Method and device for controlling intelligent sound production system to relieve carsickness based on electroencephalogram signals

Abstract

The invention provides a method and device for controlling an intelligent sound production system to relieve carsickness based on electroencephalogram signals. The method comprises the following steps that S1, electroencephalogram signal data of passengers in the carsickness process are collected; step S2, preprocessing the electroencephalogram signals, and extracting electroencephalogram characteristics including frequency bands of # imgabs0 # waves, # imgabs1 # waves, # imgabs2 # waves and # imgabs3 # waves; s3, calculating a brain wave comprehensive variation index CVEI as an objective evaluation index of the carsickness grade based on the extracted electroencephalogram characteristics; s4, performing carsickness grade classification according to the brain wave comprehensive variation index CVEI of the passenger by using a carsickness grade identification model, wherein the carsickness grade identification model is constructed by fusing a convolutional neural network and an attention mechanism; and step S5, according to the carsickness grade classification result of the passenger, playing the audio which corresponds to the carsickness grade and is used for relieving carsickness in the audio sample library. The garment can effectively relieve carsickness of passengers.

Description

Technical Field

[0001] The present invention relates to the technical field of automobiles, in particular to a method and device for relieving motion sickness by controlling an intelligent sound system based on electroencephalogram (EEG) signals. Background Art

[0002] Due to the narrow space of the enclosed cabin and the drastic changes in the postures of the occupants, the motion perceived by the visual system is inconsistent with the balance perception of the vestibular system, which is extremely likely to induce motion sickness, seriously affecting the driving and riding comfort, and further threatening the driving safety and mission execution reliability of new energy vehicles. Therefore, in order to create a comfortable driving and riding environment and meet the increasingly prominent intelligent requirements, it is urgent to explore new anti-motion sickness control methods for new energy vehicles.

[0003] Currently, common methods for relieving motion sickness mainly include drug intervention, physical therapies (such as bracelets, massages), and non-drug means such as voice or music therapies. However, traditional intervention methods often rely on the user's self-perception or static settings, lacking real-time monitoring and dynamic adjustment of the motion sickness state, and it is difficult to achieve personalized and intelligent intervention effects.

[0004] As a recognized effective physiological index, EEG signals can reflect brain cognitive activities in real time and objectively describe the changes in people's subjective perceptions. In recent years, with the rapid development of brain-computer interface technology, it has become possible to introduce EEG physiological signals and study motion sickness during driving and riding from a physiological perspective. Mapping the vocabulary items in the currently commonly used motion sickness evaluation scale to objective EEG indicators, and accurately and quickly analyzing the motion sickness perception state of the driver and rider through EEG signals, providing a new research idea for realizing anti-motion sickness control of new energy vehicles from the perspective of neuroscience. Summary of the Invention

[0005] The purpose of the present invention is to propose a method and device for relieving motion sickness by controlling an intelligent sound system based on EEG signals, which can be used to relieve the motion sickness of occupants and is mainly for use by the adult population.

[0006] To achieve the above purpose, the technical solution of the present invention is as follows:

[0007] The present invention proposes a method for relieving motion sickness by controlling an intelligent sound system based on EEG signals, and the method includes the following steps;

[0008] Step S1, collecting EEG signal data of the occupant during motion sickness;

[0009] Step S2, preprocessing the EEG signals and extracting EEG features including the frequency bands of δ waves, θ waves, α waves, and β waves;

[0010] Step S3, calculating the comprehensive variation index of electroencephalogram CVEI based on the extracted EEG features as an objective evaluation index for the motion sickness level;

[0011] Step S4: Use the motion sickness level recognition model to classify the motion sickness level according to the comprehensive variation index of electroencephalogram (CVEI) of the occupant. The motion sickness level recognition model is constructed by integrating a convolutional neural network and an attention mechanism.

[0012] Step S5: According to the classification result of the occupant's motion sickness level, play the audio for relieving motion sickness corresponding to the motion sickness level in the audio sample library.

