System

Abstract

A system is provided.SOLUTION: A system includes, in a vehicle that cooperates with an automatic driving system that performs driving operations, means for collecting facial expression data and voice data of a user in real time by a camera installed inside, means for collecting voice data of the user in real time by a microphone installed inside, means for preprocessing the collected facial expression data and voice data to extract feature points and feature amounts, means for estimating a mental state of the user from the extracted feature data by using a generative AI model, means for adjusting driving parameters based on the estimated mental state, and means for transmitting and applying the adjusted driving parameters to the automatic driving system.SELECTED DRAWING: Figure 1

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] While autonomous driving technology is advancing, there is a problem that the sense of security felt by passengers does not necessarily correspond to technical safety. In particular, excessive speeds and sudden acceleration can cause fear and anxiety in passengers, so an autonomous driving system that improves the psychological sense of security is needed. The object of the present invention is to provide an autonomous driving system that can provide passengers with a sense of security by analyzing the user's words and facial expressions, inferring their psychological state, and adjusting driving accordingly. [Means for solving the problem]

[0005] The present invention is a system for a vehicle that cooperates with an automated driving system that performs driving operations, and includes the following means: means for collecting a user's facial expressions in real time using an internally installed camera, means for collecting the user's voice in real time using an internally installed microphone, means for preprocessing the collected facial expression data and voice data to extract feature points and feature quantities, means for inferring the user's psychological state from the extracted feature data using a generative AI model, means for adjusting driving parameters based on the inferred psychological state, and means for transmitting and applying the adjusted driving parameters to the automated driving system. This makes it possible to reduce the fear and anxiety felt by passengers while driving and provide a sense of security.

[0006] An "autonomous driving system" is a system that uses computers and sensors to drive and control a vehicle without human operation.

[0007] "Vehicle" refers to a mechanical device for transporting passengers, and primarily refers to automobiles that travel on roads.

[0008] A "camera" is an optical device for capturing images or video, particularly a device installed inside a vehicle and used to collect facial expressions of a user.

[0009] "Facial expressions" are visual expressions formed by facial feature points and muscle movements to convey human emotions and psychological states.

[0010] "Audio" refers to sound waves transmitted by vibrations in the air, and specifically refers to the user's speech or voice.

[0011] A "microphone" is a device for converting sound waves in the air into electrical signals, and is used especially inside a vehicle to collect the user's voice.

[0012] "Preprocessing" refers to the initial processing performed to convert collected raw data into a format suitable for analysis and modeling, and specifically refers to the extraction of feature points and feature quantities.

[0013] "Feature points" are specific positions on the face (for example, eyes, mouth, eyebrows, etc.) and are measurement points used for facial expression analysis.

[0014] "Features" are quantitative values ​​(for example, pitch, volume, spectral characteristics, etc.) extracted from audio data, and are data used for audio analysis.

[0015] A "generative AI model" is an artificial intelligence model that uses machine learning and deep learning algorithms to generate output from input data.

[0016] "Mental state" refers to the user's emotional or mental state, and is the specific feeling or sensation inferred by the generative AI model.

[0017] "Driving parameters" are set values ​​(e.g., speed, acceleration, braking strength, etc.) used to control the operation of the automated driving system.

[0018] "Tuning" is the process of changing settings or operating parameters in response to specific circumstances or conditions.

[0019] "Transmitting" is the act of sending information or data from one device to another.

[0020] "Applying" refers to the act of reflecting new settings and parameters in the actual system and making them available for use. [Brief explanation of the drawings]

[0021] [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

[0022] 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.

[0023] First, the terms used in the following description will be explained.

[0024] 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).

[0025] 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.

[0026] 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.

[0027] 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.

[0028] 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."

[0029] [First embodiment]

[0030] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0031] 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.

[0032] 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).

[0033] 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.

[0034] 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.

[0035] 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.

[0036] 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.

[0037] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0038] 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.

[0039] 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.

[0040] 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.

[0041] 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."

[0042] System Overview

[0043] This invention is an autonomous driving system that uses cameras and microphones installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model to infer the user's psychological state.The system aims to provide passengers with a sense of security by adjusting driving parameters based on the inferred psychological state.

[0044] Explanation of program processing

[0045] Data collection

[0046] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. A microphone also records what the user says in real time.

[0047] Data Preprocessing

[0048] Device: Extracts facial features (e.g., the position of the eyes, mouth, and eyebrows) from the collected image data, and extracts features (e.g., pitch, volume, and spectral characteristics) from the audio data, converting the data into a format that is easy for the AI ​​model described below to handle.

[0049] Psychological analysis

[0050] Server: The preprocessed data is sent to the server, which uses a generative AI model to infer the user's psychological state from this data, for example, identifying emotions such as tension, relief, and surprise.

[0051] Adjusting operating parameters

[0052] Server: Generates commands to adjust driving parameters (e.g., speed, acceleration, braking) based on the estimated psychological state. For example, if the user is nervous, the server may slow down the speed or prohibit sudden acceleration.

[0053] Sending adjustment instructions

[0054] Server: Sends the adjusted driving parameters to the automated driving system. This transmission is usually done in real time, and is quickly reflected in the automated driving system inside the vehicle.

[0055] Execution of operation

[0056] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. This allows for safe and secure driving that is tailored to the user's psychological state.

[0057] Specific examples

[0058] Scenario 1: The user is surprised

[0059] Data collection:

[0060] Device: The camera captures the user's surprised expression, and the microphone records the user's high-pitched utterance, such as "I'm scared."

[0061] Data preprocessing:

[0062] Device: Extracts facial features such as wide-open eyes and an open mouth, and extracts high-pitched and high-volume features from audio data.

[0063] Psychological Analysis:

[0064] Server: The generative AI model analyzes these feature data and infers that the user is surprised.

[0065] Adjusting operating parameters:

[0066] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[0067] Sending adjustment orders:

[0068] Server: Sends new driving parameters to the vehicle.

[0069] Execution of the operation:

[0070] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[0071] Scenario 2: User is relaxed

[0072] Data collection:

[0073] Device: The camera captures the user's relaxed facial expression, and the microphone records calm conversations.

[0074] Data preprocessing:

[0075] Device: From facial features, relaxed eye and mouth shapes are extracted, and stable pitch and volume features are extracted from audio data.

[0076] Psychological Analysis:

[0077] Server: The generative AI model analyzes these feature data and infers that the user is relaxed.

[0078] Adjusting operating parameters:

[0079] Server: Maintain normal operating parameters or configure for slightly faster speeds.

[0080] Sending adjustment orders:

[0081] Server: Sends a command to the vehicle to continue normal operation.

[0082] Execution of the operation:

[0083] Terminal: The vehicle follows your commands and you can continue driving relaxed.

[0084] The above is a specific embodiment of the automated driving system of the present invention, which enables driving according to the psychological state of the passengers, providing a greater sense of security.

[0085] The processing flow will be explained below.

[0086] Step 1:

[0087] Data collection

[0088] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. A microphone also records what the user says in real time.

[0089] How it works: The camera takes pictures at 30 frames per second and stores them in storage. The microphone captures audio at high resolution of 24-bit, 96kHz and stores it in storage.

[0090] Step 2:

[0091] Data Preprocessing

[0092] Terminal: Extracts facial features from collected image data, and extracts features such as pitch, volume, and spectral characteristics from audio data.

[0093] Specific operation: The Dlib library is used to identify the positions of the eyes, mouth, and eyebrows from the image data, and 68 feature points are extracted. MFCC (Mel-Frequency Cepstrum Coefficients) is used to extract 13-dimensional feature values ​​from the audio data every 0.1 seconds.

[0094] Step 3:

[0095] Sending data

[0096] Terminal: Sends preprocessed data (facial feature points and audio features) to the server.

[0097] Specific operation: Data is sent asynchronously to a server using HTTP or WebSocket protocol over the Internet or a dedicated in-vehicle network.

[0098] Step 4:

[0099] Psychological analysis

[0100] Server: Inputs the received preprocessed data into the generative AI model and infers the user's psychological state.

[0101] Specific operation: The received data is input into a generative AI model (e.g., BERT or CNN), which outputs an emotion label, such as whether the user is nervous, relaxed, or surprised.

[0102] Step 5:

[0103] Adjusting operating parameters

[0104] Server: Determines driving parameters (speed, acceleration, braking) based on the estimated psychological state.

[0105] Specific behavior: The rule engine is executed to generate new driving parameters, such as slowing down if the psychological state is "surprised" and increasing speed if the psychological state is "relaxed."

[0106] Step 6:

[0107] Sending adjustment instructions

[0108] Server: Sends new driving parameters to the vehicle's autonomous driving system.

[0109] Specific operation: Sends command data to the terminal using HTTP or a dedicated protocol.

[0110] Step 7:

[0111] Execution of operation

[0112] Terminal: Follows instructions from the server and applies new driving parameters to the autonomous driving system.

[0113] Specific operation: The autonomous driving system applies settings such as new speed limits and prohibitions on sudden acceleration to the controller and performs actual driving operations.

[0114] The above is the specific process and operation flow for each step. This system realizes safe and secure driving that is in line with the passenger's psychological state.

[0115] Example 1

[0116] 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."

[0117] In current automated driving systems, vehicle operating parameters are adjusted mainly based on the external environment and traffic conditions, and the psychological state of passengers is not taken into consideration. This often causes passengers to feel nervous or anxious, which can hinder a comfortable journey. In particular, sudden acceleration or braking can increase the psychological burden on passengers.

[0118] 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.

[0119] In this invention, the server includes means for extracting facial feature points from collected facial expression data, means for extracting pitch, volume, and spectral characteristics from voice data, and means for inferring the user's psychological state from the feature data using a generative AI model. This enables adjustment of driving parameters based on the user's psychological state, improving passenger comfort.

[0120] The "camera" is a photographing device that is installed inside the vehicle and collects the user's facial expressions frame by frame in real time.

[0121] A "microphone" is a recording device installed inside a vehicle to collect the user's voice in real time.

[0122] "Facial expression data" is image information of the user's face collected by a camera.

[0123] "Voice data" refers to information about a user's speech or voice collected by a microphone.

[0124] "Feature points" are the main positional information of the face (e.g., the positions of the eyes, mouth, and eyebrows) extracted from the facial expression data.

[0125] "Features" are sound characteristics (e.g., pitch, volume, spectral characteristics) extracted from audio data.

[0126] The "server" is a central processing unit for processing the collected facial expression data and voice data.

[0127] A "generative AI model" is an artificial intelligence model that takes feature data as input and uses it to infer a user's psychological state.

[0128] "Driving parameters" are set values ​​related to driving operations such as vehicle speed, acceleration, and braking.

[0129] The "command" is instruction information generated by the server that includes the details of adjustment of the operating parameters.

[0130] "Real-time transmission" means that data is transmitted almost instantly to a server or to an automated driving system.

[0131] MODE FOR CARRYING OUT THE INVENTION

[0132] The present invention is an autonomous driving system that uses a camera and microphone inside the vehicle to collect a user's facial expressions and voice in real time and uses a generative AI model to infer their psychological state. The system aims to provide passengers with a sense of security by adjusting driving parameters based on the inferred psychological state. A specific embodiment of the system is described below.

[0133] Data collection

[0134] Terminal: A camera installed inside the vehicle captures the user's face frame by frame and collects image data. The camera used is a general-purpose high-resolution camera. The microphone records the user's voice in real time. Specifically, a general-purpose recording device with noise-canceling functionality is used.

[0135] Data Preprocessing

[0136] Device: The facial recognition software "dlib" is used to extract facial features (e.g., the position of the eyes, mouth, and eyebrows) from the collected image data. Also, "Librosa" is used to extract features (e.g., pitch, volume, and spectral characteristics) from the audio data. This process converts the data into a format that is easy for the AI ​​model described below to handle.

[0137] Sending data

[0138] Terminal: The preprocessed data is sent to the server. The "WebSocket" protocol is used for transmission to achieve real-time communication, which enables high-speed, low-latency data transfer.

[0139] Psychological analysis

[0140] Server: Receives the preprocessed data and uses a generative AI model to infer the user's psychological state. The AI ​​model used is a customized model based on GPT-4 (registered trademark). This model receives image and audio feature data as input and classifies the user's psychological state (e.g., nervous, relaxed, surprised).

[0141] Adjusting operating parameters

[0142] Server: Generates commands to adjust driving parameters based on the estimated psychological state. These commands include adjustments to vehicle speed, acceleration, braking, etc. For example, if the user is surprised, it generates a command to slow down the vehicle and prohibit sudden maneuvers. This is done by inputting a prompt such as "Please suggest driving parameters when the user is surprised" into the generative AI model.

[0143] Sending adjustment instructions

[0144] Server: Sends the generated driving parameters to the autonomous driving system. This transmission is also done in real time and is immediately reflected in driving operations. The "MQTT" (Message Queuing Telemetry Transport) protocol is used for transmission, achieving highly reliable and low-latency communication.

[0145] Execution of operation

[0146] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. The vehicle's autonomous driving system reduces speed and controls braking based on the received commands, thereby realizing safe and secure driving for the user.

[0147] Specific examples

[0148] A specific scenario is shown below.

[0149] Scenario 1: The user is surprised

[0150] Data collection:

[0151] Device: The camera captures the user's surprised expression, and the microphone records the user's high-pitched utterances such as "I'm scared."

[0152] Data preprocessing:

[0153] Device: Extracts facial features such as wide-open eyes and an open mouth, and extracts high-pitch and high-volume features from audio data.

[0154] Sending data:

[0155] Terminal: Sends pre-processed data to the server in real time.

[0156] Psychological Analysis:

[0157] Server: The generative AI model analyzes the feature data and infers that the user is surprised.

[0158] Adjusting operating parameters:

[0159] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[0160] Sending adjustment orders:

[0161] Server: Sends new driving parameters to the vehicle.

[0162] Execution of the operation:

[0163] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[0164] Scenario 2: User is relaxed

[0165] Data collection:

[0166] Device: The camera captures the user's relaxed facial expression, and the microphone records calm conversations.

[0167] Data preprocessing:

[0168] Device: Extracts relaxed eye and mouth shapes from facial feature points, and extracts stable pitch and volume features from audio data.

[0169] Sending data:

[0170] Terminal: Sends preprocessed data to the server.

[0171] Psychological Analysis:

[0172] Server: The generative AI model analyzes the feature data and infers that the user is relaxed.

[0173] Adjusting operating parameters:

[0174] Server: Generates commands to maintain normal operating parameters or slightly increase speed.

[0175] Sending adjustment orders:

[0176] Server: Sends a command to the vehicle to continue normal operation.

[0177] Execution of the operation:

[0178] Terminal: The vehicle follows your commands and you can continue driving relaxed.

[0179] The above is a specific embodiment of the automated driving system of the present invention, which enables driving according to the psychological state of the passengers, providing a greater sense of security.

[0180] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0181] Program processing flow

[0182] Step 1: Data collection

[0183] Terminal: The camera captures the user's face frame by frame and collects image data. Specifically, it captures 30 frames per second. The microphone also records the user's voice in real time. Specifically, it uses a general-purpose, highly sensitive recording device and has the function of reducing background noise. This allows high-resolution image data and voice data to be obtained as input.

[0184] Step 2: Data Preprocessing

[0185] Terminal: The facial recognition software "dlib" is used to extract facial feature points (the positions of the eyes, mouth, and eyebrows) from the collected image data. At the same time, "Librosa" is used to extract feature quantities such as pitch, volume, and spectral characteristics from the audio data. Specifically, feature points are identified from the image of each frame, and the audio data is quantified. This results in the extracted feature points and feature quantities being obtained as output.

[0186] Step 3: Sending data

[0187] Terminal: The "WebSocket" protocol is used to send preprocessed data to the server. This allows feature points and feature quantities to be sent to the server at high speed and with low latency. Specifically, image feature points and audio feature quantities for each frame are packaged into packets and sent in real time.

[0188] Step 4: Psychological analysis

[0189] Server: Receives the preprocessed data and uses a generative AI model to infer the user's emotional state. The AI ​​model used is a customized version of GPT-4. Specifically, feature points and feature quantities are input into the model to classify the user's emotions (tension, relaxation, surprise, etc.). During this process, an inferred emotional state is output from the feature points and feature quantities received as input.

[0190] Step 5: Adjusting the operating parameters

[0191] Server: Generates commands to adjust driving parameters based on the estimated psychological state. Based on the output from the model, commands are generated to adjust speed, acceleration, braking, etc. For example, if it is estimated that the user is surprised, a command is generated to slow down the vehicle's speed and prohibit sudden maneuvers. This is done by inputting a prompt statement such as "Please suggest driving parameters for when the user is surprised" into the generative AI model. As a result, an adjustment command is generated.

[0192] Step 6: Sending adjustment instructions

[0193] Server: Sends the generated driving parameters to the autonomous driving system. This transmission is also done in real time and is immediately reflected in driving operations. The "MQTT" protocol is used for transmission. Specifically, a packet containing the generated driving parameters is created and sent to the autonomous driving system, which then outputs an adjustment command.

[0194] Step 7: Execute the operation

[0195] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. The vehicle's autonomous driving system reduces speed and controls braking in accordance with the received commands. This allows the user to drive safely and with peace of mind. Specifically, it reduces speed to 20 km / h and maintains smooth driving. The output of this process is a safe and relaxed driving state.

[0196] (Application example 1)

[0197] 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."

[0198] Conventional automated driving systems have been unable to adjust driving behavior to take into account the psychological state of the driver and passengers, making it difficult to provide a sense of security. Furthermore, there was no mechanism to provide driving advice or suggestions based on the psychological state in real time, so the effectiveness of reducing driver stress and promoting safe driving was limited.

[0199] 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.

[0200] In this invention, the server includes, in a vehicle linked to an automated driving system that performs driving operations, means for collecting a user's facial expressions in real time using an internally installed camera, means for collecting the user's voice in real time using an internally installed microphone, means for preprocessing the collected facial expression data and voice data and extracting feature points and feature quantities, means for inferring the user's psychological state from the extracted feature data using a generative AI model, means for adjusting driving parameters based on the inferred psychological state, means for notifying the user's mobile device of the generated suggestions and providing audio guidance or display, and means for transmitting the adjusted driving parameters to the automated driving system and applying them. This enables driving adjustments and the provision of appropriate advice based on the psychological states of the driver and passengers.

[0201] An "autonomous driving system that performs driving operations" is a system that automatically controls the operation of a vehicle using data input devices such as sensors and cameras.

[0202] A "camera" is an optical device for capturing images and video.

[0203] "User's facial expression" refers to the emotions and psychological state of the user that are indicated by the movements of the user's face and changes in the facial muscles.

[0204] A "microphone" is a sound capture device that converts sound waves into electrical signals.

[0205] "User voice" refers to the voice or sound patterns produced by a user.

[0206] "Facial expression data" refers to a collection of images and feature points of a user's face captured by a camera.

[0207] "Voice data" refers to the waveform information of the user's voice or sound collected by a microphone.

[0208] "Feature points" refer to important points on the face, including the position information of the eyes, nose, mouth, etc.

[0209] "Features" refers to information in audio data that is expressed by specific parameters (e.g., pitch, volume, spectral characteristics).

[0210] A "generative AI model" is an artificial intelligence algorithm used to infer a user's psychological state from collected data.

[0211] "Mental state" refers to the emotions and psychological state that a user is feeling.

[0212] "Driving parameters" refers to adjustable driving-related settings such as vehicle speed, acceleration, braking, etc.

[0213] "Mobile device" refers to a portable computing device such as a smartphone or tablet.

[0214] An "audio guide" is a system that provides users with information and instructions through audio.

[0215] "Notification" refers to an alert or message that notifies a user of specific information.

[0216] "Preprocessing" refers to the step of converting collected data into a form suitable for subsequent analysis and processing.

[0217] "Adjustment" refers to the act of changing driving parameters based on the estimated psychological state.

[0218] "Transmit" refers to the act of transferring data or instructions to another system or device.

[0219] "Application" refers to reflecting the adjusted driving parameters in the actual vehicle operation.

[0220]

[0221] The present invention is an autonomous driving system that uses a camera and microphone inside the vehicle to collect a user's facial expressions and voice in real time, and then uses a generative AI model to infer the user's psychological state based on the collected data. The system of the present invention includes a camera, a microphone, a server, a mobile device such as a smartphone, and a generative AI model.

[0222] System configuration

[0223] 1. Data Collection

[0224] The smartphone's camera and microphone are used to capture the driver's facial expressions and voice in real time. The camera captures facial expressions frame by frame, and the microphone records the voice.

[0225] 2. Data Preprocessing

[0226] The device preprocesses the collected image data using OpenCV to extract facial features, and for audio data, Librosa is used to extract features (pitch, volume, spectral characteristics).

[0227] 3. Psychological analysis

[0228] The preprocessed data is sent to a cloud server, where a generative AI model is used to analyze the user's psychological state. The server achieves this using platforms such as Google® Cloud AI and AWS® SageMaker.

[0229] 4. Adjusting the operating parameters

[0230] The server adjusts driving parameters based on the inferred psychological state, for example generating instructions to reduce speed and avoid sudden maneuvers if the user is nervous.

[0231] 5. Proposal Generation

[0232] The server generates advice based on the user's psychological state, suggests music, and suggests rest areas. These suggestions are sent to the mobile device and provided to the user as audio guidance or on-screen displays.

[0233] 6. Sending adjustment commands and executing operations

[0234] The server then sends the adjusted driving parameters to the vehicle's autonomous driving system and applies them, enabling driving that is appropriate for the driver's psychological state.

[0235] Specific examples

[0236] Scenario 1: Tired driver

[0237] 1. Data collection: The smartphone camera captures the driver's tired expression, and the microphone collects the driver's sigh.

[0238] 2. Data preprocessing: Extract the degree of eye contraction using OpenCV and low-pitched audio using Librosa.

[0239] 3. Psychological state analysis: The generative AI model infers that the driver is tired.

[0240] 4. Proposal generation: The server notifies the mobile device of break suggestions and soothing music playback.

[0241] 5. Sending adjustment instructions and executing driving: The server adjusts the driving parameters, and the vehicle follows the instructions to drive appropriately.

[0242] Prompt Sentence Examples

[0243] Prompt: "Analyze the driver's facial expressions and voice data to infer the driver's fatigue state and generate appropriate rest advice."

[0244] With the above configuration, this system is able to adjust driving and provide appropriate advice according to the psychological state of the driver and passengers, enabling a more secure autonomous driving experience.

[0245] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0246] Step 1:

[0247] The device captures the user's face using a camera installed in the vehicle and collects video data frame by frame. The specific input is image frames captured in real time, and the output is image data frame by frame, which is used for subsequent processing to extract feature points.

[0248] Step 2:

[0249] The terminal collects the user's voice in real time using a microphone installed in the vehicle. The specific input is the user's spoken voice, and the output is voice data. The collected voice data is used to extract features in the next step.

[0250] Step 3:

[0251] The device preprocesses the collected image data and extracts facial feature points. Specifically, it uses OpenCV to detect feature points such as eyes, mouth, and eyebrows from each frame of image. The input is the image data collected in step 1, and the output is the extracted facial feature point data.

[0252] Step 4:

[0253] The device preprocesses the collected audio data and extracts features. Specifically, Librosa is used to extract pitch, volume, and spectral characteristics from the audio data. The input is the audio data collected in step 2, and the output is the extracted audio feature data.

[0254] Step 5:

[0255] The terminal sends the preprocessed feature point data and feature volume data to the server. The specific input is the feature data obtained in step 3 and step 4, and the output is the data sent to the server.

