion, and means for feeding back the generated output message to the user, thereby enabling the generation of a more natural and appropriate output message based on the user's intention and emotion.

[1514] "Electroencephalograms" are weak electrical potential changes resulting from electrical activity in the brain.

[1515] "Preprocessing" refers to processes such as data shaping, noise removal, and filtering that are carried out before data analysis.

[1516] A "feature" is significant information extracted from data for use in analysis by a machine learning model.

[1517] "Intention" refers to the specific intention or purpose of what the user wants or thinks.

[1518] "Emotion" refers to the sensations and psychological states that a user is experiencing.

[1519] An "output message" is a reply or feedback message to the user that is generated based on the recognized intent and emotion.

[1520] "Feedback" refers to the provision of information or responses from the system to the user.

[1521] A "machine learning model" is an algorithm that learns from data and uses the learning results to make predictions and classify new data.

[1522] A "generative AI model" is an artificial intelligence algorithm that generates natural-looking sentences based on the user's intentions and emotions, a type of generative artificial intelligence.

[1523] "Filtering" is the process of removing unwanted components from a data signal.

[1524] The present invention is a system that receives brain waves as input and recognizes a user's intentions and emotions. To implement this system, a user, a terminal, and a server work together. Each step uses specific hardware and software.

[1525] User

[1526] The user wears an EEG sensor and expresses their intentions and emotions naturally. This EEG sensor detects the electrical activity of the brain in real time and transmits the data to a terminal. A commercially available electroencephalograph can be used as the EEG sensor.

[1527] Terminal

[1528] The device receives EEG data in real time from an EEG sensor worn by the user. The received EEG data undergoes preprocessing such as noise removal, and the clean data is sent to the server. Specific software that can be used for the device's preprocessing includes MATLAB and Python libraries (NumPy, SciPy).

[1529] server

[1530] The server receives the preprocessed EEG data sent from the device and performs the following data processing and calculations:

[1531] 1. Feature extraction:

[1532] The server extracts features from the preprocessed data, specifically, calculates the power spectrum for each frequency band using a Fourier transform, and analyzes the time-domain data using a machine learning model.

[1533] 2. Intention Recognition:

[1534] The user's intent is recognized from the extracted features using a machine learning model (e.g., random forest, support vector machine (SVM), deep learning, etc.). Examples of intent include "yes," "no," "I want to go to the toilet," and "I want to change my position."

[1535] 3. Emotion recognition:

[1536] The emotion engine is used to recognize the user's emotions from the extracted features. Examples of emotions include pleasant / unpleasant, joy, sadness, and anger. The emotion engine uses Microsoft Azure Emotion API and IBM Watson Emotion Recognition.

[1537] 4. Message Creation:

[1538] A generative AI model (e.g., OpenAI's GPT-3 or Google's BERT) is used to generate natural-sounding output messages based on the recognized intent and emotion. By inputting a prompt sentence into the generative AI model, an appropriate output message is generated.

[1539] 5. Sending a message:

[1540] The generated output message is sent from the server to the terminal.

[1541] feedback

[1542] The device displays or provides audio feedback to the user based on the output messages received from the server. Visual feedback is displayed on the display, and audio feedback is provided using a speech synthesis engine (e.g., Google Text-to-Speech API or Amazon Polly).

[1543] Specific examples

[1544] Example 1: If the user thinks "Yes"

[1545] 1. The user thinks "yes."

[1546] 2. The EEG sensor detects brain waves and transmits them to the device.

[1547] 3. The device performs noise removal and sends the data to the server.

[1548] 4. The server extracts features from the preprocessed data.

[1549] 5. The machine learning model recognizes the intent "yes."

[1550] 6. The emotion engine recognizes the emotion of "relief."

[1551] 7. The generative AI model generates a message such as "That's right, don't worry."

[1552] 8. The terminal displays or audibly announces the generated message.

[1553] Prompt Sentence Examples

[1554] If the user thinks "yes": "Receive EEG data indicating that the user is thinking 'yes', recognize that intention, and generate a reassuring message."