[0013] [[ID=⑧]]Preferably, the electroencephalogram (EEG) signals of the occupant during the ride are collected in real time by a wearable portable EEG acquisition device. The acquisition device supports multi-channel signal acquisition, with a sampling rate of not less than 128 Hz, and covers at least the prefrontal and parietal regions. The wearable portable EEG acquisition device includes a head-mounted EEG instrument.

[0014] Preferably, step S2 specifically includes:

[0015] S2.1: Preprocess the EEG signal data, including filtering, artifact removal, data extraction by segmenting with the stimulation point, and normalization processing.

[0016] S2.2: Obtain the preprocessed multi-channel and multi-segment EEG signal data. Each segment of EEG signal data for each channel is sampled by a sliding window method to obtain the time-domain EEG signals of different time windows for each channel:

[0017] X (c) ={X1 (c) ,X2 (c) ,...,X N (c)}

[0018] where X (c) represents the time-domain EEG signal of a time window for channel c, c = 1, 2,..., C, C is the total number of channels, and N represents the number of sampling points in a time window.

[0019] S2.3: For the time-domain EEG signal of each time window for each channel, use the fast Fourier transform (FFT) to calculate the power spectral density of different frequency bands, and obtain the power spectral densities P δ (c) ,P θ (c) ,P α (c) ,P β (c) .

[0020] Preferably, step S3 specifically includes:

[0021] It should be noted that the symbol "⑧" in the original text seems to be an incorrect numbering. It is translated as "⑧" here for the sake of consistency with the original text. You may need to check and correct it if necessary.S3.1. Calculate the variance of the waveform change of the time-domain EEG signal for each channel in each time window:

[0022]

[0023] where VAR (c) EEG is the variance of the waveform change of the time-domain EEG signal for a time window of channel c, is the mean of X (c) , and X i (c) represents the data of the i-th sampling point in X (c) ;

[0024] S3.2. Calculate the fusion feature CVEI for each channel in each time window:

[0025]

[0026] where CVEI (c) is the CVEI value for a time window of channel c, and ε is a constant used to prevent the denominator from being zero. S3.3. Statistically fuse the CVEI values of all channels corresponding to the same time window to obtain the overall CVEI index of the whole brain for different time windows:

[0027]

[0028] CVEI total represents the overall CVEI index of the whole brain for a time window.

[0029] Preferably, the training of the motion sickness level recognition model is specifically as follows:

[0030] Design a motion sickness induction experiment and collect EEG signals during motion sickness; preprocess the EEG signals and extract EEG features including the δ-wave, θ-wave, α-wave, and β-wave frequency bands;

[0031] Calculate the comprehensive variation index of brain waves CVEI as an objective evaluation index of motion sickness level based on the extracted EEG features, and combine the subjective questionnaire scores to divide the motion sickness levels, and construct a training set to train the motion sickness level classification model.

[0032] Preferably, the subjective questionnaire scores are divided into levels from 0 to 10 as follows:

[0033]

[0034] The output of the motion sickness level recognition model includes four levels: no motion sickness, mild motion sickness, moderate motion sickness, and severe motion sickness:

[0035]

[0036] Preferably, the construction of the audio sample library is specifically as follows:

[0037] Based on literature research, online materials or user experiments, collect audio sound samples with the effect of alleviating motion sickness; subjectively evaluate the motion sickness alleviation effect of the collected audio sound samples, and calculate the motion sickness alleviation effect value η:

[0038] According to the η value, divide the audio sound samples into four categories of audio with different levels of motion sickness alleviation effects, and obtain an audio sample library for alleviating motion sickness.

[0039] Preferably, the calculation of the motion sickness alleviation effect value η is specifically as follows:

[0040]

[0041] In the formula, m B , m A are the subjective scores before and after alleviating motion sickness, respectively.

[0042] Preferably, the audio playback adopts a dynamic adjustment method, including volume progression and playback interval change, and combines the subsequent electroencephalogram fluctuation situation for feedback optimization to form a closed-loop intervention control.