[0256] Step 6:

[0257] The server inputs the received feature data into a generative AI model to infer the user's psychological state. The specific inputs are feature point data and voice feature data, and the output is the inferred psychological state (e.g., relaxed, tense, fatigued). The generative AI model achieves this using Google Cloud AI and AWS SageMaker.

[0258] Step 7:

[0259] The server generates instructions to adjust driving parameters based on the inferred psychological state, e.g., to reduce speed if the user is tired. The input is the psychological state inferred in step 6, and the output is the adjusted driving parameters.

[0260] Step 8:

[0261] The server sends the generated driving parameters or advice to the mobile device. For example, a break suggestion or a music playback instruction is generated in this step. The input is the driving parameters generated in step 7, and the output is the notification content.

[0262] Step 9:

[0263] The device displays or voices the received notification to the driver. Specifically, it displays the suggestion on the screen of the mobile device or informs the user as voice guidance. The input is the notification content sent in step 8, and the output is the display on the device or voice guidance.

[0264] Step 10:

[0265] The server sends the adjusted driving parameters to the automated driving system, and the vehicle performs driving based on these instructions. The input is the driving parameters generated in step 7, and the output is the actual vehicle behavior.

[0266] Example prompt sentence:

[0267] Prompt: "Analyze the driver's facial expressions and voice data to infer the driver's fatigue state and generate appropriate rest advice."

[0268] The above are the processing steps in the embodiment of the present invention.

[0269] 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.

[0270] System Overview

[0271] This invention is an autonomous driving system that uses a camera and microphone installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model and an emotion engine to infer the user's psychological state and adjust driving parameters. By combining this with an emotion engine, it becomes possible to more accurately recognize the user's diverse emotions and perform driving operations based on them.

[0272] Explanation of program processing

[0273] Data collection

[0274] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. A microphone also records what the user says in real time.

[0275] Data Preprocessing

[0276] Device: Extracts facial features (e.g., the position of the eyes, mouth, and eyebrows) from the collected image data, and extracts features (e.g., pitch, volume, and spectral characteristics) from the audio data, converting the data into a format that is easy for the AI ​​model described below to handle.

[0277] Sending data

[0278] Terminal: Sends preprocessed data (facial feature points and audio features) to the server.

[0279] Psychological analysis

[0280] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to infer the user's mental state and emotions, for example, identifying emotions such as tension, relief, surprise, anger, and joy.

[0281] Adjusting operating parameters

[0282] Server: Adjusts driving parameters (speed, acceleration, braking) based on the inferred psychological state and emotions. For example, if the user is angry, the server will slow down the vehicle to ensure safety and drive calmly.

[0283] Sending adjustment instructions

[0284] Server: Sends the adjusted driving parameters to the automated driving system. This transmission is usually done in real time, and is quickly reflected in the automated driving system inside the vehicle.

[0285] Execution of operation

[0286] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. This allows for safe and secure driving that is tailored to the user's psychological state.

[0287] Specific examples

[0288] Scenario 1: The user is angry

[0289] Data collection:

[0290] Device: The camera captures the user's angry facial expression (e.g., frown, closed mouth), and the microphone records the user's strong-tongued remarks.

[0291] Data preprocessing:

[0292] Device: Extract anger features from image data using the Dlib library, and extract high-pitch and high-volume features from audio data.

[0293] Psychological Analysis:

[0294] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is angry."

[0295] Adjusting operating parameters:

[0296] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[0297] Sending adjustment orders:

[0298] Server: Sends new driving parameters to the vehicle.

[0299] Execution of the operation:

[0300] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[0301] Scenario 2: Happy users

[0302] Data collection:

[0303] Device: The camera captures the user's smile, and the microphone records the user's conversation in a cheerful tone.

[0304] Data preprocessing:

[0305] Device: From the facial features, smiling features (such as the corners of the mouth turning up and the corners of the eyes turning up) are extracted, and stable pitch and volume features are extracted from the audio data.

[0306] Psychological Analysis:

[0307] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is happy."

[0308] Adjusting operating parameters:

[0309] Server: Maintain normal operating parameters or configure for slightly faster speeds.

[0310] Sending adjustment orders:

[0311] Server: Sends a command to the vehicle to continue normal operation.

[0312] Execution of the operation:

[0313] Terminal: The vehicle follows the command and maintains its current operating parameters.

[0314] The above is a specific embodiment of the automated driving system of the present invention. This allows driving to be performed in accordance with the diverse emotions and psychological states of passengers, thereby improving safety and peace of mind.

[0315] The processing flow will be explained below.

[0316] Specific processing steps of the system

[0317] Step 1:

[0318] Data collection

[0319] Terminal: A camera installed inside the vehicle captures the user's facial expressions frame by frame and collects image data. A microphone also records the user's speech in real time.

[0320] How it works: The camera takes pictures at 30 frames per second and stores them in storage. The microphone captures audio at high resolution of 24-bit, 96kHz and stores it in storage.

[0321] Step 2:

[0322] Data Preprocessing

[0323] Terminal: Extracts facial features (e.g., eye, mouth, and eyebrow positions) from collected image data, and extracts features (e.g., pitch, volume, and spectral characteristics) from audio data.

[0324] Specific operation: The Dlib library is used to identify the positions of the eyes, mouth, and eyebrows from the image data, and 68 feature points are extracted. MFCC (Mel-Frequency Cepstrum Coefficients) is used to extract 13-dimensional feature values ​​from the audio data every 0.1 seconds.

[0325] Step 3:

[0326] Sending data

[0327] Terminal: Sends preprocessed data (facial feature points and audio features) to the server.

[0328] Specific operation: Data is sent asynchronously to a server via the Internet or a dedicated in-vehicle network using HTTP or WebSocket protocols.

[0329] Step 4:

[0330] emotion recognition

[0331] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to recognize the user's emotions. The generative AI model estimates the psychological state, and the emotion engine identifies the emotion.

[0332] Specific behavior: The generative AI model analyzes the pre-processed data and outputs psychological states such as "tension," "relief," and "surprise." Based on this, the emotion engine classifies the user's specific emotions into "anger," "joy," "sadness," etc.

[0333] Step 5:

[0334] Adjusting operating parameters

[0335] Server: Determines driving parameters (speed, acceleration, braking) based on the inferred psychological state and emotions.

[0336] Specific behavior: Uses a rules engine to generate driving commands such as "slow down if the user is angry" or "maintain speed if the user is happy."

[0337] Step 6:

[0338] Sending adjustment instructions

[0339] Server: Sends the generated new driving parameters to the autonomous driving system.

[0340] Specific operation: Driving parameters are sent to the terminal in the vehicle using HTTP or a dedicated protocol.

[0341] Step 7:

[0342] Execution of operation

[0343] Terminal: Apply the new driving parameters received from the server to the autonomous driving system.

[0344] Specific behavior: The autonomous driving system uses the new control parameters to perform driving maneuvers such as speed, acceleration, and braking. For example, if the user is nervous, the system will reduce the speed to 20 km / h.

[0345] Specific examples

[0346] Scenario 1: The user is angry

[0347] Step 1:

[0348] Data collection

[0349] Device: The camera captures the user's angry facial expression (frown, tight-lipped mouth), and the microphone records the user's strong-tongued remarks.

[0350] Step 2:

[0351] Data Preprocessing

[0352] Terminal: Use the Dlib library to extract facial features, and high-pitch and high-volume features from audio data.

[0353] Step 3:

[0354] Sending data

[0355] Terminal: Sends preprocessed data to the server.

[0356] Step 4:

[0357] emotion recognition

[0358] Server: The generative AI model and emotion engine analyze the data and recognize that the user is feeling "anger."

[0359] Step 5:

[0360] Adjusting operating parameters

[0361] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[0362] Step 6:

[0363] Sending adjustment instructions

[0364] Server: Sends new driving parameters to the vehicle.

[0365] Step 7:

[0366] Execution of operation

[0367] Terminal: Reduce speed to 20km / h and begin driving smoothly.

[0368] Scenario 2: Happy users

[0369] Step 1:

[0370] Data collection

[0371] Device: The camera captures the user's smile. The microphone records the user's pleasant conversation.

[0372] Step 2:

[0373] Data Preprocessing

[0374] Device: Extracts features of a smile, such as eyes and mouth turned up, and extracts stable pitch and volume features from audio data.

[0375] Step 3:

[0376] Sending data

[0377] Terminal: Sends preprocessed data to the server.

[0378] Step 4:

[0379] emotion recognition

[0380] Server: The generative AI model and emotion engine analyze the data and recognize that the user is feeling "joy."

[0381] Step 5:

[0382] Adjusting operating parameters

[0383] Server: Issues a command to continue normal operation.

[0384] Step 6:

[0385] Sending adjustment instructions

[0386] Server: Sends commands to the vehicle.

[0387] Step 7:

[0388] Execution of operation

[0389] Terminal: Maintain current operating parameters and continue normal operation.

[0390] This will enable safe and secure driving based on specific emotions.

[0391] Example 2

[0392] 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."

[0393] In an automated driving system for a vehicle, there is a demand for providing a safer and more comfortable driving experience by accurately grasping the user's psychological state and adjusting driving parameters accordingly. However, with conventional technologies, it has been difficult to accurately recognize the user's emotions and psychological state in real time and dynamically adjust driving parameters based on that. This invention is intended to solve this problem.

[0394] 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.

[0395] In this invention, the server, in a vehicle linked to an automated driving system that estimates a user's psychological state and adjusts driving parameters, includes: means for collecting a user's facial expression in real time using an internally installed image capture device; means for collecting the user's voice in real time using an internally installed voice recording device; means for preprocessing the collected facial expression data and voice data and extracting feature points and feature quantities; means for estimating the user's psychological state from the extracted feature data using a generative technology model; means for adjusting the driving parameters based on the estimated psychological state; and means for transmitting the adjusted driving parameters to the automated driving system and applying them, thereby enabling appropriate adjustment of driving parameters based on the user's real-time psychological state.

[0396] A "user" is a person who uses this system.

[0397] "Mental state" refers to the user's feelings and mental state, and specifically includes emotions such as tension, relief, surprise, anger, and joy.

[0398] "Driving parameters" are control data related to vehicle driving, specifically settings such as speed, acceleration, and braking.

[0399] An "autonomous driving system" is a system that automatically controls the driving of a vehicle and has the function of allowing the vehicle to drive autonomously without the user performing any driving operations.

[0400] The "photography device" is a device that is installed inside the vehicle and captures the user's facial expressions in real time.

[0401] The "voice recording device" is a device that is installed inside the vehicle and records the user's speech in real time.

[0402] "Preprocessing" refers to the process of extracting feature points and feature quantities from collected raw data (image data and audio data).

[0403] "Feature points" refer to characteristic points on the user's face extracted from image data, such as the positions of the eyes, mouth, and eyebrows.

[0404] "Features" are acoustic features extracted from audio data, specifically pitch, volume, spectral characteristics, etc.

[0405] A "generative model" is an artificial intelligence model used to infer a user's psychological state from collected data.

[0406] "Adjusted driving parameters" refer to new driving parameters that have been adjusted based on the user's psychological state.

[0407] "Application" means reflecting the adjusted driving parameters in the actual automated driving system.

[0408] System Overview

[0409] This invention is an autonomous driving system that uses a camera and microphone installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model and emotion engine to infer the user's psychological state and adjust driving parameters. By combining this with an emotion engine, it is possible to more accurately recognize the user's diverse emotions and perform driving operations based on them.

[0410] Data collection

[0411] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. An audio recording device also records the user's speech in real time.

[0412] Hardware used:

[0413] Camera (e.g. high-resolution camera)

[0414] Microphone (e.g. high-sensitivity microphone)

[0415] Data Preprocessing

[0416] Device: Facial feature points (e.g., eye, mouth, and eyebrow positions) are extracted from the collected image data using a face recognition library (e.g., Dlib), while acoustic features (e.g., pitch, volume, and spectral characteristics) are extracted from the audio data.

[0417] Software used:

[0418] Dlib library (for face recognition)

[0419] Audio analysis tools (e.g. FFT)

[0420] Sending data

[0421] Terminal: The facial feature point data and voice feature data obtained in the preprocessing are encrypted and sent to the server using the HTTPS protocol to ensure data security.

[0422] Psychological analysis

[0423] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to infer the user's psychological state. For example, identify emotions such as tension, relief, surprise, anger, and joy.

[0424] Software used:

[0425] Generative technology models (e.g., BERT-based sentiment analysis models)

[0426] Emotion Engine

[0427] Adjusting operating parameters

[0428] Server: Optimizes driving parameters (speed, acceleration, braking) based on the estimated psychological state. If the user is nervous, the server reduces speed and smooths braking to make the driver more calm.

[0429] Sending adjustment instructions

[0430] Server: Transmits the calculated driving parameters to the vehicle's autonomous driving system in real time. This transmission also uses the HTTPS protocol to ensure data integrity and security.

[0431] Execution of operation

[0432] Terminal: The terminal applies the new driving parameters received from the server to the actual vehicle control and performs driving, thereby automatically driving according to the user's psychological state.

[0433] Specific examples

[0434] Scenario 1: The user is angry

[0435] Data collection:

[0436] Device: The camera captures the user's angry facial expression (e.g., frown, closed mouth), and the microphone records the user's strong-tongued remarks.

[0437] Data preprocessing:

[0438] Device: Extract anger features from image data using the Dlib library, and extract high-pitch and high-volume features from audio data.

[0439] Psychological Analysis:

[0440] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is angry."

[0441] Adjusting operating parameters:

[0442] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[0443] Sending adjustment orders:

[0444] Server: Sends new driving parameters to the vehicle.

[0445] Execution of the operation:

[0446] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[0447] Scenario 2: Happy users

[0448] Data collection:

[0449] Device: The camera captures the user's smile, and the microphone records the user's conversation in a cheerful tone.

[0450] Data preprocessing:

[0451] Device: From the facial features, smiling features (such as the corners of the mouth turning up and the corners of the eyes turning up) are extracted, and stable pitch and volume features are extracted from the audio data.

[0452] Psychological Analysis:

[0453] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is happy."

[0454] Adjusting operating parameters:

[0455] Server: Maintain normal operating parameters or configure for slightly faster speeds.

[0456] Sending adjustment orders:

[0457] Server: Sends a command to the vehicle to continue normal operation.

[0458] Execution of the operation:

[0459] Terminal: The vehicle follows the command and maintains its current operating parameters.

[0460] Prompt Sentence Examples

[0461] Scenario 1 prompt statement:

[0462] If the user is angry, please follow the instructions below to adjust the operating parameters:

[0463] Angry facial expression captured on camera

[0464] Logged voice data with strong verbal content

[0465] Scenario 2 prompt statement:

[0466] If the user is willing, please follow the instructions below to adjust the operating parameters:

[0467] A photo of a smiling face captured on camera

[0468] Logged voice data with a pleasant tone

[0469] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0470] Step 1: Data collection

[0471] Device: The in-car camera captures the user's face frame by frame and collects image data, while the microphone records what the user says in real time.

[0472] Input: Video frames of the user's face, audio of what the user is saying

[0473] Output: User's face image data, voice data

[0474] Specific behavior:

[0475] The interior camera captures images at 30 frames per second.

[0476] The microphone records audio at a sample rate of 16kHz.

[0477] Step 2: Data Preprocessing

[0478] Device: Facial feature points (e.g., eye, mouth, and eyebrow positions) are extracted from the collected image data using a face recognition library (e.g., Dlib), while acoustic features (e.g., pitch, volume, and spectral characteristics) are extracted from the audio data.

[0479] Input: Facial image data, audio data

[0480] Output: Facial feature data, acoustic feature data

[0481] Specific behavior:

[0482] Using the Dlib library, 68 facial feature points are extracted from each frame.

[0483] The frequency components of the audio data are analyzed using FFT (Fast Fourier Transform) and extracted as features.

[0484] Step 3: Sending data

[0485] Terminal: The facial feature point data and voice feature data obtained in the preprocessing are encrypted and sent to the server using the HTTPS protocol to ensure data security.

[0486] Input: facial feature point data, acoustic feature data

[0487] Output: Notification that encrypted data has been sent

[0488] Specific behavior:

[0489] Package the data in JSON format and send a POST request to the server.

[0490] Step 4: Psychological analysis

[0491] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to infer the user's psychological state. For example, identify emotions such as tension, relief, surprise, anger, and joy.

[0492] Input: Preprocessed data (facial feature points, acoustic features)

[0493] Output: User's mental state and emotion category

[0494] Specific behavior:

[0495] Input feature data into a generative AI model (e.g., a BERT-based sentiment analysis model) to predict sentiment categories.

[0496] The emotion engine analyzes the prediction results and determines the most appropriate emotion.

[0497] Step 5: Adjusting the operating parameters

[0498] Server: Optimizes driving parameters (speed, acceleration, braking) based on the estimated psychological state. If the user is nervous, the server reduces speed and smooths braking to make the driver more calm.

[0499] Input: User's mental state and emotion category

[0500] Output: Adjusted operating parameters

[0501] Specific behavior:

[0502] It obtains driving data such as current speed and acceleration and calculates new driving parameters based on that.

[0503] Step 6: Sending adjustment instructions

[0504] Server: Transmits the calculated driving parameters to the vehicle's autonomous driving system in real time. This transmission also uses the HTTPS protocol to ensure data integrity and security.

[0505] Input: Adjusted operating parameters

[0506] Output: Notification of completion of sending of operation parameters

[0507] Specific behavior:

[0508] The new driving parameters are packaged in JSON format and sent as a POST request to the vehicle's control system.

[0509] Step 7: Execute the operation

[0510] Terminal: The terminal applies the new driving parameters received from the server to the actual vehicle control and performs driving, thereby automatically driving according to the user's psychological state.

[0511] Input: New operating parameters

[0512] Output: Vehicle control actions based on driving parameters

[0513] Specific behavior:

[0514] The vehicle's ECU (Electronic Control Unit) receives the new driving parameters and reflects them in brake and accelerator control.

[0515] (Application example 2)

[0516] 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."

[0517] Conventional automated driving systems operate based on predetermined driving parameters without considering the user's emotions or psychological state, and therefore cannot fully guarantee the user's safety or comfort. Furthermore, they lack a mechanism for providing support content that adapts to the user's psychological state while driving, making it difficult to promote stress reduction and relaxation for the user. Therefore, there is a need for a system that can adjust driving parameters and provide support content appropriately according to the user's psychological state.

[0518] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting the user's facial expressions in real time using an internally installed camera, means for collecting the user's voice in real time using an internally installed microphone, means for preprocessing the collected facial expression data and voice data to extract feature points and feature quantities, means for inferring the user's psychological state from the extracted feature data using a generative AI model, means for adjusting driving parameters based on the inferred psychological state, means for transmitting the adjusted driving parameters to the automated driving system and applying them, and means for providing support content during driving in accordance with the user's emotional state. This makes it possible to adjust driving parameters in accordance with the user's psychological state and provide appropriate support content.

[0519] An "autonomous driving system that performs driving operations" is a system that automatically drives a vehicle, and is a general term for the programs and hardware that perform driving operations such as adjusting speed, changing direction, and braking.

[0520] The "interiorly installed camera" is a camera device that is attached to the interior of the vehicle and captures the user's facial expressions in real time.

[0521] The "interiorly installed microphone" is a microphone device that is installed inside the vehicle and collects the user's voice in real time.

[0522] "Facial expression data" is image data of the user's face captured by a camera, and includes facial feature points.

[0523] "Voice data" refers to voice data of a user's speech recorded by a microphone, and includes voice features.

[0524] "Preprocessing" refers to a data processing procedure for extracting feature points and feature quantities from collected data.

[0525] "Feature points" are data points relating to the position and shape of a face, and indicate specific positions such as the eyes, mouth, and eyebrows.

[0526] "Features" are specific numerical data such as pitch, volume, and spectral characteristics extracted from audio data.

[0527] A "generative AI model" is an artificial intelligence model that analyzes collected data and infers the user's psychological state.

[0528] The "emotion engine" is an analytical engine that identifies the user's emotions and generates appropriate responses based on them.

[0529] "Driving parameters" are set values ​​for controlling the driving of a vehicle, and include speed, acceleration, braking, and the like.

[0530] "Support content" refers to supplementary information and entertainment provided while driving according to the user's psychological state, and includes relaxing music, positive news, and the like.

[0531] This invention is an autonomous driving system that uses a camera and microphone installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model and emotion engine to infer the user's psychological state and adjust driving parameters.Furthermore, it is a system that provides support content according to the user's psychological state.

[0532] System Overview

[0533] Hardware used

[0534] Camera: An image capturing device installed inside the vehicle.

[0535] Microphone: An audio collection device installed inside the vehicle.

[0536] Smart glasses: A wearable device for capturing the user's facial expressions and voice in real time.

[0537] Minicomputer: A small computer (e.g., Raspberry Pi) for data preprocessing and analysis.

[0538] On-board computer: An on-board computer used to adjust the vehicle's operating parameters.

[0539] Software used

[0540] Facial expression recognition library: Dlib (a library for extracting facial feature points).

[0541] Speech analysis library: LibROSA (a library for extracting speech features).

[0542] Generative AI models: Natural language processing models such as GPT-3 (registered trademark).

[0543] Sentiment Engine: A sentiment analysis engine such as Azure® Cognitive Services.

[0544] Specific Embodiments

[0545] Data collection

[0546] The device (smart glasses) captures the user's facial expressions in real time with a camera installed inside the vehicle and records the user's voice with a microphone, and these data are sent to a minicomputer.

[0547] Data Preprocessing

[0548] The minicomputer uses Dlib to extract facial features from image data, and LibROSA to extract audio features such as pitch, volume, and spectral characteristics from audio data.

[0549] Psychological analysis

[0550] The server inputs the preprocessed data into a generative AI model and emotion engine to infer the user's psychological state, and determines whether the user is happy, angry, or other emotional state based on the inference.

[0551] Providing support content

[0552] If the user is inferred to be angry, the following prompt sentence can be input into the generative AI model to generate instructions to provide relaxing music:

[0553] "User's facial expression data: {facial expression data}

[0554] User's voice data: {voice data}

[0555] Inferred emotion: Anger

[0556] Suggested action: Play relaxing music

[0557] Based on this instruction, the appropriate music will be played.

[0558] If the user is deemed happy, the generative AI model is given the following prompt:

[0559] "User's facial expression data: {facial expression data}

[0560] User's voice data: {voice data}

[0561] Inferred emotion: Joy

[0562] Recommended action: Read positive news

[0563] This will read out positive news aloud.

[0564] Adjusting operating parameters

[0565] The server adjusts driving parameters (speed, braking, etc.) based on the inferred psychological state. For example, if the user is angry, it generates an instruction to slow down the vehicle and sends it to the on-board computer. This automatically adjusts the vehicle's driving, increasing the user's safety.

[0566] Example

[0567] Scenario 1: The user is angry

[0568] Data collection:

[0569] The camera captures the user's angry expression (frowns, closed mouth), and the microphone records the strong words spoken.

[0570] Sentiment Analysis and Response:

[0571] The generative AI model and emotion engine infer that the user is angry and generate instructions to provide relaxing music.

[0572] Adjusting operating parameters:

[0573] The server generates and transmits to the vehicle instructions to reduce the vehicle's speed.

[0574] Scenario 2: Happy users

[0575] Data collection:

[0576] The camera captures the user's smile, and the microphone records conversations in a playful tone.