[1555] When the user thinks, "I want to go to the toilet": "Receive EEG data when the user is thinking, 'I want to go to the toilet,' recognize that intention, and generate a message that includes a sense of urgency."

[1556] As a result, the present invention enables the generation of more natural and appropriate output messages based on the user's electroencephalogram data and emotions, providing a more effective means of communication.

[1557] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1558] Step 1:

[1559] The user wears an EEG sensor

[1560] The user wears an EEG sensor, which detects the user's brain waves in real time and converts them into digital signals, collecting brain wave data generated when the user thinks about something.

[1561] Input: User's brainwaves

[1562] Output: EEG data converted into digital signals

[1563] Specific behavior:

[1564] The user thinks, "I want to go to the toilet."

[1565] The EEG sensor detects brain waves and converts them into digital signals.

[1566] The converted digital signal is transmitted to the terminal.

[1567] Step 2:

[1568] The terminal receives and preprocesses the EEG data.

[1569] The device receives the digital signal sent from the EEG sensor and performs preprocessing such as filtering and noise removal on the received data to create clean data.

[1570] Input: Converted digital signal

[1571] Output: Preprocessed and clean EEG data

[1572] Specific behavior:

[1573] The device begins receiving brainwave data.

[1574] A band pass filter is applied to remove low and high frequency noise.

[1575] The preprocessed data is sent to the server.

[1576] Step 3:

[1577] The server extracts features

[1578] The server receives the preprocessed data sent from the device, calculates the power spectrum for each frequency band using Frue, and extracts important features from the received data.

[1579] Input: Preprocessed EEG data

[1580] Output: Extracted features

[1581] Specific behavior:

[1582] The server performs a Fourier transform on the pre-processed data.

[1583] Calculate the power spectral density of each frequency band.

[1584] Specific features are extracted and passed to the next processing step.

[1585] Step 4:

[1586] The server recognizes the user's intent

[1587] The server inputs the extracted features into a machine learning model to recognize the user's intent. The recognized intent indicates the meaning of thoughts and actions, such as "yes," "no," or "I want to go to the toilet."

[1588] Input: extracted features

[1589] Output: Recognized user intent

[1590] Specific behavior:

[1591] Input the features into a machine learning model (e.g., random forest or deep learning model).

[1592] The model performs analysis and recognizes the intention of "I want to go to the toilet."

[1593] The recognized intent is passed to the next processing step.

[1594] Step 5:

[1595] The server recognizes the user's emotions

[1596] The server uses the same features to input them into an emotion engine to recognize the user's emotions. The recognized emotions indicate the user's psychological state. Examples include "pleasant / unpleasant," "joy," "sadness," and "anger."

[1597] Input: extracted features, recognized intent

[1598] Output: Recognized user emotion

[1599] Specific behavior:

[1600] Input the features into a sentiment analysis model (e.g., Microsoft Azure Emotion API).

[1601] The model performs analysis and recognizes the emotion of "impatience."

[1602] The recognized emotion is passed on to the next processing step.

[1603] Step 6:

[1604] The server generates an output message

[1605] The server utilizes a generative AI model based on the recognized user's intention and emotion to generate natural-sounding output messages that reflect the user's intention and appropriately convey their emotions.

[1606] Input: Perceived Intent, Perceived Emotion

[1607] Output: The generated output message

[1608] Specific behavior:

[1609] The recognized intent "I want to go to the toilet" and emotion "impatience" are input as prompts into the generative AI model.

[1610] The generative AI model generates the message, "Sorry, but I need to go to the bathroom quickly."

[1611] Send the generated message to the terminal.

[1612] Step 7:

[1613] The device provides feedback to the user

[1614] The terminal provides the output messages received from the server to the user, visually on a display and using a speech synthesis engine for voice feedback.

[1615] Input: The generated output message

[1616] Output: Feedback provided to the user

[1617] Specific behavior:

[1618] The terminal receives the generated message.

[1619] The display will say, "Sorry, but I need to go to the toilet quickly."

[1620] Or use a text-to-speech engine to play the message aloud.