[0043] The present invention also proposes a device for alleviating motion sickness based on an electroencephalogram signal-controlled intelligent sound system. The device is implemented by using any one of the above methods for alleviating motion sickness based on an electroencephalogram signal-controlled intelligent sound system, and includes the following modules:

[0044] Electroencephalogram acquisition module: Real-time collect the electroencephalogram signals of the occupant during the vehicle ride;

[0045] Signal processing module: Preprocess the electroencephalogram signals and extract the electroencephalogram features including the frequency bands of δ wave, θ wave, α wave and β wave;

[0046] Evaluation module: Calculate the comprehensive electroencephalogram variation index CVEI based on the extracted electroencephalogram features;

[0047] Motion sickness recognition module: Use the motion sickness level recognition model to classify the motion sickness level according to the occupant's comprehensive electroencephalogram variation index CVEI;

[0048] Audio control module: Call the audio samples of the matching level according to the classification result of the occupant's motion sickness level;

[0049] Intelligent sound module: Output the audio samples to intervene and alleviate the discomfort of motion sickness.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] 1. This application adopts a non-invasive identification method, uses a wearable electroencephalogram device to achieve real-time and convenient monitoring of an individual's physiological state, and improves the use comfort and adaptability. The device structure proposed in this application is compact and can be embedded in portable electroencephalogram devices, vehicle-mounted systems or wearable hardware, and is applicable to various motion sickness intervention scenarios such as intelligent cockpits, virtual reality, and flight simulation, with broad application prospects.

[0052] 2. By introducing multi-modal index fusion technologies such as the Comprehensive Variation Index of Electroencephalogram (CVEI) and spectral analysis, the recognition accuracy of motion sickness status can be significantly improved.

[0053] The newly proposed "Comprehensive Variation Index of Electroencephalogram (CVEI)", as a composite physiological index for judging motion sickness status, comprehensively considers the changes in the power spectral density of the δ wave, θ wave, α wave, and β wave frequency bands in the electroencephalogram signal, and combines the time-domain volatility of the signal (i.e., the variance of changes), and can more accurately reflect the neural activation and fluctuations in the brain during motion sickness. Compared with the traditional single-frequency band electroencephalogram power value, CVEI has stronger expression ability in characterizing the complexity of neural activities.

[0054] 3. This application constructs a motion sickness level recognition model based on deep learning algorithms. CVEI is used as an important input feature and combined with a convolutional neural network and an attention mechanism to achieve intelligent classification of no motion sickness, mild, moderate, and severe states, significantly improving the recognition accuracy and stability of the model in the case of few samples, and having strong generalization ability and robustness.

[0055] 4. An intelligent grading intervention system is constructed, which can rely on machine learning algorithms to quantitatively evaluate the degree of motion sickness at four levels, and achieve intelligent grading intervention based on the recognition results. It can accurately match the intervention audio content according to the individual's motion sickness level, effectively improving the motion sickness relief effect and user comfort.

[0056] 5. The device adopts a modular design architecture, integrates a microprocessor, is compact in volume and convenient for vehicle-mounted deployment, meeting personalized needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a schematic flow chart of the method for relieving motion sickness by controlling an intelligent sound system based on electroencephalogram signals according to the present invention;

[0058] Figure 2 It is a schematic framework flow chart of the device for relieving motion sickness by controlling an intelligent sound system based on electroencephalogram signals according to the present invention;

[0059] Figure 3 It is a schematic diagram of the preprocessing waveform of electroencephalogram signals;

[0060] Figure 4 It is a schematic diagram of the CVEI calculation process according to the present invention;

[0061] Figure 5 Schematic diagram for comparing EEG features under different motion sickness levels

[0062] Figure 6 Schematic diagram of the audio matching and control logic flow of the present invention Specific embodiments

[0063] The following combines the append Figure 1-6 ices to specifically describe the technical solutions of the present invention

[0064] The present invention proposes a method for relieving motion sickness by controlling an intelligent sound - generating system based on EEG signals. As Figure 1 shown, the method includes the following steps