[0577] Sentiment Analysis and Response:

[0578] The generative AI model and emotion engine infer that the user is happy and generate instructions to deliver positive news.

[0579] Adjusting operating parameters:

[0580] The server generates and transmits instructions to the vehicle to maintain normal operating parameters.

[0581] This will enable the adjustment of driving parameters according to the user's psychological state and the provision of appropriate support content, resulting in safer and more comfortable automated driving.

[0582] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0583] Step 1:

[0584] The device (smart glasses) captures images of the user's face in real time using an internal camera and acquires image data. It also records the user's voice in real time using an internal microphone and acquires voice data. This input data (image data and voice data) is sent to a minicomputer.

[0585] Step 2:

[0586] The minicomputer extracts facial features (e.g., eye, mouth, and eyebrow positions) from the received image data using the Dlib library, and simultaneously extracts pitch, volume, and spectral characteristics from the audio data using the LibROSA library, resulting in preprocessed data (facial features and audio features).

[0587] Step 3:

[0588] The preprocessed data is sent to the server. The input data (facial features and voice features) is fed into the generative AI model and emotion engine to infer the user's psychological state. The generative AI model analyzes the combined facial and voice data and identifies emotions such as "the user is angry" or "the user is happy."

[0589] Step 4:

[0590] The server adjusts driving parameters based on the estimated psychological state. Specifically, it generates instructions to set driving parameters such as vehicle speed, acceleration, and braking. For example, if it estimates that the user is angry, it generates an instruction to slow down the vehicle.

[0591] Step 5:

[0592] The server sends the generated instructions to the vehicle's on-board computer. The input data (adjusted driving parameters) are applied by the on-board computer and reflected in actual driving operations. This allows driving to be performed in accordance with the user's psychological state.

[0593] Step 6:

[0594] The server provides support content according to the user's psychological state. For example, if the user is angry, the server generates an instruction to play relaxing music and sends it to the minicomputer. The minicomputer then plays the relaxing music according to the instruction.

[0595] Step 7:

[0596] If the user's psychological state is calmed by the support content they receive, the camera and microphone will again collect their facial expressions and voice, and the entire system will start operating again, making it possible to provide continuous driving support that responds to the user's psychological state.

[0597] As a result, safe and comfortable automated driving is achieved by adjusting driving parameters according to the user's psychological state and providing support content.

[0598] 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.

[0599] 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.

[0600] 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.

[0601] [Second embodiment]

[0602] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0603] 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.

[0604] 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).

[0605] 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.

[0606] 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.

[0607] 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).

[0608] 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.

[0609] 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.

[0610] 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.

[0611] 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.

[0612] In the smart glasses 214, the 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.

[0613] 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."

[0614] System Overview

[0615] This invention is an autonomous driving system that uses cameras and microphones installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model to infer the user's psychological state.The system aims to provide passengers with a sense of security by adjusting driving parameters based on the inferred psychological state.

[0616] Explanation of program processing

[0617] Data collection

[0618] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. A microphone also records what the user says in real time.

[0619] Data Preprocessing

[0620] Device: Extracts facial features (e.g., the position of the eyes, mouth, and eyebrows) from the collected image data, and extracts features (e.g., pitch, volume, and spectral characteristics) from the audio data, converting the data into a format that is easy for the AI ​​model described below to handle.

[0621] Psychological analysis

[0622] Server: The preprocessed data is sent to the server, which uses a generative AI model to infer the user's psychological state from this data, for example, identifying emotions such as tension, relief, and surprise.

[0623] Adjusting operating parameters

[0624] Server: Generates commands to adjust driving parameters (e.g., speed, acceleration, braking) based on the estimated psychological state. For example, if the user is nervous, the server may slow down the speed or prohibit sudden acceleration.

[0625] Sending adjustment instructions

[0626] Server: Sends the adjusted driving parameters to the automated driving system. This transmission is usually done in real time, and is quickly reflected in the automated driving system inside the vehicle.

[0627] Execution of operation

[0628] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. This allows for safe and secure driving that is tailored to the user's psychological state.

[0629] Specific examples

[0630] Scenario 1: The user is surprised

[0631] Data collection:

[0632] Device: The camera captures the user's surprised expression, and the microphone records the user's high-pitched utterance, such as "I'm scared."

[0633] Data preprocessing:

[0634] Device: Extracts facial features such as wide-open eyes and an open mouth, and extracts high-pitched and high-volume features from audio data.

[0635] Psychological Analysis:

[0636] Server: The generative AI model analyzes these feature data and infers that the user is surprised.

[0637] Adjusting operating parameters:

[0638] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[0639] Sending adjustment orders:

[0640] Server: Sends new driving parameters to the vehicle.

[0641] Execution of the operation:

[0642] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[0643] Scenario 2: User is relaxed

[0644] Data collection:

[0645] Device: The camera captures the user's relaxed facial expression, and the microphone records calm conversations.

[0646] Data preprocessing:

[0647] Device: From facial features, relaxed eye and mouth shapes are extracted, and stable pitch and volume features are extracted from audio data.

[0648] Psychological Analysis:

[0649] Server: The generative AI model analyzes these feature data and infers that the user is relaxed.

[0650] Adjusting operating parameters:

[0651] Server: Maintain normal operating parameters or configure for slightly faster speeds.

[0652] Sending adjustment orders:

[0653] Server: Sends a command to the vehicle to continue normal operation.

[0654] Execution of the operation:

[0655] Terminal: The vehicle follows your commands and you can continue driving relaxed.

[0656] The above is a specific embodiment of the automated driving system of the present invention, which enables driving according to the psychological state of the passengers, providing a greater sense of security.

[0657] The processing flow will be explained below.

[0658] Step 1:

[0659] Data collection

[0660] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. A microphone also records what the user says in real time.

[0661] How it works: The camera takes pictures at 30 frames per second and stores them in storage. The microphone captures audio at high resolution of 24-bit, 96kHz and stores it in storage.

[0662] Step 2:

[0663] Data Preprocessing

[0664] Terminal: Extracts facial features from collected image data, and extracts features such as pitch, volume, and spectral characteristics from audio data.

[0665] Specific operation: The Dlib library is used to identify the positions of the eyes, mouth, and eyebrows from the image data, and 68 feature points are extracted. MFCC (Mel-Frequency Cepstrum Coefficients) is used to extract 13-dimensional feature values ​​from the audio data every 0.1 seconds.

[0666] Step 3:

[0667] Sending data

[0668] Terminal: Sends preprocessed data (facial feature points and audio features) to the server.

[0669] Specific operation: Data is sent asynchronously to a server using HTTP or WebSocket protocol over the Internet or a dedicated in-vehicle network.

[0670] Step 4:

[0671] Psychological analysis

[0672] Server: Inputs the received preprocessed data into the generative AI model and infers the user's psychological state.

[0673] Specific operation: The received data is input into a generative AI model (e.g., BERT or CNN), which outputs an emotion label, such as whether the user is nervous, relaxed, or surprised.

[0674] Step 5:

[0675] Adjusting operating parameters

[0676] Server: Determines driving parameters (speed, acceleration, braking) based on the estimated psychological state.

[0677] Specific behavior: The rule engine is executed to generate new driving parameters, such as slowing down if the psychological state is "surprised" and increasing speed if the psychological state is "relaxed."

[0678] Step 6:

[0679] Sending adjustment instructions

[0680] Server: Sends new driving parameters to the vehicle's autonomous driving system.

[0681] Specific operation: Sends command data to the terminal using HTTP or a dedicated protocol.

[0682] Step 7:

[0683] Execution of operation

[0684] Terminal: Follows instructions from the server and applies new driving parameters to the autonomous driving system.

[0685] Specific operation: The autonomous driving system applies settings such as new speed limits and prohibitions on sudden acceleration to the controller and performs actual driving operations.

[0686] The above is the specific process and operation flow for each step. This system realizes safe and secure driving that is in line with the passenger's psychological state.

[0687] Example 1

[0688] 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."

[0689] In current automated driving systems, vehicle operating parameters are adjusted mainly based on the external environment and traffic conditions, and the psychological state of passengers is not taken into consideration. This often causes passengers to feel nervous or anxious, which can hinder a comfortable journey. In particular, sudden acceleration or braking can increase the psychological burden on passengers.

[0690] 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.

[0691] In this invention, the server includes means for extracting facial feature points from collected facial expression data, means for extracting pitch, volume, and spectral characteristics from voice data, and means for inferring the user's psychological state from the feature data using a generative AI model. This enables adjustment of driving parameters based on the user's psychological state, improving passenger comfort.

[0692] The "camera" is a photographing device that is installed inside the vehicle and collects the user's facial expressions frame by frame in real time.

[0693] A "microphone" is a recording device installed inside a vehicle to collect the user's voice in real time.

[0694] "Facial expression data" is image information of the user's face collected by a camera.

[0695] "Voice data" refers to information about a user's speech or voice collected by a microphone.

[0696] "Feature points" are the main positional information of the face (e.g., the positions of the eyes, mouth, and eyebrows) extracted from the facial expression data.

[0697] "Features" are sound characteristics (e.g., pitch, volume, spectral characteristics) extracted from audio data.

[0698] The "server" is a central processing unit for processing the collected facial expression data and voice data.

[0699] A "generative AI model" is an artificial intelligence model that takes feature data as input and uses it to infer a user's psychological state.

[0700] "Driving parameters" are set values ​​related to driving operations such as vehicle speed, acceleration, and braking.

[0701] The "command" is instruction information generated by the server that includes the details of adjustment of the operating parameters.

[0702] "Real-time transmission" means that data is transmitted almost instantly to a server or to an automated driving system.

[0703] MODE FOR CARRYING OUT THE INVENTION

[0704] The present invention is an autonomous driving system that uses a camera and microphone inside the vehicle to collect a user's facial expressions and voice in real time and uses a generative AI model to infer their psychological state. The system aims to provide passengers with a sense of security by adjusting driving parameters based on the inferred psychological state. A specific embodiment of the system is described below.

[0705] Data collection

[0706] Terminal: A camera installed inside the vehicle captures the user's face frame by frame and collects image data. The camera used is a general-purpose high-resolution camera. The microphone records the user's voice in real time. Specifically, a general-purpose recording device with noise-canceling functionality is used.

[0707] Data Preprocessing

[0708] Device: The facial recognition software "dlib" is used to extract facial features (e.g., the position of the eyes, mouth, and eyebrows) from the collected image data. Also, "Librosa" is used to extract features (e.g., pitch, volume, and spectral characteristics) from the audio data. This process converts the data into a format that is easy for the AI ​​model described below to handle.

[0709] Sending data

[0710] Terminal: The preprocessed data is sent to the server. The "WebSocket" protocol is used for transmission to achieve real-time communication, which enables high-speed, low-latency data transfer.

[0711] Psychological analysis

[0712] Server: Receives the preprocessed data and uses a generative AI model to infer the user's psychological state. The AI ​​model used is a customized model based on GPT-4. This model receives image and audio feature data as input and classifies the user's psychological state (e.g., nervous, relaxed, surprised).

[0713] Adjusting operating parameters

[0714] Server: Generates commands to adjust driving parameters based on the estimated psychological state. These commands include adjustments to vehicle speed, acceleration, braking, etc. For example, if the user is surprised, it generates a command to slow down the vehicle and prohibit sudden maneuvers. This is done by inputting a prompt such as "Please suggest driving parameters when the user is surprised" into the generative AI model.

[0715] Sending adjustment instructions

[0716] Server: Sends the generated driving parameters to the autonomous driving system. This transmission is also done in real time and is immediately reflected in driving operations. The "MQTT" (Message Queuing Telemetry Transport) protocol is used for transmission, achieving highly reliable and low-latency communication.

[0717] Execution of operation

[0718] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. The vehicle's autonomous driving system reduces speed and controls braking based on the received commands, thereby realizing safe and secure driving for the user.

[0719] Specific examples

[0720] A specific scenario is shown below.

[0721] Scenario 1: The user is surprised

[0722] Data collection:

[0723] Device: The camera captures the user's surprised expression, and the microphone records the user's high-pitched utterances such as "I'm scared."

[0724] Data preprocessing:

[0725] Device: Extracts facial features such as wide-open eyes and an open mouth, and extracts high-pitch and high-volume features from audio data.

[0726] Sending data:

[0727] Terminal: Sends pre-processed data to the server in real time.

[0728] Psychological Analysis:

[0729] Server: The generative AI model analyzes the feature data and infers that the user is surprised.

[0730] Adjusting operating parameters:

[0731] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[0732] Sending adjustment orders:

[0733] Server: Sends new driving parameters to the vehicle.

[0734] Execution of the operation:

[0735] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[0736] Scenario 2: User is relaxed

[0737] Data collection:

[0738] Device: The camera captures the user's relaxed facial expression, and the microphone records calm conversations.

[0739] Data preprocessing:

[0740] Device: Extracts relaxed eye and mouth shapes from facial feature points, and extracts stable pitch and volume features from audio data.

[0741] Sending data:

[0742] Terminal: Sends preprocessed data to the server.

[0743] Psychological Analysis:

[0744] Server: The generative AI model analyzes the feature data and infers that the user is relaxed.

[0745] Adjusting operating parameters:

[0746] Server: Generates commands to maintain normal operating parameters or slightly increase speed.

[0747] Sending adjustment orders:

[0748] Server: Sends a command to the vehicle to continue normal operation.

[0749] Execution of the operation:

[0750] Terminal: The vehicle follows your commands and you can continue driving relaxed.

[0751] The above is a specific embodiment of the automated driving system of the present invention, which enables driving according to the psychological state of the passengers, providing a greater sense of security.

[0752] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0753] Program processing flow

[0754] Step 1: Data collection

[0755] Terminal: The camera captures the user's face frame by frame and collects image data. Specifically, it captures 30 frames per second. The microphone also records the user's voice in real time. Specifically, it uses a general-purpose, highly sensitive recording device and has the function of reducing background noise. This allows high-resolution image data and voice data to be obtained as input.

[0756] Step 2: Data Preprocessing

[0757] Terminal: The facial recognition software "dlib" is used to extract facial feature points (the positions of the eyes, mouth, and eyebrows) from the collected image data. At the same time, "Librosa" is used to extract feature quantities such as pitch, volume, and spectral characteristics from the audio data. Specifically, feature points are identified from the image of each frame, and the audio data is quantified. This results in the extracted feature points and feature quantities being obtained as output.

[0758] Step 3: Sending data

[0759] Terminal: The "WebSocket" protocol is used to send preprocessed data to the server. This allows feature points and feature quantities to be sent to the server at high speed and with low latency. Specifically, image feature points and audio feature quantities for each frame are packaged into packets and sent in real time.

[0760] Step 4: Psychological analysis

[0761] Server: Receives the preprocessed data and uses a generative AI model to infer the user's emotional state. The AI ​​model used is a customized version of GPT-4. Specifically, feature points and feature quantities are input into the model to classify the user's emotions (tension, relaxation, surprise, etc.). During this process, an inferred emotional state is output from the feature points and feature quantities received as input.

[0762] Step 5: Adjusting the operating parameters

[0763] Server: Generates commands to adjust driving parameters based on the estimated psychological state. Based on the output from the model, commands are generated to adjust speed, acceleration, braking, etc. For example, if it is estimated that the user is surprised, a command is generated to slow down the vehicle's speed and prohibit sudden maneuvers. This is done by inputting a prompt statement such as "Please suggest driving parameters for when the user is surprised" into the generative AI model. As a result, an adjustment command is generated.

[0764] Step 6: Sending adjustment instructions

[0765] Server: Sends the generated driving parameters to the autonomous driving system. This transmission is also done in real time and is immediately reflected in driving operations. The "MQTT" protocol is used for transmission. Specifically, a packet containing the generated driving parameters is created and sent to the autonomous driving system, which then outputs an adjustment command.

[0766] Step 7: Execute the operation

[0767] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. The vehicle's autonomous driving system reduces speed and controls braking in accordance with the received commands. This allows the user to drive safely and with peace of mind. Specifically, it reduces speed to 20 km / h and maintains smooth driving. The output of this process is a safe and relaxed driving state.

[0768] (Application example 1)

[0769] 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."

[0770] Conventional automated driving systems have been unable to adjust driving behavior to take into account the psychological state of the driver and passengers, making it difficult to provide a sense of security. Furthermore, there was no mechanism to provide driving advice or suggestions based on the psychological state in real time, so the effectiveness of reducing driver stress and promoting safe driving was limited.

[0771] 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.

[0772] In this invention, the server includes, in a vehicle linked to an automated driving system that performs driving operations, means for collecting a user's facial expressions in real time using an internally installed camera, means for collecting the user's voice in real time using an internally installed microphone, means for preprocessing the collected facial expression data and voice data and extracting feature points and feature quantities, means for inferring the user's psychological state from the extracted feature data using a generative AI model, means for adjusting driving parameters based on the inferred psychological state, means for notifying the user's mobile device of the generated suggestions and providing audio guidance or display, and means for transmitting the adjusted driving parameters to the automated driving system and applying them. This enables driving adjustments and the provision of appropriate advice based on the psychological states of the driver and passengers.

[0773] An "autonomous driving system that performs driving operations" is a system that automatically controls the operation of a vehicle using data input devices such as sensors and cameras.

[0774] A "camera" is an optical device for capturing images and video.

[0775] "User's facial expression" refers to the emotions and psychological state of the user that are indicated by the movements of the user's face and changes in the facial muscles.

[0776] A "microphone" is a sound capture device that converts sound waves into electrical signals.

[0777] "User voice" refers to the voice or sound patterns produced by a user.

[0778] "Facial expression data" refers to a collection of images and feature points of a user's face captured by a camera.

[0779] "Voice data" refers to the waveform information of the user's voice or sound collected by a microphone.

[0780] "Feature points" refer to important points on the face, including the position information of the eyes, nose, mouth, etc.

[0781] "Features" refers to information in audio data that is expressed by specific parameters (e.g., pitch, volume, spectral characteristics).

[0782] A "generative AI model" is an artificial intelligence algorithm used to infer a user's psychological state from collected data.

[0783] "Mental state" refers to the emotions and psychological state that a user is feeling.

[0784] "Driving parameters" refers to adjustable driving-related settings such as vehicle speed, acceleration, braking, etc.

[0785] "Mobile device" refers to a portable computing device such as a smartphone or tablet.

[0786] An "audio guide" is a system that provides users with information and instructions through audio.

[0787] "Notification" refers to an alert or message that notifies a user of specific information.

[0788] "Preprocessing" refers to the step of converting collected data into a form suitable for subsequent analysis and processing.

[0789] "Adjustment" refers to the act of changing driving parameters based on the estimated psychological state.

[0790] "Transmit" refers to the act of transferring data or instructions to another system or device.

[0791] "Application" refers to reflecting the adjusted driving parameters in the actual vehicle operation.

[0792]

[0793] The present invention is an autonomous driving system that uses a camera and microphone inside the vehicle to collect a user's facial expressions and voice in real time, and then uses a generative AI model to infer the user's psychological state based on the collected data. The system of the present invention includes a camera, a microphone, a server, a mobile device such as a smartphone, and a generative AI model.

[0794] System configuration

[0795] 1. Data Collection

[0796] The smartphone's camera and microphone are used to capture the driver's facial expressions and voice in real time. The camera captures facial expressions frame by frame, and the microphone records the voice.

[0797] 2. Data Preprocessing

[0798] The device preprocesses the collected image data using OpenCV to extract facial features, and for audio data, Librosa is used to extract features (pitch, volume, spectral characteristics).

[0799] 3. Psychological analysis

[0800] The preprocessed data is sent to a cloud server, where a generative AI model is used to analyze the user's psychological state. The server achieves this using platforms such as Google Cloud AI and AWS SageMaker.

[0801] 4. Adjusting the operating parameters

[0802] The server adjusts driving parameters based on the inferred psychological state, for example generating instructions to reduce speed and avoid sudden maneuvers if the user is nervous.

[0803] 5. Proposal Generation

[0804] The server generates advice based on the user's psychological state, suggests music, and suggests rest areas. These suggestions are sent to the mobile device and provided to the user as audio guidance or on-screen displays.

[0805] 6. Sending adjustment commands and executing operations

[0806] The server then sends the adjusted driving parameters to the vehicle's autonomous driving system and applies them, enabling driving that is appropriate for the driver's psychological state.

[0807] Specific examples

[0808] Scenario 1: Tired driver

[0809] 1. Data collection: The smartphone camera captures the driver's tired expression, and the microphone collects the driver's sigh.

[0810] 2. Data preprocessing: Extract the degree of eye contraction using OpenCV and low-pitched audio using Librosa.

[0811] 3. Psychological state analysis: The generative AI model infers that the driver is tired.

[0812] 4. Proposal generation: The server notifies the mobile device of break suggestions and soothing music playback.

[0813] 5. Sending adjustment instructions and executing driving: The server adjusts the driving parameters, and the vehicle follows the instructions to drive appropriately.

[0814] Prompt Sentence Examples

[0815] Prompt: "Analyze the driver's facial expressions and voice data to infer the driver's fatigue state and generate appropriate rest advice."

[0816] With the above configuration, this system is able to adjust driving and provide appropriate advice according to the psychological state of the driver and passengers, enabling a more secure autonomous driving experience.

[0817] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0818] Step 1:

[0819] The device captures the user's face using a camera installed in the vehicle and collects video data frame by frame. The specific input is image frames captured in real time, and the output is image data frame by frame, which is used for subsequent processing to extract feature points.

[0820] Step 2:

[0821] The terminal collects the user's voice in real time using a microphone installed in the vehicle. The specific input is the user's spoken voice, and the output is voice data. The collected voice data is used to extract features in the next step.

[0822] Step 3:

[0823] The device preprocesses the collected image data and extracts facial feature points. Specifically, it uses OpenCV to detect feature points such as eyes, mouth, and eyebrows from each frame of image. The input is the image data collected in step 1, and the output is the extracted facial feature point data.

[0824] Step 4:

[0825] The device preprocesses the collected audio data and extracts features. Specifically, Librosa is used to extract pitch, volume, and spectral characteristics from the audio data. The input is the audio data collected in step 2, and the output is the extracted audio feature data.

[0826] Step 5:

[0827] The terminal sends the preprocessed feature point data and feature volume data to the server. The specific input is the feature data obtained in step 3 and step 4, and the output is the data sent to the server.

[0828] Step 6:

[0829] The server inputs the received feature data into a generative AI model to infer the user's psychological state. The specific inputs are feature point data and voice feature data, and the output is the inferred psychological state (e.g., relaxed, tense, fatigued). The generative AI model achieves this using Google Cloud AI and AWS SageMaker.

[0830] Step 7:

[0831] The server generates instructions to adjust driving parameters based on the inferred psychological state, e.g., to reduce speed if the user is tired. The input is the psychological state inferred in step 6, and the output is the adjusted driving parameters.

[0832] Step 8:

[0833] The server sends the generated driving parameters or advice to the mobile device. For example, a break suggestion or a music playback instruction is generated in this step. The input is the driving parameters generated in step 7, and the output is the notification content.

[0834] Step 9:

[0835] The device displays or voices the received notification to the driver. Specifically, it displays the suggestion on the screen of the mobile device or informs the user as voice guidance. The input is the notification content sent in step 8, and the output is the display on the device or voice guidance.

[0836] Step 10:

[0837] The server sends the adjusted driving parameters to the automated driving system, and the vehicle performs driving based on these instructions. The input is the driving parameters generated in step 7, and the output is the actual vehicle behavior.