[1621] As described above, each processing step functions in cooperation with one another, and the present invention realizes the generation of a more natural and appropriate output message based on the user's electroencephalogram data and emotions.

[1622] (Application example 2)

[1623] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1624] In manufacturing sites such as factories, it is difficult to understand the intentions and emotions of workers in real time and respond appropriately accordingly. Rapid response is particularly important when workers become fatigued, stressed, or need help, but current systems cannot adequately meet these requirements. Solving this issue directly leads to improved safety and efficiency in the work environment.

[1625] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1626] In this invention, the server includes means for receiving electroencephalograms as input, means for preprocessing the received electroencephalogram data, means for extracting features from the preprocessed data, means for recognizing a user's intention and emotion using the extracted features, means for generating a natural-sounding output message based on the recognized intention and emotion, means for feeding back the generated output message to the user, and means for adjusting the operation of the machine based on the generated output message. This makes it possible to understand the intention and emotion of the worker in real time and respond quickly and appropriately.

[1627] "Electroencephalograms" are electrical signals emitted by the human brain that reflect states such as thoughts and emotions.

[1628] "Means for receiving" refers to a device or process for receiving signals or data input from the outside.

[1629] A "preprocessing means" is a device or process that removes noise or filters the received data to make it easier to analyze.

[1630] A "means for extracting features" is a device or process that extracts essential information from raw data for use in subsequent analysis.

[1631] The "means for recognizing intentions" is a device or process for identifying the user's actions and wishes from the extracted features.

[1632] The "means for recognizing emotions" is a device or process for identifying the emotional state of the user from the extracted features.

[1633] A "means for generating an output message" is a device or process that creates an appropriate message for the user based on the perceived intent or emotion.

[1634] A "means for providing feedback to the user" is a device or process that visually or audibly conveys the generated output message to the user.

[1635] A "means for regulating the operation of a machine" is a device or process for controlling or modifying the operation of a machine based on the generated output messages.

[1636] The present invention is a system that receives electroencephalograms as input, processes the received electroencephalogram data to recognize the user's intentions and emotions, generates natural output messages based on the intentions and emotions, and then adjusts the operation of a machine based on the generated messages. This system is mainly composed of a user, a terminal, and a server.

[1637] User

[1638] The user wears an EEG sensor and can use it in a natural environment. The EEG generated by the user's thoughts is detected by the sensor in real time.

[1639] Terminal

[1640] The device receives real-time EEG data from an EEG sensor worn by the user. The received EEG data is pre-processed at regular intervals to remove noise.

[1641] server

[1642] The server performs the following steps:

[1643] 1. Data preprocessing: Upon receiving the EEG data sent from the device, the server performs preprocessing to improve the quality of the data by removing noise and filtering.

[1644] 2. Feature extraction: Extract essential features from the preprocessed data. These features are used for intent and emotion recognition.

[1645] 3. Intent and emotion recognition: The server uses a machine learning model (e.g., TensorFlow, PyTorch) and an emotion engine to analyze the extracted features and recognize the user's intent and emotion. Recognized intents include "I need help" and "I need a break," and emotions include "fatigue" and "stress."

[1646] 4. Message generation: Generative AI (e.g., GPT-4) generates natural-sounding output messages based on the perceived intent and sentiment, such as "The worker is tired, so slow down."

[1647] 5. Feedback: The server sends the generated output message to the terminal, which then feeds it back to the user and any associated machines, for example visually displaying it on a display or audibly using speech synthesis to communicate the message.

[1648] 6. Adjusting machine operation: Based on the generated output message, the terminal controls or changes the machine's operation, for example, slowing down the work pace or asking other workers for help.

[1649] Specific examples

[1650] Example 1: When a worker thinks they need help

[1651] 1. Brainwave data collection: When a worker thinks, "I need help," the brainwave sensor detects this and sends the data to the terminal.

[1652] 2. Preprocessing: Perform noise removal on the device and send the data to the server.

[1653] 3. Feature extraction: The server extracts features from the preprocessed data.