[0065] Step S1: Collect EEG signal data of the occupant during motion sickness

[0066] Step S2: Pre - process the EEG signals and extract EEG features including the frequency bands of δ - wave, θ - wave, α - wave, and β - wave

[0067] Step S3: Calculate the comprehensive variation index of brain waves CVEI based on the extracted EEG features as an objective evaluation index for motion sickness level

[0068] Refer to Figure 4 and Figure 5 . This step analyzes the pre - processed EEG signals, extracts representative frequency - domain and time - domain features, and then constructs a composite index - - the comprehensive variation index of brain waves (CVEI) - - for characterizing the change of motion sickness state

[0069] Step S4: Use the motion sickness level recognition model to classify the motion sickness level according to the occupant's comprehensive variation index of brain waves CVEI. The motion sickness level recognition model is constructed by integrating a convolutional neural network and an attention mechanism

[0070] Step S5: According to the classification result of the occupant's motion sickness level, play the audio for relieving motion sickness corresponding to the motion sickness level in the audio sample library. Refer to the audio matching and control logic flow in Figure 6 .

[0071] In this embodiment, an EEG signal of the occupant during the ride is collected in real - time by a wearable portable EEG acquisition device (such as a head - mounted EEG device). The acquisition device supports multi - channel signal acquisition (such as 20 - 128 channels), the sampling rate is not less than 128 Hz, and at least covers the prefrontal and parietal regions, so as to reflect the changes in psychological states such as the occupant's attention, tension, and fatigue

[0072] In this embodiment, the specific content of step S2 includes

[0073] S2.1, Reference Figure 3 , preprocess the EEG signal data, including filtering, artifact removal, segmenting and extracting data based on the stimulation points, and normalization processing;

[0074] In this embodiment, band-pass filtering with a frequency range of 0.1 - 40 Hz is used for filtering to remove power frequency interference and high-frequency noise; then independent component analysis (ICA) is used for artifact removal; segmenting and extracting data based on the stimulation points means marking the stimulation points of the EEG signal data and extracting the fluctuating data before and after the marks to improve the accuracy of subsequent analysis; finally, normalization processing is performed to improve the accuracy of subsequent analysis.

[0075] S2.2, Obtain the preprocessed multi-channel and multi-segment EEG signal data. Each segment of EEG signal data for each channel is segmented and sampled at data points in a sliding window manner to obtain the time-domain EEG signals of different time windows for each channel:

[0076] X (c) ={X1 (c) , X2 (c) ,..., X N (c)}

[0077] where X (c) represents the time-domain EEG signal of a time window for channel c, c = 1, 2,..., C, C is the total number of channels, and N represents the number of sampling points in a time window; assuming each time window is 2 seconds and the sampling rate is 256 Hz, then the length of each segment of data is 512 points, N = 512;

[0078] S2.3, Frequency-domain feature extraction: For the time-domain EEG signal of each time window for each channel, use Butterworth filtering to divide the data into 5 bands, divided into the δ wave (0.5 - 4 Hz), θ wave (4 - 8 Hz), α wave (8 - 12 Hz), β wave (12 - 30 Hz) four frequency bands; use the fast Fourier transform (FFT) to calculate the power spectral density (PSD) of different frequency bands to obtain the power spectral densities P δ (c) , P θ (c) , P α (c) , P β (c) .

[0079] In this embodiment, the specific steps of step S3 include:

[0080] S3.1. Time-domain feature calculation: Calculate the variance (Signal Variance) of the waveform changes of the time-domain EEG signals for each channel in each time window, which is used to characterize the overall volatility of the neural signals:

[0081]

[0082] where VAR (c) EEG is the variance of the waveform changes of the time-domain EEG signals in a time window of channel c, reflecting the signal stability; is the mean of X (c) , and X i (c) represents the data of the i-th sampling point in X (c) ;

[0083] S3.2. Calculate the fusion feature CVEI for each channel in each time window:

[0084]

[0085] where CVEI (c) is the CVEI value of a time window of channel c, and ε is a small constant (such as 10 -6 ) used to prevent the denominator from being zero;

[0086] The above formula combines:

[0087] Slow wave enhancement (θ, δ): Usually appears enhanced during motion sickness or fatigue;

[0088] Fast wave suppression (α, β): More significant in the awake and attentive states;

[0089] Increased volatility (VAR): The instability of neural signals may reflect the stress state of the nervous system.