[0838] Example prompt sentence:

[0839] Prompt: "Analyze the driver's facial expressions and voice data to infer the driver's fatigue state and generate appropriate rest advice."

[0840] The above are the processing steps in the embodiment of the present invention.

[0841] 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.

[0842] System Overview

[0843] This invention is an autonomous driving system that uses a camera and microphone installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model and an emotion engine to infer the user's psychological state and adjust driving parameters. By combining this with an emotion engine, it becomes possible to more accurately recognize the user's diverse emotions and perform driving operations based on them.

[0844] Explanation of program processing

[0845] Data collection

[0846] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. A microphone also records what the user says in real time.

[0847] Data Preprocessing

[0848] Device: Extracts facial features (e.g., the position of the eyes, mouth, and eyebrows) from the collected image data, and extracts features (e.g., pitch, volume, and spectral characteristics) from the audio data, converting the data into a format that is easy for the AI ​​model described below to handle.

[0849] Sending data

[0850] Terminal: Sends preprocessed data (facial feature points and audio features) to the server.

[0851] Psychological analysis

[0852] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to infer the user's mental state and emotions, for example, identifying emotions such as tension, relief, surprise, anger, and joy.

[0853] Adjusting operating parameters

[0854] Server: Adjusts driving parameters (speed, acceleration, braking) based on the inferred psychological state and emotions. For example, if the user is angry, the server will slow down the vehicle to ensure safety and drive calmly.

[0855] Sending adjustment instructions

[0856] Server: Sends the adjusted driving parameters to the automated driving system. This transmission is usually done in real time, and is quickly reflected in the automated driving system inside the vehicle.

[0857] Execution of operation

[0858] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. This allows for safe and secure driving that is tailored to the user's psychological state.

[0859] Specific examples

[0860] Scenario 1: The user is angry

[0861] Data collection:

[0862] Device: The camera captures the user's angry facial expression (e.g., frown, closed mouth), and the microphone records the user's strong-tongued remarks.

[0863] Data preprocessing:

[0864] Device: Extract anger features from image data using the Dlib library, and extract high-pitch and high-volume features from audio data.

[0865] Psychological Analysis:

[0866] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is angry."

[0867] Adjusting operating parameters:

[0868] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[0869] Sending adjustment orders:

[0870] Server: Sends new driving parameters to the vehicle.

[0871] Execution of the operation:

[0872] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[0873] Scenario 2: Happy users

[0874] Data collection:

[0875] Device: The camera captures the user's smile, and the microphone records the user's conversation in a cheerful tone.

[0876] Data preprocessing:

[0877] Device: From the facial features, smiling features (such as the corners of the mouth turning up and the corners of the eyes turning up) are extracted, and stable pitch and volume features are extracted from the audio data.

[0878] Psychological Analysis:

[0879] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is happy."

[0880] Adjusting operating parameters:

[0881] Server: Maintain normal operating parameters or configure for slightly faster speeds.

[0882] Sending adjustment orders:

[0883] Server: Sends a command to the vehicle to continue normal operation.

[0884] Execution of the operation:

[0885] Terminal: The vehicle follows the command and maintains its current operating parameters.

[0886] The above is a specific embodiment of the automated driving system of the present invention. This allows driving to be performed in accordance with the diverse emotions and psychological states of passengers, thereby improving safety and peace of mind.

[0887] The processing flow will be explained below.

[0888] Specific processing steps of the system

[0889] Step 1:

[0890] Data collection

[0891] Terminal: A camera installed inside the vehicle captures the user's facial expressions frame by frame and collects image data. A microphone also records the user's speech in real time.

[0892] How it works: The camera takes pictures at 30 frames per second and stores them in storage. The microphone captures audio at high resolution of 24-bit, 96kHz and stores it in storage.

[0893] Step 2:

[0894] Data Preprocessing

[0895] Terminal: Extracts facial features (e.g., eye, mouth, and eyebrow positions) from collected image data, and extracts features (e.g., pitch, volume, and spectral characteristics) from audio data.

[0896] Specific operation: The Dlib library is used to identify the positions of the eyes, mouth, and eyebrows from the image data, and 68 feature points are extracted. MFCC (Mel-Frequency Cepstrum Coefficients) is used to extract 13-dimensional feature values ​​from the audio data every 0.1 seconds.

[0897] Step 3:

[0898] Sending data

[0899] Terminal: Sends preprocessed data (facial feature points and audio features) to the server.

[0900] Specific operation: Data is sent asynchronously to a server via the Internet or a dedicated in-vehicle network using HTTP or WebSocket protocols.

[0901] Step 4:

[0902] emotion recognition

[0903] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to recognize the user's emotions. The generative AI model estimates the psychological state, and the emotion engine identifies the emotion.

[0904] Specific behavior: The generative AI model analyzes the pre-processed data and outputs psychological states such as "tension," "relief," and "surprise." Based on this, the emotion engine classifies the user's specific emotions into "anger," "joy," "sadness," etc.

[0905] Step 5:

[0906] Adjusting operating parameters

[0907] Server: Determines driving parameters (speed, acceleration, braking) based on the inferred psychological state and emotions.

[0908] Specific behavior: Uses a rules engine to generate driving commands such as "slow down if the user is angry" or "maintain speed if the user is happy."

[0909] Step 6:

[0910] Sending adjustment instructions

[0911] Server: Sends the generated new driving parameters to the autonomous driving system.

[0912] Specific operation: Driving parameters are sent to the terminal in the vehicle using HTTP or a dedicated protocol.

[0913] Step 7:

[0914] Execution of operation

[0915] Terminal: Apply the new driving parameters received from the server to the autonomous driving system.

[0916] Specific behavior: The autonomous driving system uses the new control parameters to perform driving maneuvers such as speed, acceleration, and braking. For example, if the user is nervous, the system will reduce the speed to 20 km / h.

[0917] Specific examples

[0918] Scenario 1: The user is angry

[0919] Step 1:

[0920] Data collection

[0921] Device: The camera captures the user's angry facial expression (frown, tight-lipped mouth), and the microphone records the user's strong-tongued remarks.

[0922] Step 2:

[0923] Data Preprocessing

[0924] Terminal: Use the Dlib library to extract facial features, and high-pitch and high-volume features from audio data.

[0925] Step 3:

[0926] Sending data

[0927] Terminal: Sends preprocessed data to the server.

[0928] Step 4:

[0929] emotion recognition

[0930] Server: The generative AI model and emotion engine analyze the data and recognize that the user is feeling "anger."

[0931] Step 5:

[0932] Adjusting operating parameters

[0933] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[0934] Step 6:

[0935] Sending adjustment instructions

[0936] Server: Sends new driving parameters to the vehicle.

[0937] Step 7:

[0938] Execution of operation

[0939] Terminal: Reduce speed to 20km / h and begin driving smoothly.

[0940] Scenario 2: Happy users

[0941] Step 1:

[0942] Data collection

[0943] Device: The camera captures the user's smile. The microphone records the user's pleasant conversation.

[0944] Step 2:

[0945] Data Preprocessing

[0946] Device: Extracts features of a smile, such as eyes and mouth turned up, and extracts stable pitch and volume features from audio data.

[0947] Step 3:

[0948] Sending data

[0949] Terminal: Sends preprocessed data to the server.

[0950] Step 4:

[0951] emotion recognition

[0952] Server: The generative AI model and emotion engine analyze the data and recognize that the user is feeling "joy."

[0953] Step 5:

[0954] Adjusting operating parameters

[0955] Server: Issues a command to continue normal operation.

[0956] Step 6:

[0957] Sending adjustment instructions

[0958] Server: Sends commands to the vehicle.

[0959] Step 7:

[0960] Execution of operation

[0961] Terminal: Maintain current operating parameters and continue normal operation.

[0962] This will enable safe and secure driving based on specific emotions.

[0963] Example 2

[0964] 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."

[0965] In an automated driving system for a vehicle, there is a demand for providing a safer and more comfortable driving experience by accurately grasping the user's psychological state and adjusting driving parameters accordingly. However, with conventional technologies, it has been difficult to accurately recognize the user's emotions and psychological state in real time and dynamically adjust driving parameters based on that. This invention is intended to solve this problem.

[0966] 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.

[0967] In this invention, the server, in a vehicle linked to an automated driving system that estimates a user's psychological state and adjusts driving parameters, includes: means for collecting a user's facial expression in real time using an internally installed image capture device; means for collecting the user's voice in real time using an internally installed voice recording device; means for preprocessing the collected facial expression data and voice data and extracting feature points and feature quantities; means for estimating the user's psychological state from the extracted feature data using a generative technology model; means for adjusting the driving parameters based on the estimated psychological state; and means for transmitting the adjusted driving parameters to the automated driving system and applying them, thereby enabling appropriate adjustment of driving parameters based on the user's real-time psychological state.

[0968] A "user" is a person who uses this system.

[0969] "Mental state" refers to the user's feelings and mental state, and specifically includes emotions such as tension, relief, surprise, anger, and joy.

[0970] "Driving parameters" are control data related to vehicle driving, specifically settings such as speed, acceleration, and braking.

[0971] An "autonomous driving system" is a system that automatically controls the driving of a vehicle and has the function of allowing the vehicle to drive autonomously without the user performing any driving operations.

[0972] The "photography device" is a device that is installed inside the vehicle and captures the user's facial expressions in real time.

[0973] The "voice recording device" is a device that is installed inside the vehicle and records the user's speech in real time.

[0974] "Preprocessing" refers to the process of extracting feature points and feature quantities from collected raw data (image data and audio data).

[0975] "Feature points" refer to characteristic points on the user's face extracted from image data, such as the positions of the eyes, mouth, and eyebrows.

[0976] "Features" are acoustic features extracted from audio data, specifically pitch, volume, spectral characteristics, etc.

[0977] A "generative model" is an artificial intelligence model used to infer a user's psychological state from collected data.

[0978] "Adjusted driving parameters" refer to new driving parameters that have been adjusted based on the user's psychological state.

[0979] "Application" means reflecting the adjusted driving parameters in the actual automated driving system.

[0980] System Overview

[0981] This invention is an autonomous driving system that uses a camera and microphone installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model and emotion engine to infer the user's psychological state and adjust driving parameters. By combining this with an emotion engine, it is possible to more accurately recognize the user's diverse emotions and perform driving operations based on them.

[0982] Data collection

[0983] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. An audio recording device also records the user's speech in real time.

[0984] Hardware used:

[0985] Camera (e.g. high-resolution camera)

[0986] Microphone (e.g. high-sensitivity microphone)

[0987] Data Preprocessing

[0988] Device: Facial feature points (e.g., eye, mouth, and eyebrow positions) are extracted from the collected image data using a face recognition library (e.g., Dlib), while acoustic features (e.g., pitch, volume, and spectral characteristics) are extracted from the audio data.

[0989] Software used:

[0990] Dlib library (for face recognition)

[0991] Audio analysis tools (e.g. FFT)

[0992] Sending data

[0993] Terminal: The facial feature point data and voice feature data obtained in the preprocessing are encrypted and sent to the server using the HTTPS protocol to ensure data security.

[0994] Psychological analysis

[0995] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to infer the user's psychological state. For example, identify emotions such as tension, relief, surprise, anger, and joy.

[0996] Software used:

[0997] Generative technology models (e.g., BERT-based sentiment analysis models)

[0998] Emotion Engine

[0999] Adjusting operating parameters

[1000] Server: Optimizes driving parameters (speed, acceleration, braking) based on the estimated psychological state. If the user is nervous, the server reduces speed and smooths braking to make the driver more calm.

[1001] Sending adjustment instructions

[1002] Server: Transmits the calculated driving parameters to the vehicle's autonomous driving system in real time. This transmission also uses the HTTPS protocol to ensure data integrity and security.

[1003] Execution of operation

[1004] Terminal: The terminal applies the new driving parameters received from the server to the actual vehicle control and performs driving, thereby automatically driving according to the user's psychological state.

[1005] Specific examples

[1006] Scenario 1: The user is angry

[1007] Data collection:

[1008] Device: The camera captures the user's angry facial expression (e.g., frown, closed mouth), and the microphone records the user's strong-tongued remarks.

[1009] Data preprocessing:

[1010] Device: Extract anger features from image data using the Dlib library, and extract high-pitch and high-volume features from audio data.

[1011] Psychological Analysis:

[1012] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is angry."

[1013] Adjusting operating parameters:

[1014] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[1015] Sending adjustment orders:

[1016] Server: Sends new driving parameters to the vehicle.

[1017] Execution of the operation:

[1018] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[1019] Scenario 2: Happy users

[1020] Data collection:

[1021] Device: The camera captures the user's smile, and the microphone records the user's conversation in a cheerful tone.

[1022] Data preprocessing:

[1023] Device: From the facial features, smiling features (such as the corners of the mouth turning up and the corners of the eyes turning up) are extracted, and stable pitch and volume features are extracted from the audio data.

[1024] Psychological Analysis:

[1025] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is happy."

[1026] Adjusting operating parameters:

[1027] Server: Maintain normal operating parameters or configure for slightly faster speeds.

[1028] Sending adjustment orders:

[1029] Server: Sends a command to the vehicle to continue normal operation.

[1030] Execution of the operation:

[1031] Terminal: The vehicle follows the command and maintains its current operating parameters.

[1032] Prompt Sentence Examples

[1033] Scenario 1 prompt statement:

[1034] If the user is angry, please follow the instructions below to adjust the operating parameters:

[1035] Angry facial expression captured on camera

[1036] Logged voice data with strong verbal content

[1037] Scenario 2 prompt statement:

[1038] If the user is willing, please follow the instructions below to adjust the operating parameters:

[1039] A photo of a smiling face captured on camera

[1040] Logged voice data with a pleasant tone

[1041] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1042] Step 1: Data collection

[1043] Device: The in-car camera captures the user's face frame by frame and collects image data, while the microphone records what the user says in real time.

[1044] Input: Video frames of the user's face, audio of what the user is saying

[1045] Output: User's face image data, voice data

[1046] Specific behavior:

[1047] The interior camera captures images at 30 frames per second.

[1048] The microphone records audio at a sample rate of 16kHz.

[1049] Step 2: Data Preprocessing

[1050] Device: Facial feature points (e.g., eye, mouth, and eyebrow positions) are extracted from the collected image data using a face recognition library (e.g., Dlib), while acoustic features (e.g., pitch, volume, and spectral characteristics) are extracted from the audio data.

[1051] Input: Facial image data, audio data

[1052] Output: Facial feature data, acoustic feature data

[1053] Specific behavior:

[1054] Using the Dlib library, 68 facial feature points are extracted from each frame.

[1055] The frequency components of the audio data are analyzed using FFT (Fast Fourier Transform) and extracted as features.

[1056] Step 3: Sending data

[1057] Terminal: The facial feature point data and voice feature data obtained in the preprocessing are encrypted and sent to the server using the HTTPS protocol to ensure data security.

[1058] Input: facial feature point data, acoustic feature data

[1059] Output: Notification that encrypted data has been sent

[1060] Specific behavior:

[1061] Package the data in JSON format and send a POST request to the server.

[1062] Step 4: Psychological analysis

[1063] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to infer the user's psychological state. For example, identify emotions such as tension, relief, surprise, anger, and joy.

[1064] Input: Preprocessed data (facial feature points, acoustic features)

[1065] Output: User's mental state and emotion category

[1066] Specific behavior:

[1067] Input feature data into a generative AI model (e.g., a BERT-based sentiment analysis model) to predict sentiment categories.

[1068] The emotion engine analyzes the prediction results and determines the most appropriate emotion.

[1069] Step 5: Adjusting the operating parameters

[1070] Server: Optimizes driving parameters (speed, acceleration, braking) based on the estimated psychological state. If the user is nervous, the server reduces speed and smooths braking to make the driver more calm.

[1071] Input: User's mental state and emotion category

[1072] Output: Adjusted operating parameters

[1073] Specific behavior:

[1074] It obtains driving data such as current speed and acceleration and calculates new driving parameters based on that.

[1075] Step 6: Sending adjustment instructions

[1076] Server: Transmits the calculated driving parameters to the vehicle's autonomous driving system in real time. This transmission also uses the HTTPS protocol to ensure data integrity and security.

[1077] Input: Adjusted operating parameters

[1078] Output: Notification of completion of sending of operation parameters

[1079] Specific behavior:

[1080] The new driving parameters are packaged in JSON format and sent as a POST request to the vehicle's control system.

[1081] Step 7: Execute the operation

[1082] Terminal: The terminal applies the new driving parameters received from the server to the actual vehicle control and performs driving, thereby automatically driving according to the user's psychological state.

[1083] Input: New operating parameters

[1084] Output: Vehicle control actions based on driving parameters

[1085] Specific behavior:

[1086] The vehicle's ECU (Electronic Control Unit) receives the new driving parameters and reflects them in brake and accelerator control.

[1087] (Application example 2)

[1088] 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."

[1089] Conventional automated driving systems operate based on predetermined driving parameters without considering the user's emotions or psychological state, and therefore cannot fully guarantee the user's safety or comfort. Furthermore, they lack a mechanism for providing support content that adapts to the user's psychological state while driving, making it difficult to promote stress reduction and relaxation for the user. Therefore, there is a need for a system that can adjust driving parameters and provide support content appropriately according to the user's psychological state.

[1090] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting the user's facial expressions in real time using an internally installed camera, means for collecting the user's voice in real time using an internally installed microphone, means for preprocessing the collected facial expression data and voice data to extract feature points and feature quantities, means for inferring the user's psychological state from the extracted feature data using a generative AI model, means for adjusting driving parameters based on the inferred psychological state, means for transmitting the adjusted driving parameters to the automated driving system and applying them, and means for providing support content during driving in accordance with the user's emotional state. This makes it possible to adjust driving parameters in accordance with the user's psychological state and provide appropriate support content.

[1091] An "autonomous driving system that performs driving operations" is a system that automatically drives a vehicle, and is a general term for the programs and hardware that perform driving operations such as adjusting speed, changing direction, and braking.

[1092] The "interiorly installed camera" is a camera device that is attached to the interior of the vehicle and captures the user's facial expressions in real time.

[1093] The "interiorly installed microphone" is a microphone device that is installed inside the vehicle and collects the user's voice in real time.

[1094] "Facial expression data" is image data of the user's face captured by a camera, and includes facial feature points.

[1095] "Voice data" refers to voice data of a user's speech recorded by a microphone, and includes voice features.

[1096] "Preprocessing" refers to a data processing procedure for extracting feature points and feature quantities from collected data.

[1097] "Feature points" are data points relating to the position and shape of a face, and indicate specific positions such as the eyes, mouth, and eyebrows.

[1098] "Features" are specific numerical data such as pitch, volume, and spectral characteristics extracted from audio data.

[1099] A "generative AI model" is an artificial intelligence model that analyzes collected data and infers the user's psychological state.

[1100] The "emotion engine" is an analytical engine that identifies the user's emotions and generates appropriate responses based on them.

[1101] "Driving parameters" are set values ​​for controlling the driving of a vehicle, and include speed, acceleration, braking, and the like.

[1102] "Support content" refers to supplementary information and entertainment provided while driving according to the user's psychological state, and includes relaxing music, positive news, and the like.

[1103] This invention is an autonomous driving system that uses a camera and microphone installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model and emotion engine to infer the user's psychological state and adjust driving parameters.Furthermore, it is a system that provides support content according to the user's psychological state.

[1104] System Overview

[1105] Hardware used

[1106] Camera: An image capturing device installed inside the vehicle.

[1107] Microphone: An audio collection device installed inside the vehicle.

[1108] Smart glasses: A wearable device for capturing the user's facial expressions and voice in real time.

[1109] Minicomputer: A small computer (e.g., Raspberry Pi) for data preprocessing and analysis.

[1110] On-board computer: An on-board computer used to adjust the vehicle's operating parameters.

[1111] Software used

[1112] Facial expression recognition library: Dlib (a library for extracting facial feature points).

[1113] Speech analysis library: LibROSA (a library for extracting speech features).

[1114] Generative AI models: Natural language processing models such as GPT-3.

[1115] Sentiment engine: A sentiment analysis engine such as Azure Cognitive Services.

[1116] Specific Embodiments

[1117] Data collection

[1118] The device (smart glasses) captures the user's facial expressions in real time with a camera installed inside the vehicle and records the user's voice with a microphone, and these data are sent to a minicomputer.

[1119] Data Preprocessing

[1120] The minicomputer uses Dlib to extract facial features from image data, and LibROSA to extract audio features such as pitch, volume, and spectral characteristics from audio data.

[1121] Psychological analysis

[1122] The server inputs the preprocessed data into a generative AI model and emotion engine to infer the user's psychological state, and determines whether the user is happy, angry, or other emotional state based on the inference.

[1123] Providing support content

[1124] If the user is inferred to be angry, the following prompt sentence can be input into the generative AI model to generate instructions to provide relaxing music:

[1125] "User's facial expression data: {facial expression data}

[1126] User's voice data: {voice data}

[1127] Inferred emotion: Anger

[1128] Suggested action: Play relaxing music

[1129] Based on this instruction, the appropriate music will be played.

[1130] If the user is deemed happy, the generative AI model is given the following prompt:

[1131] "User's facial expression data: {facial expression data}

[1132] User's voice data: {voice data}

[1133] Inferred emotion: Joy

[1134] Recommended action: Read positive news

[1135] This will read out positive news aloud.

[1136] Adjusting operating parameters

[1137] The server adjusts driving parameters (speed, braking, etc.) based on the inferred psychological state. For example, if the user is angry, it generates an instruction to slow down the vehicle and sends it to the on-board computer. This automatically adjusts the vehicle's driving, increasing the user's safety.

[1138] Example

[1139] Scenario 1: The user is angry

[1140] Data collection:

[1141] The camera captures the user's angry expression (frowns, closed mouth), and the microphone records the strong words spoken.

[1142] Sentiment Analysis and Response:

[1143] The generative AI model and emotion engine infer that the user is angry and generate instructions to provide relaxing music.

[1144] Adjusting operating parameters:

[1145] The server generates and transmits to the vehicle instructions to reduce the vehicle's speed.

[1146] Scenario 2: Happy users

[1147] Data collection:

[1148] The camera captures the user's smile, and the microphone records conversations in a playful tone.

[1149] Sentiment Analysis and Response:

[1150] The generative AI model and emotion engine infer that the user is happy and generate instructions to deliver positive news.

[1151] Adjusting operating parameters:

[1152] The server generates and transmits instructions to the vehicle to maintain normal operating parameters.

[1153] This will enable the adjustment of driving parameters according to the user's psychological state and the provision of appropriate support content, resulting in safer and more comfortable automated driving.

[1154] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1155] Step 1:

[1156] The device (smart glasses) captures images of the user's face in real time using an internal camera and acquires image data. It also records the user's voice in real time using an internal microphone and acquires voice data. This input data (image data and voice data) is sent to a minicomputer.

[1157] Step 2:

[1158] The minicomputer extracts facial features (e.g., eye, mouth, and eyebrow positions) from the received image data using the Dlib library, and simultaneously extracts pitch, volume, and spectral characteristics from the audio data using the LibROSA library, resulting in preprocessed data (facial features and audio features).

[1159] Step 3:

[1160] The preprocessed data is sent to the server. The input data (facial features and voice features) is fed into the generative AI model and emotion engine to infer the user's psychological state. The generative AI model analyzes the combined facial and voice data and identifies emotions such as "the user is angry" or "the user is happy."