[1654] 4. Intent and emotion recognition: The server uses a machine learning model to recognize the intent of "need help" from the features, and an emotion engine to recognize the emotion of "tired."

[1655] 5. Message generation: The generation AI generates a natural message such as, "The worker needs help and seems tired. Please provide assistance immediately."

[1656] 6. Feedback: The terminal provides visual or audio feedback of the generated message and adjusts the machine's operation.

[1657] Example prompt sentence:

[1658] The intent of "I need help" and the emotion of "fatigue" have been recognized. Please respond.

[1659] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1660] Step 1:

[1661] The user wears an EEG sensor and works in a natural environment. The EEG sensor detects the brain waves generated by the user's thoughts in real time. The input is the user's brain waves, and the output is the brain wave data sent from the EEG sensor to the terminal.

[1662] Step 2:

[1663] The device acquires EEG data received from the EEG sensor in real time. The acquired EEG data is preprocessed at regular intervals to remove noise. Specifically, noise filtering and noise removal algorithms (e.g., bandpass filters) are applied. The input is raw EEG data, and the output is preprocessed EEG data.

[1664] Step 3:

[1665] The device sends the preprocessed EEG data to a server, which receives the preprocessed EEG data and performs additional noise removal and filtering to improve the quality of the data. The input is the preprocessed EEG data, and the output is the quality-improved EEG data.

[1666] Step 4:

[1667] The server extracts essential features from the improved EEG data. Specifically, it uses machine learning algorithms (e.g., PCA, ICA) to extract features from the EEG data. The input is the improved EEG data, and the output is feature data.

[1668] Step 5:

[1669] The server uses the extracted features to recognize the user's intention and emotion. It uses a machine learning model (e.g., TensorFlow, PyTorch) to recognize intents such as "I need help" and "I need a break" from the features, and an emotion engine to recognize emotions such as "fatigue" and "stress." The input is feature data, and the output is recognized intent and emotion data.

[1670] Step 6:

[1671] The server generates a natural-sounding output message using a generative AI model (e.g., GPT-4) based on the recognized intent and emotion. For example, a message such as "A worker needs help and appears tired. Please provide assistance immediately" is generated. The input is the recognized intent and emotion data, and the output is the generated output message.

[1672] Step 7:

[1673] The server sends the generated output message to the terminal. The terminal provides visual or audio feedback of the received message to the user. Visual feedback uses a display, and audio feedback uses speech synthesis technology. The input is the generated output message, and the output is the feedback to the user.

[1674] Step 8:

[1675] The terminal adjusts the machine's operation based on the generated output message. For example, it issues instructions to adjust the work pace or to ask other workers for help. The input is the generated output message, and the output is an instruction to control the machine's operation.

[1676] 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 control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1677] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1678] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1679] The emotion identification model 59 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 an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1680] FIG. 9 illustrates an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and behaviors arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1681] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1682] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1683] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1684] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs 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 a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1685] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1686] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1687] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1688] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1689] 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.

[1690] It is not necessary to store all 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 all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1691] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1692] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1693] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1694] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1695] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1696] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1697] The following is further disclosed regarding the above embodiment.

[1698] (Claim 1)

[1699] means for receiving brain waves as input;

[1700] means for pre-processing the received electroencephalogram data;

[1701] means for extracting features from the preprocessed data;

[1702] a means for recognizing a user's intention using the extracted features;

[1703] means for generating a natural-sounding output message based on the recognized intent;

[1704] means for feeding back the generated output message to a user;

[1705] A system including:

[1706] (Claim 2)

[1707] 2. The system according to claim 1, wherein the preprocessing means performs filtering and noise removal on the electroencephalogram data.

[1708] (Claim 3)

[1709] 2. The system according to claim 1, wherein the intention recognition means recognizes the user's intention using a machine learning model.

[1710] "Example 1"

[1711] (Claim 1)

[1712] means for receiving brain waves as input;

[1713] means for pre-processing the received electroencephalogram data;

[1714] means for extracting features from the preprocessed data;

[1715] a means for recognizing a user's intention using the extracted features;

[1716] means for generating a natural-sounding output message based on the recognized intent;

[1717] means for feeding back the generated output message to a user;

[1718] A system including:

[1719] (Claim 2)

[1720] 2. The system according to claim 1, wherein the preprocessing means performs filtering and noise removal on the electroencephalogram data.