[0090] S3.3. Statistically fuse the CVEI values of all channels corresponding to the same time window (such as taking the average or maximum value) to obtain the whole-brain overall CVEI index for different time windows:

[0091]

[0092] CVEI total represents the whole-brain overall CVEI index for a time window; this index can comprehensively reflect the sympathetic-parasympathetic activity, the degree of neural tension and its change trend, and has a correlation with the degree of motion sickness.

[0093] The final result can be sent into the motion sickness level recognition model as a key physiological index, or can be used alone to judge whether the current brain is in a motion sickness abnormal state.

[0094] In this embodiment, the training of the motion sickness level recognition model is specifically as follows:

[0095] Design a motion sickness induction experiment and collect electroencephalogram (EEG) signals during motion sickness; preprocess the EEG signals and extract EEG features including the frequency bands of delta waves, theta waves, alpha waves, and beta waves;

[0096] Calculate the comprehensive variation index of electroencephalogram (CVEI) based on the extracted EEG features as an objective evaluation index of the motion sickness level, and combine the subjective questionnaire scores (such as MSQ or Likert scale 1-10 scores) to classify (label) the motion sickness level, and construct a training set to train the motion sickness level classification model, so as to perform multi-level classification and recognition of the motion sickness state by constructing a motion sickness level classification model based on deep learning;

[0097] The motion sickness level classification model is a structure that combines a convolutional neural network (CNN) and an attention mechanism. The CNN is used to extract spatial frequency features, and the attention module is used to focus on key time points or frequency bands, thereby improving the recognition accuracy and generalization ability of the model.

[0098] In this embodiment, the subjective questionnaire scores are classified from 0 to 10 as follows:

[0099]

[0100] The output of the motion sickness level recognition model includes four levels: no motion sickness, mild motion sickness, moderate motion sickness, and severe motion sickness.

[0101] The classification results are output to the control system in the form of numbers or labels for subsequent audio selection logic to call:

[0102]

[0103] In this embodiment, the construction of the audio sample library is specifically as follows:

[0104] Based on literature research, online materials, or user experiments, collect audio samples with the characteristics of relieving motion sickness and soothing, relaxing, and diverting attention, such as natural sounds (ocean waves, rain, bird calls), music segments (low frequency, slow rhythm), guiding voices (meditation, relaxation, psychological suggestion), etc.; subjectively evaluate the motion sickness relief effect of the collected audio samples, and calculate the motion sickness relief effect value η:

[0105] According to the η value, the audio samples are divided into four categories of audio with different levels of motion sickness relief effects to obtain an audio sample library for relieving motion sickness.

[0106] In this embodiment, the calculation of the motion sickness relief effect value η is specifically as follows:

[0107]

[0108] where m B and m A are the subjective scores before and after relieving motion sickness, respectively.

[0109] In this embodiment, the audio playback adopts a dynamic adjustment method, including volume progression and playback interval variation, and is combined with the subsequent electroencephalogram (EEG) fluctuation conditions for feedback optimization to form a closed-loop intervention control.

[0110] Referring to Figure 2 , the present invention also proposes a device for relieving motion sickness based on an EEG signal-controlled intelligent sound generation system. The device is implemented by using any of the above methods for relieving motion sickness based on an EEG signal-controlled intelligent sound generation system, and includes the following modules:

[0111] EEG acquisition module: continuously acquire the EEG signals of the occupant during the vehicle ride;

[0112] Signal processing module: preprocess the EEG signals and extract the EEG features including the frequency bands of delta waves, theta waves, alpha waves, and beta waves;

[0113] Evaluation module: calculate the comprehensive variation index of electroencephalogram (CVEI) based on the extracted EEG features;

[0114] Motion sickness recognition module: classify the motion sickness levels of the occupant according to the comprehensive variation index of electroencephalogram (CVEI) of the occupant by using a motion sickness level recognition model;

[0115] Audio control module: call the audio samples of the matching level according to the classification result of the occupant's motion sickness level;

[0116] Intelligent sound generation module: output the audio samples to intervene and relieve the discomfort of motion sickness.