[1161] Step 4:

[1162] The server adjusts driving parameters based on the estimated psychological state. Specifically, it generates instructions to set driving parameters such as vehicle speed, acceleration, and braking. For example, if it estimates that the user is angry, it generates an instruction to slow down the vehicle.

[1163] Step 5:

[1164] The server sends the generated instructions to the vehicle's on-board computer. The input data (adjusted driving parameters) are applied by the on-board computer and reflected in actual driving operations. This allows driving to be performed in accordance with the user's psychological state.

[1165] Step 6:

[1166] The server provides support content according to the user's psychological state. For example, if the user is angry, the server generates an instruction to play relaxing music and sends it to the minicomputer. The minicomputer then plays the relaxing music according to the instruction.

[1167] Step 7:

[1168] If the user's psychological state is calmed by the support content they receive, the camera and microphone will again collect their facial expressions and voice, and the entire system will start operating again, making it possible to provide continuous driving support that responds to the user's psychological state.

[1169] As a result, safe and comfortable automated driving is achieved by adjusting driving parameters according to the user's psychological state and providing support content.

[1170] 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.

[1171] 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.

[1172] 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.

[1173] [Third embodiment]

[1174] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1175] 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.

[1176] 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).

[1177] 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.

[1178] 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.

[1179] 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).

[1180] 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.

[1181] 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.

[1182] 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.

[1183] 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.

[1184] 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.

[1185] 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."

[1186] System Overview

[1187] This invention is an autonomous driving system that uses cameras and microphones installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model to infer the user's psychological state.The system aims to provide passengers with a sense of security by adjusting driving parameters based on the inferred psychological state.

[1188] Explanation of program processing

[1189] Data collection

[1190] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. A microphone also records what the user says in real time.

[1191] Data Preprocessing

[1192] Device: Extracts facial features (e.g., the position of the eyes, mouth, and eyebrows) from the collected image data, and extracts features (e.g., pitch, volume, and spectral characteristics) from the audio data, converting the data into a format that is easy for the AI ​​model described below to handle.

[1193] Psychological analysis

[1194] Server: The preprocessed data is sent to the server, which uses a generative AI model to infer the user's psychological state from this data, for example, identifying emotions such as tension, relief, and surprise.

[1195] Adjusting operating parameters

[1196] Server: Generates commands to adjust driving parameters (e.g., speed, acceleration, braking) based on the estimated psychological state. For example, if the user is nervous, the server may slow down the speed or prohibit sudden acceleration.

[1197] Sending adjustment instructions

[1198] Server: Sends the adjusted driving parameters to the automated driving system. This transmission is usually done in real time, and is quickly reflected in the automated driving system inside the vehicle.

[1199] Execution of operation

[1200] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. This allows for safe and secure driving that is tailored to the user's psychological state.

[1201] Specific examples

[1202] Scenario 1: The user is surprised

[1203] Data collection:

[1204] Device: The camera captures the user's surprised expression, and the microphone records the user's high-pitched utterance, such as "I'm scared."

[1205] Data preprocessing:

[1206] Device: Extracts facial features such as wide-open eyes and an open mouth, and extracts high-pitched and high-volume features from audio data.

[1207] Psychological Analysis:

[1208] Server: The generative AI model analyzes these feature data and infers that the user is surprised.

[1209] Adjusting operating parameters:

[1210] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[1211] Sending adjustment orders:

[1212] Server: Sends new driving parameters to the vehicle.

[1213] Execution of the operation:

[1214] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[1215] Scenario 2: User is relaxed

[1216] Data collection:

[1217] Device: The camera captures the user's relaxed facial expression, and the microphone records calm conversations.

[1218] Data preprocessing:

[1219] Device: From facial features, relaxed eye and mouth shapes are extracted, and stable pitch and volume features are extracted from audio data.

[1220] Psychological Analysis:

[1221] Server: The generative AI model analyzes these feature data and infers that the user is relaxed.

[1222] Adjusting operating parameters:

[1223] Server: Maintain normal operating parameters or configure for slightly faster speeds.

[1224] Sending adjustment orders:

[1225] Server: Sends a command to the vehicle to continue normal operation.

[1226] Execution of the operation:

[1227] Terminal: The vehicle follows your commands and you can continue driving relaxed.

[1228] The above is a specific embodiment of the automated driving system of the present invention, which enables driving according to the psychological state of the passengers, providing a greater sense of security.

[1229] The processing flow will be explained below.

[1230] Step 1:

[1231] Data collection

[1232] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. A microphone also records what the user says in real time.

[1233] How it works: The camera takes pictures at 30 frames per second and stores them in storage. The microphone captures audio at high resolution of 24-bit, 96kHz and stores it in storage.

[1234] Step 2:

[1235] Data Preprocessing

[1236] Terminal: Extracts facial features from collected image data, and extracts features such as pitch, volume, and spectral characteristics from audio data.

[1237] Specific operation: The Dlib library is used to identify the positions of the eyes, mouth, and eyebrows from the image data, and 68 feature points are extracted. MFCC (Mel-Frequency Cepstrum Coefficients) is used to extract 13-dimensional feature values ​​from the audio data every 0.1 seconds.

[1238] Step 3:

[1239] Sending data

[1240] Terminal: Sends preprocessed data (facial feature points and audio features) to the server.

[1241] Specific operation: Data is sent asynchronously to a server using HTTP or WebSocket protocol over the Internet or a dedicated in-vehicle network.

[1242] Step 4:

[1243] Psychological analysis

[1244] Server: Inputs the received preprocessed data into the generative AI model and infers the user's psychological state.

[1245] Specific operation: The received data is input into a generative AI model (e.g., BERT or CNN), which outputs an emotion label, such as whether the user is nervous, relaxed, or surprised.

[1246] Step 5:

[1247] Adjusting operating parameters

[1248] Server: Determines driving parameters (speed, acceleration, braking) based on the estimated psychological state.

[1249] Specific behavior: The rule engine is executed to generate new driving parameters, such as slowing down if the psychological state is "surprised" and increasing speed if the psychological state is "relaxed."

[1250] Step 6:

[1251] Sending adjustment instructions

[1252] Server: Sends new driving parameters to the vehicle's autonomous driving system.

[1253] Specific operation: Sends command data to the terminal using HTTP or a dedicated protocol.

[1254] Step 7:

[1255] Execution of operation

[1256] Terminal: Follows instructions from the server and applies new driving parameters to the autonomous driving system.

[1257] Specific operation: The autonomous driving system applies settings such as new speed limits and prohibitions on sudden acceleration to the controller and performs actual driving operations.

[1258] The above is the specific process and operation flow for each step. This system realizes safe and secure driving that is in line with the passenger's psychological state.

[1259] Example 1

[1260] 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."

[1261] In current automated driving systems, vehicle operating parameters are adjusted mainly based on the external environment and traffic conditions, and the psychological state of passengers is not taken into consideration. This often causes passengers to feel nervous or anxious, which can hinder a comfortable journey. In particular, sudden acceleration or braking can increase the psychological burden on passengers.

[1262] 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.

[1263] In this invention, the server includes means for extracting facial feature points from collected facial expression data, means for extracting pitch, volume, and spectral characteristics from voice data, and means for inferring the user's psychological state from the feature data using a generative AI model. This enables adjustment of driving parameters based on the user's psychological state, improving passenger comfort.

[1264] The "camera" is a photographing device that is installed inside the vehicle and collects the user's facial expressions frame by frame in real time.

[1265] A "microphone" is a recording device installed inside a vehicle to collect the user's voice in real time.

[1266] "Facial expression data" is image information of the user's face collected by a camera.

[1267] "Voice data" refers to information about a user's speech or voice collected by a microphone.

[1268] "Feature points" are the main positional information of the face (e.g., the positions of the eyes, mouth, and eyebrows) extracted from the facial expression data.

[1269] "Features" are sound characteristics (e.g., pitch, volume, spectral characteristics) extracted from audio data.

[1270] The "server" is a central processing unit for processing the collected facial expression data and voice data.

[1271] A "generative AI model" is an artificial intelligence model that takes feature data as input and uses it to infer a user's psychological state.

[1272] "Driving parameters" are set values ​​related to driving operations such as vehicle speed, acceleration, and braking.

[1273] The "command" is instruction information generated by the server that includes the details of adjustment of the operating parameters.

[1274] "Real-time transmission" means that data is transmitted almost instantly to a server or to an automated driving system.

[1275] MODE FOR CARRYING OUT THE INVENTION

[1276] The present invention is an autonomous driving system that uses a camera and microphone inside the vehicle to collect a user's facial expressions and voice in real time and uses a generative AI model to infer their psychological state. The system aims to provide passengers with a sense of security by adjusting driving parameters based on the inferred psychological state. A specific embodiment of the system is described below.

[1277] Data collection

[1278] Terminal: A camera installed inside the vehicle captures the user's face frame by frame and collects image data. The camera used is a general-purpose high-resolution camera. The microphone records the user's voice in real time. Specifically, a general-purpose recording device with noise-canceling functionality is used.

[1279] Data Preprocessing

[1280] Device: The facial recognition software "dlib" is used to extract facial features (e.g., the position of the eyes, mouth, and eyebrows) from the collected image data. Also, "Librosa" is used to extract features (e.g., pitch, volume, and spectral characteristics) from the audio data. This process converts the data into a format that is easy for the AI ​​model described below to handle.

[1281] Sending data

[1282] Terminal: The preprocessed data is sent to the server. The "WebSocket" protocol is used for transmission to achieve real-time communication, which enables high-speed, low-latency data transfer.

[1283] Psychological analysis

[1284] Server: Receives the preprocessed data and uses a generative AI model to infer the user's psychological state. The AI ​​model used is a customized model based on GPT-4. This model receives image and audio feature data as input and classifies the user's psychological state (e.g., nervous, relaxed, surprised).

[1285] Adjusting operating parameters

[1286] Server: Generates commands to adjust driving parameters based on the estimated psychological state. These commands include adjustments to vehicle speed, acceleration, braking, etc. For example, if the user is surprised, it generates a command to slow down the vehicle and prohibit sudden maneuvers. This is done by inputting a prompt such as "Please suggest driving parameters when the user is surprised" into the generative AI model.

[1287] Sending adjustment instructions

[1288] Server: Sends the generated driving parameters to the autonomous driving system. This transmission is also done in real time and is immediately reflected in driving operations. The "MQTT" (Message Queuing Telemetry Transport) protocol is used for transmission, achieving highly reliable and low-latency communication.

[1289] Execution of operation

[1290] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. The vehicle's autonomous driving system reduces speed and controls braking based on the received commands, thereby realizing safe and secure driving for the user.

[1291] Specific examples

[1292] A specific scenario is shown below.

[1293] Scenario 1: The user is surprised

[1294] Data collection:

[1295] Device: The camera captures the user's surprised expression, and the microphone records the user's high-pitched utterances such as "I'm scared."

[1296] Data preprocessing:

[1297] Device: Extracts facial features such as wide-open eyes and an open mouth, and extracts high-pitch and high-volume features from audio data.

[1298] Sending data:

[1299] Terminal: Sends pre-processed data to the server in real time.

[1300] Psychological Analysis:

[1301] Server: The generative AI model analyzes the feature data and infers that the user is surprised.

[1302] Adjusting operating parameters:

[1303] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[1304] Sending adjustment orders:

[1305] Server: Sends new driving parameters to the vehicle.

[1306] Execution of the operation:

[1307] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[1308] Scenario 2: User is relaxed

[1309] Data collection:

[1310] Device: The camera captures the user's relaxed facial expression, and the microphone records calm conversations.

[1311] Data preprocessing:

[1312] Device: Extracts relaxed eye and mouth shapes from facial feature points, and extracts stable pitch and volume features from audio data.

[1313] Sending data:

[1314] Terminal: Sends preprocessed data to the server.

[1315] Psychological Analysis:

[1316] Server: The generative AI model analyzes the feature data and infers that the user is relaxed.

[1317] Adjusting operating parameters:

[1318] Server: Generates commands to maintain normal operating parameters or slightly increase speed.

[1319] Sending adjustment orders:

[1320] Server: Sends a command to the vehicle to continue normal operation.

[1321] Execution of the operation:

[1322] Terminal: The vehicle follows your commands and you can continue driving relaxed.

[1323] The above is a specific embodiment of the automated driving system of the present invention, which enables driving according to the psychological state of the passengers, providing a greater sense of security.

[1324] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1325] Program processing flow

[1326] Step 1: Data collection

[1327] Terminal: The camera captures the user's face frame by frame and collects image data. Specifically, it captures 30 frames per second. The microphone also records the user's voice in real time. Specifically, it uses a general-purpose, highly sensitive recording device and has the function of reducing background noise. This allows high-resolution image data and voice data to be obtained as input.

[1328] Step 2: Data Preprocessing

[1329] Terminal: The facial recognition software "dlib" is used to extract facial feature points (the positions of the eyes, mouth, and eyebrows) from the collected image data. At the same time, "Librosa" is used to extract feature quantities such as pitch, volume, and spectral characteristics from the audio data. Specifically, feature points are identified from the image of each frame, and the audio data is quantified. This results in the extracted feature points and feature quantities being obtained as output.

[1330] Step 3: Sending data

[1331] Terminal: The "WebSocket" protocol is used to send preprocessed data to the server. This allows feature points and feature quantities to be sent to the server at high speed and with low latency. Specifically, image feature points and audio feature quantities for each frame are packaged into packets and sent in real time.

[1332] Step 4: Psychological analysis

[1333] Server: Receives the preprocessed data and uses a generative AI model to infer the user's emotional state. The AI ​​model used is a customized version of GPT-4. Specifically, feature points and feature quantities are input into the model to classify the user's emotions (tension, relaxation, surprise, etc.). During this process, an inferred emotional state is output from the feature points and feature quantities received as input.

[1334] Step 5: Adjusting the operating parameters

[1335] Server: Generates commands to adjust driving parameters based on the estimated psychological state. Based on the output from the model, commands are generated to adjust speed, acceleration, braking, etc. For example, if it is estimated that the user is surprised, a command is generated to slow down the vehicle's speed and prohibit sudden maneuvers. This is done by inputting a prompt statement such as "Please suggest driving parameters for when the user is surprised" into the generative AI model. As a result, an adjustment command is generated.

[1336] Step 6: Sending adjustment instructions

[1337] Server: Sends the generated driving parameters to the autonomous driving system. This transmission is also done in real time and is immediately reflected in driving operations. The "MQTT" protocol is used for transmission. Specifically, a packet containing the generated driving parameters is created and sent to the autonomous driving system, which then outputs an adjustment command.

[1338] Step 7: Execute the operation

[1339] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. The vehicle's autonomous driving system reduces speed and controls braking in accordance with the received commands. This allows the user to drive safely and with peace of mind. Specifically, it reduces speed to 20 km / h and maintains smooth driving. The output of this process is a safe and relaxed driving state.

[1340] (Application example 1)

[1341] 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."

[1342] Conventional automated driving systems have been unable to adjust driving behavior to take into account the psychological state of the driver and passengers, making it difficult to provide a sense of security. Furthermore, there was no mechanism to provide driving advice or suggestions based on the psychological state in real time, so the effectiveness of reducing driver stress and promoting safe driving was limited.

[1343] 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.

[1344] In this invention, the server includes, in a vehicle linked to an automated driving system that performs driving operations, means for collecting a user's facial expressions in real time using an internally installed camera, means for collecting the user's voice in real time using an internally installed microphone, means for preprocessing the collected facial expression data and voice data and extracting feature points and feature quantities, means for inferring the user's psychological state from the extracted feature data using a generative AI model, means for adjusting driving parameters based on the inferred psychological state, means for notifying the user's mobile device of the generated suggestions and providing audio guidance or display, and means for transmitting the adjusted driving parameters to the automated driving system and applying them. This enables driving adjustments and the provision of appropriate advice based on the psychological states of the driver and passengers.

[1345] An "autonomous driving system that performs driving operations" is a system that automatically controls the operation of a vehicle using data input devices such as sensors and cameras.

[1346] A "camera" is an optical device for capturing images and video.

[1347] "User's facial expression" refers to the emotions and psychological state of the user that are indicated by the movements of the user's face and changes in the facial muscles.

[1348] A "microphone" is a sound capture device that converts sound waves into electrical signals.

[1349] "User voice" refers to the voice or sound patterns produced by a user.

[1350] "Facial expression data" refers to a collection of images and feature points of a user's face captured by a camera.

[1351] "Voice data" refers to the waveform information of the user's voice or sound collected by a microphone.

[1352] "Feature points" refer to important points on the face, including the position information of the eyes, nose, mouth, etc.

[1353] "Features" refers to information in audio data that is expressed by specific parameters (e.g., pitch, volume, spectral characteristics).

[1354] A "generative AI model" is an artificial intelligence algorithm used to infer a user's psychological state from collected data.

[1355] "Mental state" refers to the emotions and psychological state that a user is feeling.

[1356] "Driving parameters" refers to adjustable driving-related settings such as vehicle speed, acceleration, braking, etc.

[1357] "Mobile device" refers to a portable computing device such as a smartphone or tablet.

[1358] An "audio guide" is a system that provides users with information and instructions through audio.

[1359] "Notification" refers to an alert or message that notifies a user of specific information.

[1360] "Preprocessing" refers to the step of converting collected data into a form suitable for subsequent analysis and processing.

[1361] "Adjustment" refers to the act of changing driving parameters based on the estimated psychological state.

[1362] "Transmit" refers to the act of transferring data or instructions to another system or device.

[1363] "Application" refers to reflecting the adjusted driving parameters in the actual vehicle operation.

[1364]

[1365] The present invention is an autonomous driving system that uses a camera and microphone inside the vehicle to collect a user's facial expressions and voice in real time, and then uses a generative AI model to infer the user's psychological state based on the collected data. The system of the present invention includes a camera, a microphone, a server, a mobile device such as a smartphone, and a generative AI model.

[1366] System configuration

[1367] 1. Data Collection

[1368] The smartphone's camera and microphone are used to capture the driver's facial expressions and voice in real time. The camera captures facial expressions frame by frame, and the microphone records the voice.

[1369] 2. Data Preprocessing

[1370] The device preprocesses the collected image data using OpenCV to extract facial features, and for audio data, Librosa is used to extract features (pitch, volume, spectral characteristics).

[1371] 3. Psychological analysis

[1372] The preprocessed data is sent to a cloud server, where a generative AI model is used to analyze the user's psychological state. The server achieves this using platforms such as Google Cloud AI and AWS SageMaker.

[1373] 4. Adjusting the operating parameters

[1374] The server adjusts driving parameters based on the inferred psychological state, for example generating instructions to reduce speed and avoid sudden maneuvers if the user is nervous.

[1375] 5. Proposal Generation

[1376] The server generates advice based on the user's psychological state, suggests music, and suggests rest areas. These suggestions are sent to the mobile device and provided to the user as audio guidance or on-screen displays.

[1377] 6. Sending adjustment commands and executing operations

[1378] The server then sends the adjusted driving parameters to the vehicle's autonomous driving system and applies them, enabling driving that is appropriate for the driver's psychological state.

[1379] Specific examples

[1380] Scenario 1: Tired driver

[1381] 1. Data collection: The smartphone camera captures the driver's tired expression, and the microphone collects the driver's sigh.

[1382] 2. Data preprocessing: Extract the degree of eye contraction using OpenCV and low-pitched audio using Librosa.

[1383] 3. Psychological state analysis: The generative AI model infers that the driver is tired.

[1384] 4. Proposal generation: The server notifies the mobile device of break suggestions and soothing music playback.

[1385] 5. Sending adjustment instructions and executing driving: The server adjusts the driving parameters, and the vehicle follows the instructions to drive appropriately.

[1386] Prompt Sentence Examples

[1387] Prompt: "Analyze the driver's facial expressions and voice data to infer the driver's fatigue state and generate appropriate rest advice."

[1388] With the above configuration, this system is able to adjust driving and provide appropriate advice according to the psychological state of the driver and passengers, enabling a more secure autonomous driving experience.

[1389] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1390] Step 1:

[1391] The device captures the user's face using a camera installed in the vehicle and collects video data frame by frame. The specific input is image frames captured in real time, and the output is image data frame by frame, which is used for subsequent processing to extract feature points.

[1392] Step 2:

[1393] The terminal collects the user's voice in real time using a microphone installed in the vehicle. The specific input is the user's spoken voice, and the output is voice data. The collected voice data is used to extract features in the next step.

[1394] Step 3:

[1395] The device preprocesses the collected image data and extracts facial feature points. Specifically, it uses OpenCV to detect feature points such as eyes, mouth, and eyebrows from each frame of image. The input is the image data collected in step 1, and the output is the extracted facial feature point data.

[1396] Step 4:

[1397] The device preprocesses the collected audio data and extracts features. Specifically, Librosa is used to extract pitch, volume, and spectral characteristics from the audio data. The input is the audio data collected in step 2, and the output is the extracted audio feature data.

[1398] Step 5:

[1399] The terminal sends the preprocessed feature point data and feature volume data to the server. The specific input is the feature data obtained in step 3 and step 4, and the output is the data sent to the server.

[1400] Step 6:

[1401] The server inputs the received feature data into a generative AI model to infer the user's psychological state. The specific inputs are feature point data and voice feature data, and the output is the inferred psychological state (e.g., relaxed, tense, fatigued). The generative AI model achieves this using Google Cloud AI and AWS SageMaker.

[1402] Step 7:

[1403] The server generates instructions to adjust driving parameters based on the inferred psychological state, e.g., to reduce speed if the user is tired. The input is the psychological state inferred in step 6, and the output is the adjusted driving parameters.

[1404] Step 8:

[1405] The server sends the generated driving parameters or advice to the mobile device. For example, a break suggestion or a music playback instruction is generated in this step. The input is the driving parameters generated in step 7, and the output is the notification content.

[1406] Step 9:

[1407] The device displays or voices the received notification to the driver. Specifically, it displays the suggestion on the screen of the mobile device or informs the user as voice guidance. The input is the notification content sent in step 8, and the output is the display on the device or voice guidance.

[1408] Step 10:

[1409] The server sends the adjusted driving parameters to the automated driving system, and the vehicle performs driving based on these instructions. The input is the driving parameters generated in step 7, and the output is the actual vehicle behavior.

[1410] Example prompt sentence:

[1411] Prompt: "Analyze the driver's facial expressions and voice data to infer the driver's fatigue state and generate appropriate rest advice."

[1412] The above are the processing steps in the embodiment of the present invention.

[1413] 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.

[1414] System Overview

[1415] This invention is an autonomous driving system that uses a camera and microphone installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model and an emotion engine to infer the user's psychological state and adjust driving parameters. By combining this with an emotion engine, it becomes possible to more accurately recognize the user's diverse emotions and perform driving operations based on them.

[1416] Explanation of program processing

[1417] Data collection

[1418] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. A microphone also records what the user says in real time.

[1419] Data Preprocessing

[1420] Device: Extracts facial features (e.g., the position of the eyes, mouth, and eyebrows) from the collected image data, and extracts features (e.g., pitch, volume, and spectral characteristics) from the audio data, converting the data into a format that is easy for the AI ​​model described below to handle.

[1421] Sending data

[1422] Terminal: Sends preprocessed data (facial feature points and audio features) to the server.

[1423] Psychological analysis

[1424] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to infer the user's mental state and emotions, for example, identifying emotions such as tension, relief, surprise, anger, and joy.

[1425] Adjusting operating parameters

[1426] Server: Adjusts driving parameters (speed, acceleration, braking) based on the inferred psychological state and emotions. For example, if the user is angry, the server will slow down the vehicle to ensure safety and drive calmly.