[1721] (Claim 3)

[1722] 2. The system according to claim 1, wherein the intention recognition means recognizes the user's intention using a machine learning model.

[1723] (Claim 4)

[1724] 10. The system of claim 1, wherein a generative AI model is used to generate natural-sounding output messages based on the recognized intent.

[1725] (Claim 5)

[1726] 3. The system of claim 2, wherein the pre-processing means includes means for processing the received electroencephalogram data in real time and transmitting the data to a server.

[1727] "Application Example 1"

[1728] (Claim 1)

[1729] means for receiving brain waves as input;

[1730] means for pre-processing the received electroencephalogram data;

[1731] means for extracting features from the preprocessed data;

[1732] a means for recognizing a user's intention using the extracted features;

[1733] means for generating a natural-sounding output message based on the recognized intent;

[1734] means for feeding back the generated output message to a user;

[1735] means for analyzing a user's intention and generating instructions for operating a machine based on the intention;

[1736] means for transmitting the generated operation instructions to a machine;

[1737] A system including:

[1738] (Claim 2)

[1739] 2. The system according to claim 1, wherein the preprocessing means performs filtering and noise removal on the electroencephalogram data.

[1740] (Claim 3)

[1741] 2. The system according to claim 1, wherein the intention recognition means recognizes the user's intention using a machine learning model.

[1742] "Example 2: Combining Emotion Engines"

[1743] (Claim 1)

[1744] means for receiving brain waves as input;

[1745] means for pre-processing the received electroencephalogram data;

[1746] means for extracting features from the preprocessed data;

[1747] a means for recognizing a user's intention using the extracted features;

[1748] a means for recognizing a user's emotion using the extracted features;

[1749] means for generating natural-sounding output messages based on the recognized intent and emotion;

[1750] means for feeding back the generated output message to a user;

[1751] A system including:

[1752] (Claim 2)

[1753] 2. The system according to claim 1, wherein the preprocessing means performs filtering and noise removal on the electroencephalogram data.

[1754] (Claim 3)

[1755] 2. The system according to claim 1, wherein the intention recognition means recognizes the user's intention using a machine learning model.

[1756] (Claim 4)

[1757] 2. The system according to claim 1, wherein the emotion recognition means recognizes the user's emotions using a machine learning model.

[1758] (Claim 5)

[1759] 2. The system of claim 1, wherein the output message generating means uses a generative AI model to generate a message based on the user's intentions and emotions.

[1760] "Application example 2 when combining emotion engines"

[1761] (Claim 1)

[1762] means for receiving brain waves as input;

[1763] means for pre-processing the received electroencephalogram data;

[1764] means for extracting features from the preprocessed data;

[1765] a means for recognizing a user's intention and emotion using the extracted features;

[1766] means for generating natural-sounding output messages based on the recognized intent and emotion;

[1767] means for feeding back the generated output message to a user;

[1768] means for adjusting operation of the machine based on the generated output messages;

[1769] A system including:

[1770] (Claim 2)

[1771] 2. The system according to claim 1, wherein the preprocessing means performs filtering and noise removal on the electroencephalogram data.

[1772] (Claim 3)

[1773] 2. The system according to claim 1, wherein the intention and emotion recognition means recognizes the user's intention and emotion using a machine learning model. [Explanation of symbols]

[1774] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. means for receiving brain waves as input; means for pre-processing the received electroencephalogram data; means for extracting features from the preprocessed data; a means for recognizing a user's intention using the extracted features; means for generating a natural-sounding output message based on the recognized intent; means for feeding back the generated output message to a user; A system including:

2. 2. The system according to claim 1, wherein the pre-processing means performs filtering and noise removal on the electroencephalogram data.

3. 2. The system according to claim 1, wherein the intention recognition means recognizes the user's intention using a machine learning model.

Images