[1427] Sending adjustment instructions

[1428] Server: Sends the adjusted driving parameters to the automated driving system. This transmission is usually done in real time, and is quickly reflected in the automated driving system inside the vehicle.

[1429] Execution of operation

[1430] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. This allows for safe and secure driving that is tailored to the user's psychological state.

[1431] Specific examples

[1432] Scenario 1: The user is angry

[1433] Data collection:

[1434] Device: The camera captures the user's angry facial expression (e.g., frown, closed mouth), and the microphone records the user's strong-tongued remarks.

[1435] Data preprocessing:

[1436] Device: Extract anger features from image data using the Dlib library, and extract high-pitch and high-volume features from audio data.

[1437] Psychological Analysis:

[1438] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is angry."

[1439] Adjusting operating parameters:

[1440] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[1441] Sending adjustment orders:

[1442] Server: Sends new driving parameters to the vehicle.

[1443] Execution of the operation:

[1444] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[1445] Scenario 2: Happy users

[1446] Data collection:

[1447] Device: The camera captures the user's smile, and the microphone records the user's conversation in a cheerful tone.

[1448] Data preprocessing:

[1449] Device: From the facial features, smiling features (such as the corners of the mouth turning up and the corners of the eyes turning up) are extracted, and stable pitch and volume features are extracted from the audio data.

[1450] Psychological Analysis:

[1451] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is happy."

[1452] Adjusting operating parameters:

[1453] Server: Maintain normal operating parameters or configure for slightly faster speeds.

[1454] Sending adjustment orders:

[1455] Server: Sends a command to the vehicle to continue normal operation.

[1456] Execution of the operation:

[1457] Terminal: The vehicle follows the command and maintains its current operating parameters.

[1458] The above is a specific embodiment of the automated driving system of the present invention. This allows driving to be performed in accordance with the diverse emotions and psychological states of passengers, thereby improving safety and peace of mind.

[1459] The processing flow will be explained below.

[1460] Specific processing steps of the system

[1461] Step 1:

[1462] Data collection

[1463] Terminal: A camera installed inside the vehicle captures the user's facial expressions frame by frame and collects image data. A microphone also records the user's speech in real time.

[1464] How it works: The camera takes pictures at 30 frames per second and stores them in storage. The microphone captures audio at high resolution of 24-bit, 96kHz and stores it in storage.

[1465] Step 2:

[1466] Data Preprocessing

[1467] Terminal: Extracts facial features (e.g., eye, mouth, and eyebrow positions) from collected image data, and extracts features (e.g., pitch, volume, and spectral characteristics) from audio data.

[1468] Specific operation: The Dlib library is used to identify the positions of the eyes, mouth, and eyebrows from the image data, and 68 feature points are extracted. MFCC (Mel-Frequency Cepstrum Coefficients) is used to extract 13-dimensional feature values ​​from the audio data every 0.1 seconds.

[1469] Step 3:

[1470] Sending data

[1471] Terminal: Sends preprocessed data (facial feature points and audio features) to the server.

[1472] Specific operation: Data is sent asynchronously to a server via the Internet or a dedicated in-vehicle network using HTTP or WebSocket protocols.

[1473] Step 4:

[1474] emotion recognition

[1475] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to recognize the user's emotions. The generative AI model estimates the psychological state, and the emotion engine identifies the emotion.

[1476] Specific behavior: The generative AI model analyzes the pre-processed data and outputs psychological states such as "tension," "relief," and "surprise." Based on this, the emotion engine classifies the user's specific emotions into "anger," "joy," "sadness," etc.

[1477] Step 5:

[1478] Adjusting operating parameters

[1479] Server: Determines driving parameters (speed, acceleration, braking) based on the inferred psychological state and emotions.

[1480] Specific behavior: Uses a rules engine to generate driving commands such as "slow down if the user is angry" or "maintain speed if the user is happy."

[1481] Step 6:

[1482] Sending adjustment instructions

[1483] Server: Sends the generated new driving parameters to the autonomous driving system.

[1484] Specific operation: Driving parameters are sent to the terminal in the vehicle using HTTP or a dedicated protocol.

[1485] Step 7:

[1486] Execution of operation

[1487] Terminal: Apply the new driving parameters received from the server to the autonomous driving system.

[1488] Specific behavior: The autonomous driving system uses the new control parameters to perform driving maneuvers such as speed, acceleration, and braking. For example, if the user is nervous, the system will reduce the speed to 20 km / h.

[1489] Specific examples

[1490] Scenario 1: The user is angry

[1491] Step 1:

[1492] Data collection

[1493] Device: The camera captures the user's angry facial expression (frown, tight-lipped mouth), and the microphone records the user's strong-tongued remarks.

[1494] Step 2:

[1495] Data Preprocessing

[1496] Terminal: Use the Dlib library to extract facial features, and high-pitch and high-volume features from audio data.

[1497] Step 3:

[1498] Sending data

[1499] Terminal: Sends preprocessed data to the server.

[1500] Step 4:

[1501] emotion recognition

[1502] Server: The generative AI model and emotion engine analyze the data and recognize that the user is feeling "anger."

[1503] Step 5:

[1504] Adjusting operating parameters

[1505] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[1506] Step 6:

[1507] Sending adjustment instructions

[1508] Server: Sends new driving parameters to the vehicle.

[1509] Step 7:

[1510] Execution of operation

[1511] Terminal: Reduce speed to 20km / h and begin driving smoothly.

[1512] Scenario 2: Happy users

[1513] Step 1:

[1514] Data collection

[1515] Device: The camera captures the user's smile. The microphone records the user's pleasant conversation.

[1516] Step 2:

[1517] Data Preprocessing

[1518] Device: Extracts features of a smile, such as eyes and mouth turned up, and extracts stable pitch and volume features from audio data.

[1519] Step 3:

[1520] Sending data

[1521] Terminal: Sends preprocessed data to the server.

[1522] Step 4:

[1523] emotion recognition

[1524] Server: The generative AI model and emotion engine analyze the data and recognize that the user is feeling "joy."

[1525] Step 5:

[1526] Adjusting operating parameters

[1527] Server: Issues a command to continue normal operation.

[1528] Step 6:

[1529] Sending adjustment instructions

[1530] Server: Sends commands to the vehicle.

[1531] Step 7:

[1532] Execution of operation

[1533] Terminal: Maintain current operating parameters and continue normal operation.

[1534] This will enable safe and secure driving based on specific emotions.

[1535] Example 2

[1536] 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."

[1537] In an automated driving system for a vehicle, there is a demand for providing a safer and more comfortable driving experience by accurately grasping the user's psychological state and adjusting driving parameters accordingly. However, with conventional technologies, it has been difficult to accurately recognize the user's emotions and psychological state in real time and dynamically adjust driving parameters based on that. This invention is intended to solve this problem.

[1538] 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.

[1539] In this invention, the server, in a vehicle linked to an automated driving system that estimates a user's psychological state and adjusts driving parameters, includes: means for collecting a user's facial expression in real time using an internally installed image capture device; means for collecting the user's voice in real time using an internally installed voice recording device; means for preprocessing the collected facial expression data and voice data and extracting feature points and feature quantities; means for estimating the user's psychological state from the extracted feature data using a generative technology model; means for adjusting the driving parameters based on the estimated psychological state; and means for transmitting the adjusted driving parameters to the automated driving system and applying them, thereby enabling appropriate adjustment of driving parameters based on the user's real-time psychological state.

[1540] A "user" is a person who uses this system.

[1541] "Mental state" refers to the user's feelings and mental state, and specifically includes emotions such as tension, relief, surprise, anger, and joy.

[1542] "Driving parameters" are control data related to vehicle driving, specifically settings such as speed, acceleration, and braking.

[1543] An "autonomous driving system" is a system that automatically controls the driving of a vehicle and has the function of allowing the vehicle to drive autonomously without the user performing any driving operations.

[1544] The "photography device" is a device that is installed inside the vehicle and captures the user's facial expressions in real time.

[1545] The "voice recording device" is a device that is installed inside the vehicle and records the user's speech in real time.

[1546] "Preprocessing" refers to the process of extracting feature points and feature quantities from collected raw data (image data and audio data).

[1547] "Feature points" refer to characteristic points on the user's face extracted from image data, such as the positions of the eyes, mouth, and eyebrows.

[1548] "Features" are acoustic features extracted from audio data, specifically pitch, volume, spectral characteristics, etc.

[1549] A "generative model" is an artificial intelligence model used to infer a user's psychological state from collected data.

[1550] "Adjusted driving parameters" refer to new driving parameters that have been adjusted based on the user's psychological state.

[1551] "Application" means reflecting the adjusted driving parameters in the actual automated driving system.

[1552] System Overview

[1553] This invention is an autonomous driving system that uses a camera and microphone installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model and emotion engine to infer the user's psychological state and adjust driving parameters. By combining this with an emotion engine, it is possible to more accurately recognize the user's diverse emotions and perform driving operations based on them.

[1554] Data collection

[1555] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. An audio recording device also records the user's speech in real time.

[1556] Hardware used:

[1557] Camera (e.g. high-resolution camera)

[1558] Microphone (e.g. high-sensitivity microphone)

[1559] Data Preprocessing

[1560] Device: Facial feature points (e.g., eye, mouth, and eyebrow positions) are extracted from the collected image data using a face recognition library (e.g., Dlib), while acoustic features (e.g., pitch, volume, and spectral characteristics) are extracted from the audio data.

[1561] Software used:

[1562] Dlib library (for face recognition)

[1563] Audio analysis tools (e.g. FFT)

[1564] Sending data

[1565] Terminal: The facial feature point data and voice feature data obtained in the preprocessing are encrypted and sent to the server using the HTTPS protocol to ensure data security.

[1566] Psychological analysis

[1567] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to infer the user's psychological state. For example, identify emotions such as tension, relief, surprise, anger, and joy.

[1568] Software used:

[1569] Generative technology models (e.g., BERT-based sentiment analysis models)

[1570] Emotion Engine

[1571] Adjusting operating parameters

[1572] Server: Optimizes driving parameters (speed, acceleration, braking) based on the estimated psychological state. If the user is nervous, the server reduces speed and smooths braking to make the driver more calm.

[1573] Sending adjustment instructions

[1574] Server: Transmits the calculated driving parameters to the vehicle's autonomous driving system in real time. This transmission also uses the HTTPS protocol to ensure data integrity and security.

[1575] Execution of operation

[1576] Terminal: The terminal applies the new driving parameters received from the server to the actual vehicle control and performs driving, thereby automatically driving according to the user's psychological state.

[1577] Specific examples

[1578] Scenario 1: The user is angry

[1579] Data collection:

[1580] Device: The camera captures the user's angry facial expression (e.g., frown, closed mouth), and the microphone records the user's strong-tongued remarks.

[1581] Data preprocessing:

[1582] Device: Extract anger features from image data using the Dlib library, and extract high-pitch and high-volume features from audio data.

[1583] Psychological Analysis:

[1584] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is angry."

[1585] Adjusting operating parameters:

[1586] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[1587] Sending adjustment orders:

[1588] Server: Sends new driving parameters to the vehicle.

[1589] Execution of the operation:

[1590] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[1591] Scenario 2: Happy users

[1592] Data collection:

[1593] Device: The camera captures the user's smile, and the microphone records the user's conversation in a cheerful tone.

[1594] Data preprocessing:

[1595] Device: From the facial features, smiling features (such as the corners of the mouth turning up and the corners of the eyes turning up) are extracted, and stable pitch and volume features are extracted from the audio data.

[1596] Psychological Analysis:

[1597] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is happy."

[1598] Adjusting operating parameters:

[1599] Server: Maintain normal operating parameters or configure for slightly faster speeds.

[1600] Sending adjustment orders:

[1601] Server: Sends a command to the vehicle to continue normal operation.

[1602] Execution of the operation:

[1603] Terminal: The vehicle follows the command and maintains its current operating parameters.

[1604] Prompt Sentence Examples

[1605] Scenario 1 prompt statement:

[1606] If the user is angry, please follow the instructions below to adjust the operating parameters:

[1607] Angry facial expression captured on camera

[1608] Logged voice data with strong verbal content

[1609] Scenario 2 prompt statement:

[1610] If the user is willing, please follow the instructions below to adjust the operating parameters:

[1611] A photo of a smiling face captured on camera

[1612] Logged voice data with a pleasant tone

[1613] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1614] Step 1: Data collection

[1615] Device: The in-car camera captures the user's face frame by frame and collects image data, while the microphone records what the user says in real time.

[1616] Input: Video frames of the user's face, audio of what the user is saying

[1617] Output: User's face image data, voice data

[1618] Specific behavior:

[1619] The interior camera captures images at 30 frames per second.

[1620] The microphone records audio at a sample rate of 16kHz.

[1621] Step 2: Data Preprocessing

[1622] Device: Facial feature points (e.g., eye, mouth, and eyebrow positions) are extracted from the collected image data using a face recognition library (e.g., Dlib), while acoustic features (e.g., pitch, volume, and spectral characteristics) are extracted from the audio data.

[1623] Input: Facial image data, audio data

[1624] Output: Facial feature data, acoustic feature data

[1625] Specific behavior:

[1626] Using the Dlib library, 68 facial feature points are extracted from each frame.

[1627] The frequency components of the audio data are analyzed using FFT (Fast Fourier Transform) and extracted as features.

[1628] Step 3: Sending data

[1629] Terminal: The facial feature point data and voice feature data obtained in the preprocessing are encrypted and sent to the server using the HTTPS protocol to ensure data security.

[1630] Input: facial feature point data, acoustic feature data

[1631] Output: Notification that encrypted data has been sent

[1632] Specific behavior:

[1633] Package the data in JSON format and send a POST request to the server.

[1634] Step 4: Psychological analysis

[1635] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to infer the user's psychological state. For example, identify emotions such as tension, relief, surprise, anger, and joy.

[1636] Input: Preprocessed data (facial feature points, acoustic features)

[1637] Output: User's mental state and emotion category

[1638] Specific behavior:

[1639] Input feature data into a generative AI model (e.g., a BERT-based sentiment analysis model) to predict sentiment categories.

[1640] The emotion engine analyzes the prediction results and determines the most appropriate emotion.

[1641] Step 5: Adjusting the operating parameters

[1642] Server: Optimizes driving parameters (speed, acceleration, braking) based on the estimated psychological state. If the user is nervous, the server reduces speed and smooths braking to make the driver more calm.

[1643] Input: User's mental state and emotion category

[1644] Output: Adjusted operating parameters

[1645] Specific behavior:

[1646] It obtains driving data such as current speed and acceleration and calculates new driving parameters based on that.

[1647] Step 6: Sending adjustment instructions

[1648] Server: Transmits the calculated driving parameters to the vehicle's autonomous driving system in real time. This transmission also uses the HTTPS protocol to ensure data integrity and security.

[1649] Input: Adjusted operating parameters

[1650] Output: Notification of completion of sending of operation parameters

[1651] Specific behavior:

[1652] The new driving parameters are packaged in JSON format and sent as a POST request to the vehicle's control system.

[1653] Step 7: Execute the operation

[1654] Terminal: The terminal applies the new driving parameters received from the server to the actual vehicle control and performs driving, thereby automatically driving according to the user's psychological state.

[1655] Input: New operating parameters

[1656] Output: Vehicle control actions based on driving parameters

[1657] Specific behavior:

[1658] The vehicle's ECU (Electronic Control Unit) receives the new driving parameters and reflects them in brake and accelerator control.

[1659] (Application example 2)

[1660] 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."

[1661] Conventional automated driving systems operate based on predetermined driving parameters without considering the user's emotions or psychological state, and therefore cannot fully guarantee the user's safety or comfort. Furthermore, they lack a mechanism for providing support content that adapts to the user's psychological state while driving, making it difficult to promote stress reduction and relaxation for the user. Therefore, there is a need for a system that can adjust driving parameters and provide support content appropriately according to the user's psychological state.

[1662] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting the user's facial expressions in real time using an internally installed camera, means for collecting the user's voice in real time using an internally installed microphone, means for preprocessing the collected facial expression data and voice data to extract feature points and feature quantities, means for inferring the user's psychological state from the extracted feature data using a generative AI model, means for adjusting driving parameters based on the inferred psychological state, means for transmitting the adjusted driving parameters to the automated driving system and applying them, and means for providing support content during driving in accordance with the user's emotional state. This makes it possible to adjust driving parameters in accordance with the user's psychological state and provide appropriate support content.

[1663] An "autonomous driving system that performs driving operations" is a system that automatically drives a vehicle, and is a general term for the programs and hardware that perform driving operations such as adjusting speed, changing direction, and braking.

[1664] The "interiorly installed camera" is a camera device that is attached to the interior of the vehicle and captures the user's facial expressions in real time.

[1665] The "interiorly installed microphone" is a microphone device that is installed inside the vehicle and collects the user's voice in real time.

[1666] "Facial expression data" is image data of the user's face captured by a camera, and includes facial feature points.

[1667] "Voice data" refers to voice data of a user's speech recorded by a microphone, and includes voice features.

[1668] "Preprocessing" refers to a data processing procedure for extracting feature points and feature quantities from collected data.

[1669] "Feature points" are data points relating to the position and shape of a face, and indicate specific positions such as the eyes, mouth, and eyebrows.

[1670] "Features" are specific numerical data such as pitch, volume, and spectral characteristics extracted from audio data.

[1671] A "generative AI model" is an artificial intelligence model that analyzes collected data and infers the user's psychological state.

[1672] The "emotion engine" is an analytical engine that identifies the user's emotions and generates appropriate responses based on them.

[1673] "Driving parameters" are set values ​​for controlling the driving of a vehicle, and include speed, acceleration, braking, and the like.

[1674] "Support content" refers to supplementary information and entertainment provided while driving according to the user's psychological state, and includes relaxing music, positive news, and the like.

[1675] This invention is an autonomous driving system that uses a camera and microphone installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model and emotion engine to infer the user's psychological state and adjust driving parameters.Furthermore, it is a system that provides support content according to the user's psychological state.

[1676] System Overview

[1677] Hardware used

[1678] Camera: An image capturing device installed inside the vehicle.

[1679] Microphone: An audio collection device installed inside the vehicle.

[1680] Smart glasses: A wearable device for capturing the user's facial expressions and voice in real time.

[1681] Minicomputer: A small computer (e.g., Raspberry Pi) for data preprocessing and analysis.

[1682] On-board computer: An on-board computer used to adjust the vehicle's operating parameters.

[1683] Software used

[1684] Facial expression recognition library: Dlib (a library for extracting facial feature points).

[1685] Speech analysis library: LibROSA (a library for extracting speech features).

[1686] Generative AI models: Natural language processing models such as GPT-3.

[1687] Sentiment engine: A sentiment analysis engine such as Azure Cognitive Services.

[1688] Specific Embodiments

[1689] Data collection

[1690] The device (smart glasses) captures the user's facial expressions in real time with a camera installed inside the vehicle and records the user's voice with a microphone, and these data are sent to a minicomputer.

[1691] Data Preprocessing

[1692] The minicomputer uses Dlib to extract facial features from image data, and LibROSA to extract audio features such as pitch, volume, and spectral characteristics from audio data.

[1693] Psychological analysis

[1694] The server inputs the preprocessed data into a generative AI model and emotion engine to infer the user's psychological state, and determines whether the user is happy, angry, or other emotional state based on the inference.

[1695] Providing support content

[1696] If the user is inferred to be angry, the following prompt sentence can be input into the generative AI model to generate instructions to provide relaxing music:

[1697] "User's facial expression data: {facial expression data}

[1698] User's voice data: {voice data}

[1699] Inferred emotion: Anger

[1700] Suggested action: Play relaxing music

[1701] Based on this instruction, the appropriate music will be played.

[1702] If the user is deemed happy, the generative AI model is given the following prompt:

[1703] "User's facial expression data: {facial expression data}

[1704] User's voice data: {voice data}

[1705] Inferred emotion: Joy

[1706] Recommended action: Read positive news

[1707] This will read out positive news aloud.

[1708] Adjusting operating parameters

[1709] The server adjusts driving parameters (speed, braking, etc.) based on the inferred psychological state. For example, if the user is angry, it generates an instruction to slow down the vehicle and sends it to the on-board computer. This automatically adjusts the vehicle's driving, increasing the user's safety.

[1710] Example

[1711] Scenario 1: The user is angry

[1712] Data collection:

[1713] The camera captures the user's angry expression (frowns, closed mouth), and the microphone records the strong words spoken.

[1714] Sentiment Analysis and Response:

[1715] The generative AI model and emotion engine infer that the user is angry and generate instructions to provide relaxing music.

[1716] Adjusting operating parameters:

[1717] The server generates and transmits to the vehicle instructions to reduce the vehicle's speed.

[1718] Scenario 2: Happy users

[1719] Data collection:

[1720] The camera captures the user's smile, and the microphone records conversations in a playful tone.

[1721] Sentiment Analysis and Response:

[1722] The generative AI model and emotion engine infer that the user is happy and generate instructions to deliver positive news.

[1723] Adjusting operating parameters:

[1724] The server generates and transmits instructions to the vehicle to maintain normal operating parameters.

[1725] This will enable the adjustment of driving parameters according to the user's psychological state and the provision of appropriate support content, resulting in safer and more comfortable automated driving.

[1726] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1727] Step 1:

[1728] The device (smart glasses) captures images of the user's face in real time using an internal camera and acquires image data. It also records the user's voice in real time using an internal microphone and acquires voice data. This input data (image data and voice data) is sent to a minicomputer.

[1729] Step 2:

[1730] The minicomputer extracts facial features (e.g., eye, mouth, and eyebrow positions) from the received image data using the Dlib library, and simultaneously extracts pitch, volume, and spectral characteristics from the audio data using the LibROSA library, resulting in preprocessed data (facial features and audio features).

[1731] Step 3:

[1732] The preprocessed data is sent to the server. The input data (facial features and voice features) is fed into the generative AI model and emotion engine to infer the user's psychological state. The generative AI model analyzes the combined facial and voice data and identifies emotions such as "the user is angry" or "the user is happy."

[1733] Step 4:

[1734] The server adjusts driving parameters based on the estimated psychological state. Specifically, it generates instructions to set driving parameters such as vehicle speed, acceleration, and braking. For example, if it estimates that the user is angry, it generates an instruction to slow down the vehicle.

[1735] Step 5:

[1736] The server sends the generated instructions to the vehicle's on-board computer. The input data (adjusted driving parameters) are applied by the on-board computer and reflected in actual driving operations. This allows driving to be performed in accordance with the user's psychological state.

[1737] Step 6:

[1738] The server provides support content according to the user's psychological state. For example, if the user is angry, the server generates an instruction to play relaxing music and sends it to the minicomputer. The minicomputer then plays the relaxing music according to the instruction.

[1739] Step 7:

[1740] If the user's psychological state is calmed by the support content they receive, the camera and microphone will again collect their facial expressions and voice, and the entire system will start operating again, making it possible to provide continuous driving support that responds to the user's psychological state.

[1741] As a result, safe and comfortable automated driving is achieved by adjusting driving parameters according to the user's psychological state and providing support content.

[1742] 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.

[1743] 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.

[1744] 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.

[1745] [Fourth embodiment]

[1746] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1747] 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.

[1748] 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).

[1749] 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.

[1750] 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.

[1751] 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).

[1752] 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.

[1753] 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.

[1754] 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.

[1755] 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.

[1756] 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.

[1757] 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.

[1758] 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."

[1759] System Overview

[1760] This invention is an autonomous driving system that uses cameras and microphones installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model to infer the user's psychological state.The system aims to provide passengers with a sense of security by adjusting driving parameters based on the inferred psychological state.

[1761] Explanation of program processing

[1762] Data collection

[1763] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. A microphone also records what the user says in real time.

[1764] Data Preprocessing

[1765] Device: Extracts facial features (e.g., the position of the eyes, mouth, and eyebrows) from the collected image data, and extracts features (e.g., pitch, volume, and spectral characteristics) from the audio data, converting the data into a format that is easy for the AI ​​model described below to handle.

[1766] Psychological analysis

[1767] Server: The preprocessed data is sent to the server, which uses a generative AI model to infer the user's psychological state from this data, for example, identifying emotions such as tension, relief, and surprise.

[1768] Adjusting operating parameters

[1769] Server: Generates commands to adjust driving parameters (e.g., speed, acceleration, braking) based on the estimated psychological state. For example, if the user is nervous, the server may slow down the speed or prohibit sudden acceleration.

[1770] Sending adjustment instructions

[1771] Server: Sends the adjusted driving parameters to the automated driving system. This transmission is usually done in real time, and is quickly reflected in the automated driving system inside the vehicle.

[1772] Execution of operation

[1773] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. This allows for safe and secure driving that is tailored to the user's psychological state.

[1774] Specific examples

[1775] Scenario 1: The user is surprised

[1776] Data collection:

[1777] Device: The camera captures the user's surprised expression, and the microphone records the user's high-pitched utterance, such as "I'm scared."

[1778] Data preprocessing:

[1779] Device: Extracts facial features such as wide-open eyes and an open mouth, and extracts high-pitched and high-volume features from audio data.

[1780] Psychological Analysis:

[1781] Server: The generative AI model analyzes these feature data and infers that the user is surprised.

[1782] Adjusting operating parameters:

[1783] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[1784] Sending adjustment orders:

[1785] Server: Sends new driving parameters to the vehicle.

[1786] Execution of the operation:

[1787] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[1788] Scenario 2: User is relaxed

[1789] Data collection:

[1790] Device: The camera captures the user's relaxed facial expression, and the microphone records calm conversations.

[1791] Data preprocessing:

[1792] Device: From facial features, relaxed eye and mouth shapes are extracted, and stable pitch and volume features are extracted from audio data.

[1793] Psychological Analysis:

[1794] Server: The generative AI model analyzes these feature data and infers that the user is relaxed.

[1795] Adjusting operating parameters:

[1796] Server: Maintain normal operating parameters or configure for slightly faster speeds.

[1797] Sending adjustment orders:

[1798] Server: Sends a command to the vehicle to continue normal operation.

[1799] Execution of the operation:

[1800] Terminal: The vehicle follows your commands and you can continue driving relaxed.

[1801] The above is a specific embodiment of the automated driving system of the present invention, which enables driving according to the psychological state of the passengers, providing a greater sense of security.

[1802] The processing flow will be explained below.

[1803] Step 1:

[1804] Data collection

[1805] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. A microphone also records what the user says in real time.

[1806] How it works: The camera takes pictures at 30 frames per second and stores them in storage. The microphone captures audio at high resolution of 24-bit, 96kHz and stores it in storage.

[1807] Step 2:

[1808] Data Preprocessing

[1809] Terminal: Extracts facial features from collected image data, and extracts features such as pitch, volume, and spectral characteristics from audio data.

[1810] Specific operation: The Dlib library is used to identify the positions of the eyes, mouth, and eyebrows from the image data, and 68 feature points are extracted. MFCC (Mel-Frequency Cepstrum Coefficients) is used to extract 13-dimensional feature values ​​from the audio data every 0.1 seconds.

[1811] Step 3:

[1812] Sending data

[1813] Terminal: Sends preprocessed data (facial feature points and audio features) to the server.

[1814] Specific operation: Data is sent asynchronously to a server using HTTP or WebSocket protocol over the Internet or a dedicated in-vehicle network.

[1815] Step 4:

[1816] Psychological analysis

[1817] Server: Inputs the received preprocessed data into the generative AI model and infers the user's psychological state.

[1818] Specific operation: The received data is input into a generative AI model (e.g., BERT or CNN), which outputs an emotion label, such as whether the user is nervous, relaxed, or surprised.

[1819] Step 5:

[1820] Adjusting operating parameters

[1821] Server: Determines driving parameters (speed, acceleration, braking) based on the estimated psychological state.

[1822] Specific behavior: The rule engine is executed to generate new driving parameters, such as slowing down if the psychological state is "surprised" and increasing speed if the psychological state is "relaxed."

[1823] Step 6:

[1824] Sending adjustment instructions

[1825] Server: Sends new driving parameters to the vehicle's autonomous driving system.

[1826] Specific operation: Sends command data to the terminal using HTTP or a dedicated protocol.

[1827] Step 7:

[1828] Execution of operation

[1829] Terminal: Follows instructions from the server and applies new driving parameters to the autonomous driving system.

[1830] Specific operation: The autonomous driving system applies settings such as new speed limits and prohibitions on sudden acceleration to the controller and performs actual driving operations.

[1831] The above is the specific process and operation flow for each step. This system realizes safe and secure driving that is in line with the passenger's psychological state.

[1832] Example 1

[1833] 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."

[1834] In current automated driving systems, vehicle operating parameters are adjusted mainly based on the external environment and traffic conditions, and the psychological state of passengers is not taken into consideration. This often causes passengers to feel nervous or anxious, which can hinder a comfortable journey. In particular, sudden acceleration or braking can increase the psychological burden on passengers.

[1835] 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.

[1836] In this invention, the server includes means for extracting facial feature points from collected facial expression data, means for extracting pitch, volume, and spectral characteristics from voice data, and means for inferring the user's psychological state from the feature data using a generative AI model. This enables adjustment of driving parameters based on the user's psychological state, improving passenger comfort.

[1837] The "camera" is a photographing device that is installed inside the vehicle and collects the user's facial expressions frame by frame in real time.

[1838] A "microphone" is a recording device installed inside a vehicle to collect the user's voice in real time.

[1839] "Facial expression data" is image information of the user's face collected by a camera.

[1840] "Voice data" refers to information about a user's speech or voice collected by a microphone.

[1841] "Feature points" are the main positional information of the face (e.g., the positions of the eyes, mouth, and eyebrows) extracted from the facial expression data.

[1842] "Features" are sound characteristics (e.g., pitch, volume, spectral characteristics) extracted from audio data.

[1843] The "server" is a central processing unit for processing the collected facial expression data and voice data.

[1844] A "generative AI model" is an artificial intelligence model that takes feature data as input and uses it to infer a user's psychological state.

[1845] "Driving parameters" are set values ​​related to driving operations such as vehicle speed, acceleration, and braking.

[1846] The "command" is instruction information generated by the server that includes the details of adjustment of the operating parameters.

[1847] "Real-time transmission" means that data is transmitted almost instantly to a server or to an automated driving system.

[1848] MODE FOR CARRYING OUT THE INVENTION

[1849] The present invention is an autonomous driving system that uses a camera and microphone inside the vehicle to collect a user's facial expressions and voice in real time and uses a generative AI model to infer their psychological state. The system aims to provide passengers with a sense of security by adjusting driving parameters based on the inferred psychological state. A specific embodiment of the system is described below.

[1850] Data collection

[1851] Terminal: A camera installed inside the vehicle captures the user's face frame by frame and collects image data. The camera used is a general-purpose high-resolution camera. The microphone records the user's voice in real time. Specifically, a general-purpose recording device with noise-canceling functionality is used.

[1852] Data Preprocessing

[1853] Device: The facial recognition software "dlib" is used to extract facial features (e.g., the position of the eyes, mouth, and eyebrows) from the collected image data. Also, "Librosa" is used to extract features (e.g., pitch, volume, and spectral characteristics) from the audio data. This process converts the data into a format that is easy for the AI ​​model described below to handle.

[1854] Sending data

[1855] Terminal: The preprocessed data is sent to the server. The "WebSocket" protocol is used for transmission to achieve real-time communication, which enables high-speed, low-latency data transfer.

[1856] Psychological analysis

[1857] Server: Receives the preprocessed data and uses a generative AI model to infer the user's psychological state. The AI ​​model used is a customized model based on GPT-4. This model receives image and audio feature data as input and classifies the user's psychological state (e.g., nervous, relaxed, surprised).

[1858] Adjusting operating parameters

[1859] Server: Generates commands to adjust driving parameters based on the estimated psychological state. These commands include adjustments to vehicle speed, acceleration, braking, etc. For example, if the user is surprised, it generates a command to slow down the vehicle and prohibit sudden maneuvers. This is done by inputting a prompt such as "Please suggest driving parameters when the user is surprised" into the generative AI model.

[1860] Sending adjustment instructions

[1861] Server: Sends the generated driving parameters to the autonomous driving system. This transmission is also done in real time and is immediately reflected in driving operations. The "MQTT" (Message Queuing Telemetry Transport) protocol is used for transmission, achieving highly reliable and low-latency communication.

[1862] Execution of operation

[1863] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. The vehicle's autonomous driving system reduces speed and controls braking based on the received commands, thereby realizing safe and secure driving for the user.

[1864] Specific examples

[1865] A specific scenario is shown below.

[1866] Scenario 1: The user is surprised

[1867] Data collection:

[1868] Device: The camera captures the user's surprised expression, and the microphone records the user's high-pitched utterances such as "I'm scared."

[1869] Data preprocessing:

[1870] Device: Extracts facial features such as wide-open eyes and an open mouth, and extracts high-pitch and high-volume features from audio data.

[1871] Sending data:

[1872] Terminal: Sends pre-processed data to the server in real time.

[1873] Psychological Analysis:

[1874] Server: The generative AI model analyzes the feature data and infers that the user is surprised.

[1875] Adjusting operating parameters:

[1876] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[1877] Sending adjustment orders:

[1878] Server: Sends new driving parameters to the vehicle.

[1879] Execution of the operation:

[1880] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[1881] Scenario 2: User is relaxed

[1882] Data collection:

[1883] Device: The camera captures the user's relaxed facial expression, and the microphone records calm conversations.

[1884] Data preprocessing:

[1885] Device: Extracts relaxed eye and mouth shapes from facial feature points, and extracts stable pitch and volume features from audio data.

[1886] Sending data:

[1887] Terminal: Sends preprocessed data to the server.

[1888] Psychological Analysis:

[1889] Server: The generative AI model analyzes the feature data and infers that the user is relaxed.

[1890] Adjusting operating parameters:

[1891] Server: Generates commands to maintain normal operating parameters or slightly increase speed.

[1892] Sending adjustment orders:

[1893] Server: Sends a command to the vehicle to continue normal operation.

[1894] Execution of the operation:

[1895] Terminal: The vehicle follows your commands and you can continue driving relaxed.

[1896] The above is a specific embodiment of the automated driving system of the present invention, which enables driving according to the psychological state of the passengers, providing a greater sense of security.

[1897] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1898] Program processing flow

[1899] Step 1: Data collection

[1900] Terminal: The camera captures the user's face frame by frame and collects image data. Specifically, it captures 30 frames per second. The microphone also records the user's voice in real time. Specifically, it uses a general-purpose, highly sensitive recording device and has the function of reducing background noise. This allows high-resolution image data and voice data to be obtained as input.

[1901] Step 2: Data Preprocessing

[1902] Terminal: The facial recognition software "dlib" is used to extract facial feature points (the positions of the eyes, mouth, and eyebrows) from the collected image data. At the same time, "Librosa" is used to extract feature quantities such as pitch, volume, and spectral characteristics from the audio data. Specifically, feature points are identified from the image of each frame, and the audio data is quantified. This results in the extracted feature points and feature quantities being obtained as output.

[1903] Step 3: Sending data

[1904] Terminal: The "WebSocket" protocol is used to send preprocessed data to the server. This allows feature points and feature quantities to be sent to the server at high speed and with low latency. Specifically, image feature points and audio feature quantities for each frame are packaged into packets and sent in real time.

[1905] Step 4: Psychological analysis

[1906] Server: Receives the preprocessed data and uses a generative AI model to infer the user's emotional state. The AI ​​model used is a customized version of GPT-4. Specifically, feature points and feature quantities are input into the model to classify the user's emotions (tension, relaxation, surprise, etc.). During this process, an inferred emotional state is output from the feature points and feature quantities received as input.

[1907] Step 5: Adjusting the operating parameters

[1908] Server: Generates commands to adjust driving parameters based on the estimated psychological state. Based on the output from the model, commands are generated to adjust speed, acceleration, braking, etc. For example, if it is estimated that the user is surprised, a command is generated to slow down the vehicle's speed and prohibit sudden maneuvers. This is done by inputting a prompt statement such as "Please suggest driving parameters for when the user is surprised" into the generative AI model. As a result, an adjustment command is generated.

[1909] Step 6: Sending adjustment instructions

[1910] Server: Sends the generated driving parameters to the autonomous driving system. This transmission is also done in real time and is immediately reflected in driving operations. The "MQTT" protocol is used for transmission. Specifically, a packet containing the generated driving parameters is created and sent to the autonomous driving system, which then outputs an adjustment command.

[1911] Step 7: Execute the operation

[1912] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. The vehicle's autonomous driving system reduces speed and controls braking in accordance with the received commands. This allows the user to drive safely and with peace of mind. Specifically, it reduces speed to 20 km / h and maintains smooth driving. The output of this process is a safe and relaxed driving state.

[1913] (Application example 1)

[1914] 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."

[1915] Conventional automated driving systems have been unable to adjust driving behavior to take into account the psychological state of the driver and passengers, making it difficult to provide a sense of security. Furthermore, there was no mechanism to provide driving advice or suggestions based on the psychological state in real time, so the effectiveness of reducing driver stress and promoting safe driving was limited.

[1916] 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.

[1917] In this invention, the server includes, in a vehicle linked to an automated driving system that performs driving operations, means for collecting a user's facial expressions in real time using an internally installed camera, means for collecting the user's voice in real time using an internally installed microphone, means for preprocessing the collected facial expression data and voice data and extracting feature points and feature quantities, means for inferring the user's psychological state from the extracted feature data using a generative AI model, means for adjusting driving parameters based on the inferred psychological state, means for notifying the user's mobile device of the generated suggestions and providing audio guidance or display, and means for transmitting the adjusted driving parameters to the automated driving system and applying them. This enables driving adjustments and the provision of appropriate advice based on the psychological states of the driver and passengers.

[1918] An "autonomous driving system that performs driving operations" is a system that automatically controls the operation of a vehicle using data input devices such as sensors and cameras.

[1919] A "camera" is an optical device for capturing images and video.

[1920] "User's facial expression" refers to the emotions and psychological state of the user that are indicated by the movements of the user's face and changes in the facial muscles.

[1921] A "microphone" is a sound capture device that converts sound waves into electrical signals.

[1922] "User voice" refers to the voice or sound patterns produced by a user.

[1923] "Facial expression data" refers to a collection of images and feature points of a user's face captured by a camera.

[1924] "Voice data" refers to the waveform information of the user's voice or sound collected by a microphone.

[1925] "Feature points" refer to important points on the face, including the position information of the eyes, nose, mouth, etc.

[1926] "Features" refers to information in audio data that is expressed by specific parameters (e.g., pitch, volume, spectral characteristics).

[1927] A "generative AI model" is an artificial intelligence algorithm used to infer a user's psychological state from collected data.

[1928] "Mental state" refers to the emotions and psychological state that a user is feeling.

[1929] "Driving parameters" refers to adjustable driving-related settings such as vehicle speed, acceleration, braking, etc.

[1930] "Mobile device" refers to a portable computing device such as a smartphone or tablet.

[1931] An "audio guide" is a system that provides users with information and instructions through audio.

[1932] "Notification" refers to an alert or message that notifies a user of specific information.

[1933] "Preprocessing" refers to the step of converting collected data into a form suitable for subsequent analysis and processing.

[1934] "Adjustment" refers to the act of changing driving parameters based on the estimated psychological state.

[1935] "Transmit" refers to the act of transferring data or instructions to another system or device.

[1936] "Application" refers to reflecting the adjusted driving parameters in the actual vehicle operation.

[1937]

[1938] The present invention is an autonomous driving system that uses a camera and microphone inside the vehicle to collect a user's facial expressions and voice in real time, and then uses a generative AI model to infer the user's psychological state based on the collected data. The system of the present invention includes a camera, a microphone, a server, a mobile device such as a smartphone, and a generative AI model.

[1939] System configuration

[1940] 1. Data Collection

[1941] The smartphone's camera and microphone are used to capture the driver's facial expressions and voice in real time. The camera captures facial expressions frame by frame, and the microphone records the voice.

[1942] 2. Data Preprocessing

[1943] The device preprocesses the collected image data using OpenCV to extract facial features, and for audio data, Librosa is used to extract features (pitch, volume, spectral characteristics).

[1944] 3. Psychological analysis

[1945] The preprocessed data is sent to a cloud server, where a generative AI model is used to analyze the user's psychological state. The server achieves this using platforms such as Google Cloud AI and AWS SageMaker.

[1946] 4. Adjusting the operating parameters

[1947] The server adjusts driving parameters based on the inferred psychological state, for example generating instructions to reduce speed and avoid sudden maneuvers if the user is nervous.

[1948] 5. Proposal Generation

[1949] The server generates advice based on the user's psychological state, suggests music, and suggests rest areas. These suggestions are sent to the mobile device and provided to the user as audio guidance or on-screen displays.

[1950] 6. Sending adjustment commands and executing operations

[1951] The server then sends the adjusted driving parameters to the vehicle's autonomous driving system and applies them, enabling driving that is appropriate for the driver's psychological state.

[1952] Specific examples

[1953] Scenario 1: Tired driver

[1954] 1. Data collection: The smartphone camera captures the driver's tired expression, and the microphone collects the driver's sigh.

[1955] 2. Data preprocessing: Extract the degree of eye contraction using OpenCV and low-pitched audio using Librosa.

[1956] 3. Psychological state analysis: The generative AI model infers that the driver is tired.

[1957] 4. Proposal generation: The server notifies the mobile device of break suggestions and soothing music playback.

[1958] 5. Sending adjustment instructions and executing driving: The server adjusts the driving parameters, and the vehicle follows the instructions to drive appropriately.

[1959] Prompt Sentence Examples

[1960] Prompt: "Analyze the driver's facial expressions and voice data to infer the driver's fatigue state and generate appropriate rest advice."

[1961] With the above configuration, this system is able to adjust driving and provide appropriate advice according to the psychological state of the driver and passengers, enabling a more secure autonomous driving experience.

[1962] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1963] Step 1:

[1964] The device captures the user's face using a camera installed in the vehicle and collects video data frame by frame. The specific input is image frames captured in real time, and the output is image data frame by frame, which is used for subsequent processing to extract feature points.

[1965] Step 2:

[1966] The terminal collects the user's voice in real time using a microphone installed in the vehicle. The specific input is the user's spoken voice, and the output is voice data. The collected voice data is used to extract features in the next step.

[1967] Step 3:

[1968] The device preprocesses the collected image data and extracts facial feature points. Specifically, it uses OpenCV to detect feature points such as eyes, mouth, and eyebrows from each frame of image. The input is the image data collected in step 1, and the output is the extracted facial feature point data.

[1969] Step 4:

[1970] The device preprocesses the collected audio data and extracts features. Specifically, Librosa is used to extract pitch, volume, and spectral characteristics from the audio data. The input is the audio data collected in step 2, and the output is the extracted audio feature data.

[1971] Step 5:

[1972] The terminal sends the preprocessed feature point data and feature volume data to the server. The specific input is the feature data obtained in step 3 and step 4, and the output is the data sent to the server.

[1973] Step 6:

[1974] The server inputs the received feature data into a generative AI model to infer the user's psychological state. The specific inputs are feature point data and voice feature data, and the output is the inferred psychological state (e.g., relaxed, tense, fatigued). The generative AI model achieves this using Google Cloud AI and AWS SageMaker.

[1975] Step 7:

[1976] The server generates instructions to adjust driving parameters based on the inferred psychological state, e.g., to reduce speed if the user is tired. The input is the psychological state inferred in step 6, and the output is the adjusted driving parameters.

[1977] Step 8:

[1978] The server sends the generated driving parameters or advice to the mobile device. For example, a break suggestion or a music playback instruction is generated in this step. The input is the driving parameters generated in step 7, and the output is the notification content.

[1979] Step 9:

[1980] The device displays or voices the received notification to the driver. Specifically, it displays the suggestion on the screen of the mobile device or informs the user as voice guidance. The input is the notification content sent in step 8, and the output is the display on the device or voice guidance.

[1981] Step 10:

[1982] The server sends the adjusted driving parameters to the automated driving system, and the vehicle performs driving based on these instructions. The input is the driving parameters generated in step 7, and the output is the actual vehicle behavior.

[1983] Example prompt sentence:

[1984] Prompt: "Analyze the driver's facial expressions and voice data to infer the driver's fatigue state and generate appropriate rest advice."

[1985] The above are the processing steps in the embodiment of the present invention.

[1986] 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.

[1987] System Overview

[1988] This invention is an autonomous driving system that uses a camera and microphone installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model and an emotion engine to infer the user's psychological state and adjust driving parameters. By combining this with an emotion engine, it becomes possible to more accurately recognize the user's diverse emotions and perform driving operations based on them.

[1989] Explanation of program processing

[1990] Data collection

[1991] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. A microphone also records what the user says in real time.

[1992] Data Preprocessing

[1993] Device: Extracts facial features (e.g., the position of the eyes, mouth, and eyebrows) from the collected image data, and extracts features (e.g., pitch, volume, and spectral characteristics) from the audio data, converting the data into a format that is easy for the AI ​​model described below to handle.

[1994] Sending data

[1995] Terminal: Sends preprocessed data (facial feature points and audio features) to the server.

[1996] Psychological analysis

[1997] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to infer the user's mental state and emotions, for example, identifying emotions such as tension, relief, surprise, anger, and joy.

[1998] Adjusting operating parameters

[1999] Server: Adjusts driving parameters (speed, acceleration, braking) based on the inferred psychological state and emotions. For example, if the user is angry, the server will slow down the vehicle to ensure safety and drive calmly.

[2000] Sending adjustment instructions

[2001] Server: Sends the adjusted driving parameters to the automated driving system. This transmission is usually done in real time, and is quickly reflected in the automated driving system inside the vehicle.

[2002] Execution of operation

[2003] Terminal: The terminal applies the new driving parameters received from the server and performs actual driving operations. This allows for safe and secure driving that is tailored to the user's psychological state.

[2004] Specific examples

[2005] Scenario 1: The user is angry

[2006] Data collection:

[2007] Device: The camera captures the user's angry facial expression (e.g., frown, closed mouth), and the microphone records the user's strong-tongued remarks.

[2008] Data preprocessing:

[2009] Device: Extract anger features from image data using the Dlib library, and extract high-pitch and high-volume features from audio data.

[2010] Psychological Analysis:

[2011] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is angry."

[2012] Adjusting operating parameters:

[2013] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[2014] Sending adjustment orders:

[2015] Server: Sends new driving parameters to the vehicle.

[2016] Execution of the operation:

[2017] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[2018] Scenario 2: Happy users

[2019] Data collection:

[2020] Device: The camera captures the user's smile, and the microphone records the user's conversation in a cheerful tone.

[2021] Data preprocessing:

[2022] Device: From the facial features, smiling features (such as the corners of the mouth turning up and the corners of the eyes turning up) are extracted, and stable pitch and volume features are extracted from the audio data.

[2023] Psychological Analysis:

[2024] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is happy."

[2025] Adjusting operating parameters:

[2026] Server: Maintain normal operating parameters or configure for slightly faster speeds.

[2027] Sending adjustment orders:

[2028] Server: Sends a command to the vehicle to continue normal operation.

[2029] Execution of the operation:

[2030] Terminal: The vehicle follows the command and maintains its current operating parameters.

[2031] The above is a specific embodiment of the automated driving system of the present invention. This allows driving to be performed in accordance with the diverse emotions and psychological states of passengers, thereby improving safety and peace of mind.

[2032] The processing flow will be explained below.

[2033] Specific processing steps of the system

[2034] Step 1:

[2035] Data collection

[2036] Terminal: A camera installed inside the vehicle captures the user's facial expressions frame by frame and collects image data. A microphone also records the user's speech in real time.

[2037] How it works: The camera takes pictures at 30 frames per second and stores them in storage. The microphone captures audio at high resolution of 24-bit, 96kHz and stores it in storage.

[2038] Step 2:

[2039] Data Preprocessing

[2040] Terminal: Extracts facial features (e.g., eye, mouth, and eyebrow positions) from collected image data, and extracts features (e.g., pitch, volume, and spectral characteristics) from audio data.

[2041] Specific operation: The Dlib library is used to identify the positions of the eyes, mouth, and eyebrows from the image data, and 68 feature points are extracted. MFCC (Mel-Frequency Cepstrum Coefficients) is used to extract 13-dimensional feature values ​​from the audio data every 0.1 seconds.

[2042] Step 3:

[2043] Sending data

[2044] Terminal: Sends preprocessed data (facial feature points and audio features) to the server.

[2045] Specific operation: Data is sent asynchronously to a server via the Internet or a dedicated in-vehicle network using HTTP or WebSocket protocols.

[2046] Step 4:

[2047] emotion recognition

[2048] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to recognize the user's emotions. The generative AI model estimates the psychological state, and the emotion engine identifies the emotion.

[2049] Specific behavior: The generative AI model analyzes the pre-processed data and outputs psychological states such as "tension," "relief," and "surprise." Based on this, the emotion engine classifies the user's specific emotions into "anger," "joy," "sadness," etc.

[2050] Step 5:

[2051] Adjusting operating parameters

[2052] Server: Determines driving parameters (speed, acceleration, braking) based on the inferred psychological state and emotions.

[2053] Specific behavior: Uses a rules engine to generate driving commands such as "slow down if the user is angry" or "maintain speed if the user is happy."

[2054] Step 6:

[2055] Sending adjustment instructions

[2056] Server: Sends the generated new driving parameters to the autonomous driving system.

[2057] Specific operation: Driving parameters are sent to the terminal in the vehicle using HTTP or a dedicated protocol.

[2058] Step 7:

[2059] Execution of operation

[2060] Terminal: Apply the new driving parameters received from the server to the autonomous driving system.

[2061] Specific behavior: The autonomous driving system uses the new control parameters to perform driving maneuvers such as speed, acceleration, and braking. For example, if the user is nervous, the system will reduce the speed to 20 km / h.

[2062] Specific examples

[2063] Scenario 1: The user is angry

[2064] Step 1:

[2065] Data collection

[2066] Device: The camera captures the user's angry facial expression (frown, tight-lipped mouth), and the microphone records the user's strong-tongued remarks.

[2067] Step 2:

[2068] Data Preprocessing

[2069] Terminal: Use the Dlib library to extract facial features, and high-pitch and high-volume features from audio data.

[2070] Step 3:

[2071] Sending data

[2072] Terminal: Sends preprocessed data to the server.

[2073] Step 4:

[2074] emotion recognition

[2075] Server: The generative AI model and emotion engine analyze the data and recognize that the user is feeling "anger."

[2076] Step 5:

[2077] Adjusting operating parameters

[2078] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[2079] Step 6:

[2080] Sending adjustment instructions

[2081] Server: Sends new driving parameters to the vehicle.

[2082] Step 7:

[2083] Execution of operation

[2084] Terminal: Reduce speed to 20km / h and begin driving smoothly.

[2085] Scenario 2: Happy users

[2086] Step 1:

[2087] Data collection

[2088] Device: The camera captures the user's smile. The microphone records the user's pleasant conversation.

[2089] Step 2:

[2090] Data Preprocessing

[2091] Device: Extracts features of a smile, such as eyes and mouth turned up, and extracts stable pitch and volume features from audio data.

[2092] Step 3:

[2093] Sending data

[2094] Terminal: Sends preprocessed data to the server.

[2095] Step 4:

[2096] emotion recognition

[2097] Server: The generative AI model and emotion engine analyze the data and recognize that the user is feeling "joy."

[2098] Step 5:

[2099] Adjusting operating parameters

[2100] Server: Issues a command to continue normal operation.

[2101] Step 6:

[2102] Sending adjustment instructions

[2103] Server: Sends commands to the vehicle.

[2104] Step 7:

[2105] Execution of operation

[2106] Terminal: Maintain current operating parameters and continue normal operation.

[2107] This will enable safe and secure driving based on specific emotions.

[2108] Example 2

[2109] 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."

[2110] In an automated driving system for a vehicle, there is a demand for providing a safer and more comfortable driving experience by accurately grasping the user's psychological state and adjusting driving parameters accordingly. However, with conventional technologies, it has been difficult to accurately recognize the user's emotions and psychological state in real time and dynamically adjust driving parameters based on that. This invention is intended to solve this problem.

[2111] 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.

[2112] In this invention, the server, in a vehicle linked to an automated driving system that estimates a user's psychological state and adjusts driving parameters, includes: means for collecting a user's facial expression in real time using an internally installed image capture device; means for collecting the user's voice in real time using an internally installed voice recording device; means for preprocessing the collected facial expression data and voice data and extracting feature points and feature quantities; means for estimating the user's psychological state from the extracted feature data using a generative technology model; means for adjusting the driving parameters based on the estimated psychological state; and means for transmitting the adjusted driving parameters to the automated driving system and applying them, thereby enabling appropriate adjustment of driving parameters based on the user's real-time psychological state.

[2113] A "user" is a person who uses this system.

[2114] "Mental state" refers to the user's feelings and mental state, and specifically includes emotions such as tension, relief, surprise, anger, and joy.

[2115] "Driving parameters" are control data related to vehicle driving, specifically settings such as speed, acceleration, and braking.

[2116] An "autonomous driving system" is a system that automatically controls the driving of a vehicle and has the function of allowing the vehicle to drive autonomously without the user performing any driving operations.

[2117] The "photography device" is a device that is installed inside the vehicle and captures the user's facial expressions in real time.

[2118] The "voice recording device" is a device that is installed inside the vehicle and records the user's speech in real time.

[2119] "Preprocessing" refers to the process of extracting feature points and feature quantities from collected raw data (image data and audio data).

[2120] "Feature points" refer to characteristic points on the user's face extracted from image data, such as the positions of the eyes, mouth, and eyebrows.

[2121] "Features" are acoustic features extracted from audio data, specifically pitch, volume, spectral characteristics, etc.

[2122] A "generative model" is an artificial intelligence model used to infer a user's psychological state from collected data.

[2123] "Adjusted driving parameters" refer to new driving parameters that have been adjusted based on the user's psychological state.

[2124] "Application" means reflecting the adjusted driving parameters in the actual automated driving system.

[2125] System Overview

[2126] This invention is an autonomous driving system that uses a camera and microphone installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model and emotion engine to infer the user's psychological state and adjust driving parameters. By combining this with an emotion engine, it is possible to more accurately recognize the user's diverse emotions and perform driving operations based on them.

[2127] Data collection

[2128] Terminal: A camera installed inside the vehicle captures frame-by-frame images of the user's face and collects image data. An audio recording device also records the user's speech in real time.

[2129] Hardware used:

[2130] Camera (e.g. high-resolution camera)

[2131] Microphone (e.g. high-sensitivity microphone)

[2132] Data Preprocessing

[2133] Device: Facial feature points (e.g., eye, mouth, and eyebrow positions) are extracted from the collected image data using a face recognition library (e.g., Dlib), while acoustic features (e.g., pitch, volume, and spectral characteristics) are extracted from the audio data.

[2134] Software used:

[2135] Dlib library (for face recognition)

[2136] Audio analysis tools (e.g. FFT)

[2137] Sending data

[2138] Terminal: The facial feature point data and voice feature data obtained in the preprocessing are encrypted and sent to the server using the HTTPS protocol to ensure data security.

[2139] Psychological analysis

[2140] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to infer the user's psychological state. For example, identify emotions such as tension, relief, surprise, anger, and joy.

[2141] Software used:

[2142] Generative technology models (e.g., BERT-based sentiment analysis models)

[2143] Emotion Engine

[2144] Adjusting operating parameters

[2145] Server: Optimizes driving parameters (speed, acceleration, braking) based on the estimated psychological state. If the user is nervous, the server reduces speed and smooths braking to make the driver more calm.

[2146] Sending adjustment instructions

[2147] Server: Transmits the calculated driving parameters to the vehicle's autonomous driving system in real time. This transmission also uses the HTTPS protocol to ensure data integrity and security.

[2148] Execution of operation

[2149] Terminal: The terminal applies the new driving parameters received from the server to the actual vehicle control and performs driving, thereby automatically driving according to the user's psychological state.

[2150] Specific examples

[2151] Scenario 1: The user is angry

[2152] Data collection:

[2153] Device: The camera captures the user's angry facial expression (e.g., frown, closed mouth), and the microphone records the user's strong-tongued remarks.

[2154] Data preprocessing:

[2155] Device: Extract anger features from image data using the Dlib library, and extract high-pitch and high-volume features from audio data.

[2156] Psychological Analysis:

[2157] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is angry."

[2158] Adjusting operating parameters:

[2159] Server: Generates commands to slow down the vehicle and avoid sudden maneuvers.

[2160] Sending adjustment orders:

[2161] Server: Sends new driving parameters to the vehicle.

[2162] Execution of the operation:

[2163] Terminal: The vehicle obeys the command, slows down to 20km / h, and begins driving smoothly.

[2164] Scenario 2: Happy users

[2165] Data collection:

[2166] Device: The camera captures the user's smile, and the microphone records the user's conversation in a cheerful tone.

[2167] Data preprocessing:

[2168] Device: From the facial features, smiling features (such as the corners of the mouth turning up and the corners of the eyes turning up) are extracted, and stable pitch and volume features are extracted from the audio data.

[2169] Psychological Analysis:

[2170] Server: The generative AI model and emotion engine analyze these feature data and infer that "the user is happy."

[2171] Adjusting operating parameters:

[2172] Server: Maintain normal operating parameters or configure for slightly faster speeds.

[2173] Sending adjustment orders:

[2174] Server: Sends a command to the vehicle to continue normal operation.

[2175] Execution of the operation:

[2176] Terminal: The vehicle follows the command and maintains its current operating parameters.

[2177] Prompt Sentence Examples

[2178] Scenario 1 prompt statement:

[2179] If the user is angry, please follow the instructions below to adjust the operating parameters:

[2180] Angry facial expression captured on camera

[2181] Logged voice data with strong verbal content

[2182] Scenario 2 prompt statement:

[2183] If the user is willing, please follow the instructions below to adjust the operating parameters:

[2184] A photo of a smiling face captured on camera

[2185] Logged voice data with a pleasant tone

[2186] The flow of the identification process in the second embodiment will be described with reference to FIG.

[2187] Step 1: Data collection

[2188] Device: The in-car camera captures the user's face frame by frame and collects image data, while the microphone records what the user says in real time.

[2189] Input: Video frames of the user's face, audio of what the user is saying

[2190] Output: User's face image data, voice data

[2191] Specific behavior:

[2192] The interior camera captures images at 30 frames per second.

[2193] The microphone records audio at a sample rate of 16kHz.

[2194] Step 2: Data Preprocessing

[2195] Device: Facial feature points (e.g., eye, mouth, and eyebrow positions) are extracted from the collected image data using a face recognition library (e.g., Dlib), while acoustic features (e.g., pitch, volume, and spectral characteristics) are extracted from the audio data.

[2196] Input: Facial image data, audio data

[2197] Output: Facial feature data, acoustic feature data

[2198] Specific behavior:

[2199] Using the Dlib library, 68 facial feature points are extracted from each frame.

[2200] The frequency components of the audio data are analyzed using FFT (Fast Fourier Transform) and extracted as features.

[2201] Step 3: Sending data

[2202] Terminal: The facial feature point data and voice feature data obtained in the preprocessing are encrypted and sent to the server using the HTTPS protocol to ensure data security.

[2203] Input: facial feature point data, acoustic feature data

[2204] Output: Notification that encrypted data has been sent

[2205] Specific behavior:

[2206] Package the data in JSON format and send a POST request to the server.

[2207] Step 4: Psychological analysis

[2208] Server: Inputs the received preprocessed data into the generative AI model and emotion engine to infer the user's psychological state. For example, identify emotions such as tension, relief, surprise, anger, and joy.

[2209] Input: Preprocessed data (facial feature points, acoustic features)

[2210] Output: User's mental state and emotion category

[2211] Specific behavior:

[2212] Input feature data into a generative AI model (e.g., a BERT-based sentiment analysis model) to predict sentiment categories.

[2213] The emotion engine analyzes the prediction results and determines the most appropriate emotion.

[2214] Step 5: Adjusting the operating parameters

[2215] Server: Optimizes driving parameters (speed, acceleration, braking) based on the estimated psychological state. If the user is nervous, the server reduces speed and smooths braking to make the driver more calm.

[2216] Input: User's mental state and emotion category

[2217] Output: Adjusted operating parameters

[2218] Specific behavior:

[2219] It obtains driving data such as current speed and acceleration and calculates new driving parameters based on that.

[2220] Step 6: Sending adjustment instructions

[2221] Server: Transmits the calculated driving parameters to the vehicle's autonomous driving system in real time. This transmission also uses the HTTPS protocol to ensure data integrity and security.

[2222] Input: Adjusted operating parameters

[2223] Output: Notification of completion of sending of operation parameters

[2224] Specific behavior:

[2225] The new driving parameters are packaged in JSON format and sent as a POST request to the vehicle's control system.

[2226] Step 7: Execute the operation

[2227] Terminal: The terminal applies the new driving parameters received from the server to the actual vehicle control and performs driving, thereby automatically driving according to the user's psychological state.

[2228] Input: New operating parameters

[2229] Output: Vehicle control actions based on driving parameters

[2230] Specific behavior:

[2231] The vehicle's ECU (Electronic Control Unit) receives the new driving parameters and reflects them in brake and accelerator control.

[2232] (Application example 2)

[2233] 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."

[2234] Conventional automated driving systems operate based on predetermined driving parameters without considering the user's emotions or psychological state, and therefore cannot fully guarantee the user's safety or comfort. Furthermore, they lack a mechanism for providing support content that adapts to the user's psychological state while driving, making it difficult to promote stress reduction and relaxation for the user. Therefore, there is a need for a system that can adjust driving parameters and provide support content appropriately according to the user's psychological state.

[2235] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting the user's facial expressions in real time using an internally installed camera, means for collecting the user's voice in real time using an internally installed microphone, means for preprocessing the collected facial expression data and voice data to extract feature points and feature quantities, means for inferring the user's psychological state from the extracted feature data using a generative AI model, means for adjusting driving parameters based on the inferred psychological state, means for transmitting the adjusted driving parameters to the automated driving system and applying them, and means for providing support content during driving in accordance with the user's emotional state. This makes it possible to adjust driving parameters in accordance with the user's psychological state and provide appropriate support content.

[2236] An "autonomous driving system that performs driving operations" is a system that automatically drives a vehicle, and is a general term for the programs and hardware that perform driving operations such as adjusting speed, changing direction, and braking.

[2237] The "interiorly installed camera" is a camera device that is attached to the interior of the vehicle and captures the user's facial expressions in real time.

[2238] The "interiorly installed microphone" is a microphone device that is installed inside the vehicle and collects the user's voice in real time.

[2239] "Facial expression data" is image data of the user's face captured by a camera, and includes facial feature points.

[2240] "Voice data" refers to voice data of a user's speech recorded by a microphone, and includes voice features.

[2241] "Preprocessing" refers to a data processing procedure for extracting feature points and feature quantities from collected data.

[2242] "Feature points" are data points relating to the position and shape of a face, and indicate specific positions such as the eyes, mouth, and eyebrows.

[2243] "Features" are specific numerical data such as pitch, volume, and spectral characteristics extracted from audio data.

[2244] A "generative AI model" is an artificial intelligence model that analyzes collected data and infers the user's psychological state.

[2245] The "emotion engine" is an analytical engine that identifies the user's emotions and generates appropriate responses based on them.

[2246] "Driving parameters" are set values ​​for controlling the driving of a vehicle, and include speed, acceleration, braking, and the like.

[2247] "Support content" refers to supplementary information and entertainment provided while driving according to the user's psychological state, and includes relaxing music, positive news, and the like.

[2248] This invention is an autonomous driving system that uses a camera and microphone installed inside the vehicle to collect the user's facial expressions and voice in real time, and uses a generative AI model and emotion engine to infer the user's psychological state and adjust driving parameters.Furthermore, it is a system that provides support content according to the user's psychological state.

[2249] System Overview

[2250] Hardware used

[2251] Camera: An image capturing device installed inside the vehicle.

[2252] Microphone: An audio collection device installed inside the vehicle.

[2253] Smart glasses: A wearable device for capturing the user's facial expressions and voice in real time.

[2254] Minicomputer: A small computer (e.g., Raspberry Pi) for data preprocessing and analysis.

[2255] On-board computer: An on-board computer used to adjust the vehicle's operating parameters.

[2256] Software used

[2257] Facial expression recognition library: Dlib (a library for extracting facial feature points).

[2258] Speech analysis library: LibROSA (a library for extracting speech features).

[2259] Generative AI models: Natural language processing models such as GPT-3.

[2260] Sentiment engine: A sentiment analysis engine such as Azure Cognitive Services.

[2261] Specific Embodiments

[2262] Data collection

[2263] The device (smart glasses) captures the user's facial expressions in real time with a camera installed inside the vehicle and records the user's voice with a microphone, and these data are sent to a minicomputer.

[2264] Data Preprocessing

[2265] The minicomputer uses Dlib to extract facial features from image data, and LibROSA to extract audio features such as pitch, volume, and spectral characteristics from audio data.

[2266] Psychological analysis

[2267] The server inputs the preprocessed data into a generative AI model and emotion engine to infer the user's psychological state, and determines whether the user is happy, angry, or other emotional state based on the inference.

[2268] Providing support content

[2269] If the user is inferred to be angry, the following prompt sentence can be input into the generative AI model to generate instructions to provide relaxing music:

[2270] "User's facial expression data: {facial expression data}

[2271] User's voice data: {voice data}

[2272] Inferred emotion: Anger

[2273] Suggested action: Play relaxing music

[2274] Based on this instruction, the appropriate music will be played.

[2275] If the user is deemed happy, the generative AI model is given the following prompt:

[2276] "User's facial expression data: {facial expression data}

[2277] User's voice data: {voice data}

[2278] Inferred emotion: Joy

[2279] Recommended action: Read positive news

[2280] This will read out positive news aloud.

[2281] Adjusting operating parameters

[2282] The server adjusts driving parameters (speed, braking, etc.) based on the inferred psychological state. For example, if the user is angry, it generates an instruction to slow down the vehicle and sends it to the on-board computer. This automatically adjusts the vehicle's driving, increasing the user's safety.

[2283] Example

[2284] Scenario 1: The user is angry

[2285] Data collection:

[2286] The camera captures the user's angry expression (frowns, closed mouth), and the microphone records the strong words spoken.

[2287] Sentiment Analysis and Response:

[2288] The generative AI model and emotion engine infer that the user is angry and generate instructions to provide relaxing music.

[2289] Adjusting operating parameters:

[2290] The server generates and transmits to the vehicle instructions to reduce the vehicle's speed.

[2291] Scenario 2: Happy users

[2292] Data collection:

[2293] The camera captures the user's smile, and the microphone records conversations in a playful tone.

[2294] Sentiment Analysis and Response:

[2295] The generative AI model and emotion engine infer that the user is happy and generate instructions to deliver positive news.

[2296] Adjusting operating parameters:

[2297] The server generates and transmits instructions to the vehicle to maintain normal operating parameters.

[2298] This will enable the adjustment of driving parameters according to the user's psychological state and the provision of appropriate support content, resulting in safer and more comfortable automated driving.

[2299] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[2300] Step 1:

[2301] The device (smart glasses) captures images of the user's face in real time using an internal camera and acquires image data. It also records the user's voice in real time using an internal microphone and acquires voice data. This input data (image data and voice data) is sent to a minicomputer.

[2302] Step 2:

[2303] The minicomputer extracts facial features (e.g., eye, mouth, and eyebrow positions) from the received image data using the Dlib library, and simultaneously extracts pitch, volume, and spectral characteristics from the audio data using the LibROSA library, resulting in preprocessed data (facial features and audio features).

[2304] Step 3:

[2305] The preprocessed data is sent to the server. The input data (facial features and voice features) is fed into the generative AI model and emotion engine to infer the user's psychological state. The generative AI model analyzes the combined facial and voice data and identifies emotions such as "the user is angry" or "the user is happy."

[2306] Step 4:

[2307] The server adjusts driving parameters based on the estimated psychological state. Specifically, it generates instructions to set driving parameters such as vehicle speed, acceleration, and braking. For example, if it estimates that the user is angry, it generates an instruction to slow down the vehicle.

[2308] Step 5:

[2309] The server sends the generated instructions to the vehicle's on-board computer. The input data (adjusted driving parameters) are applied by the on-board computer and reflected in actual driving operations. This allows driving to be performed in accordance with the user's psychological state.

[2310] Step 6:

[2311] The server provides support content according to the user's psychological state. For example, if the user is angry, the server generates an instruction to play relaxing music and sends it to the minicomputer. The minicomputer then plays the relaxing music according to the instruction.

[2312] Step 7:

[2313] If the user's psychological state is calmed by the support content they receive, the camera and microphone will again collect their facial expressions and voice, and the entire system will start operating again, making it possible to provide continuous driving support that responds to the user's psychological state.

[2314] As a result, safe and comfortable automated driving is achieved by adjusting driving parameters according to the user's psychological state and providing support content.

[2315] 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.

[2316] 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.

[2317] 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 robot 414.

[2318] 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.

[2319] 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.

[2320] 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.

[2321] 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).

[2322] 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.

[2323] 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."

[2324] 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.

[2325] The sy...

Claims

1. In vehicles that work in conjunction with automated driving systems that perform driving operations, A means of collecting the user's facial expressions in real time using a camera installed inside, A means for collecting the user's voice in real time using an internally installed microphone; A means for preprocessing the collected facial expression data and voice data to extract feature points and feature quantities; A means for inferring a user's psychological state from feature data extracted using a generative AI model; means for adjusting driving parameters based on the estimated psychological state; means for transmitting and applying the adjusted driving parameters to the automated driving system; A system including:

2. The means of collecting the user's facial expressions in real time using a camera installed inside is Capture image data frame by frame and detect facial feature points. The system of claim 1 .

3. The means of collecting the user's voice in real time using a microphone installed inside is Extracting pitch, volume, and spectral characteristics from audio data, The system of claim 1 .

Images