[0117] (1) In the embodiments of the present application, the "comprehensive variation index of electroencephalogram (CVEI)" proposed for the first time, as a composite physiological index for judging the motion sickness state, comprehensively considers the changes in the power spectral density of the delta waves, theta waves, alpha waves, and beta waves in the EEG signals, and combines the time-domain volatility (i.e., the variance of change) of the signals, and can more accurately reflect the neural activation and fluctuation conditions of the brain during motion sickness. Compared with the traditional single-band EEG power value, CVEI has stronger expression ability in characterizing the complexity of neural activities.

[0118] (2) The present application constructs a motion sickness level recognition model based on a deep learning algorithm. CVEI is used as an important input feature and combined with the use of a convolutional neural network and an attention mechanism to realize intelligent classification of the states of no motion sickness, mild, moderate, and severe, significantly improving the recognition accuracy and stability of the model in the case of few samples, and having strong generalization ability and robustness.

[0119] (3)According to the identified motion sickness level, this application can accurately match the intervention audio content according to the individual motion sickness level, achieve hierarchical intervention, and effectively improve the motion sickness relief effect and user comfort.

[0120] (4)This application adopts a non-invasive identification method and uses a wearable electroencephalogram device to realize real-time and convenient monitoring of the individual's physiological state, improving the use comfort and adaptability; the device proposed in this application has a compact structure and can be embedded in portable electroencephalogram devices, vehicle-mounted systems or wearable hardware, and is applicable to various motion sickness intervention scenarios such as intelligent cockpits, virtual reality, and simulated flight, with broad application prospects.

[0121] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the purpose and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A method for controlling an intelligent voice system based on electroencephalogram signals to relieve motion sickness, characterized in that The method includes the following steps; Step S1: Collect the electroencephalogram (EEG) signal data of the occupant during motion sickness; Step S2: Preprocess the EEG signals and extract the EEG features including the frequency bands of delta wave, theta wave, alpha wave and beta wave; Step S3: Calculate the comprehensive variation index of electroencephalogram (CVEI) based on the extracted EEG features as an objective evaluation index of motion sickness level; Step S4: Use the motion sickness level recognition model to classify the motion sickness level according to the CVEI of the occupant's EEG, and the motion sickness level recognition model is constructed by integrating a convolutional neural network and an attention mechanism; Step S5: According to the classification result of the occupant's motion sickness level, play the audio for relieving motion sickness corresponding to the motion sickness level in the audio sample library.

2. The method for relieving motion sickness by a brain-computer interface-controlled intelligent sound system according to claim 1, wherein The EEG signals of the occupant during the ride are collected in real time through a wearable portable EEG acquisition device. The acquisition device supports multi-channel signal acquisition, the sampling rate is not less than 128 Hz, and at least covers the prefrontal lobe and parietal lobe regions; the wearable portable EEG acquisition device includes a head-mounted EEG instrument.

3. The method for relieving motion sickness by controlling an intelligent sound generation system based on electroencephalogram signals according to claim 1, wherein The specific content of step S2 includes: S2.1: Preprocess the EEG signal data, including filtering, artifact removal, segmenting the data by stimulation points and normalization processing; S2.2: Obtain the multi-channel multi-segment EEG signal data after preprocessing. Each segment of EEG signal data of each channel is sampled by a sliding window method to obtain the time-domain EEG signals of different time windows of each channel: X (c) = {X1 (c) , X2 (c) ,..., X N (c)} where, X (c) represents the time-domain EEG signal of a time window of channel c, where c = 1, 2, …, C and C is the total number of channels, and N represents the number of sampling points of a time window; S2.

3. For the time-domain EEG signals of each time window in each channel, calculate the power spectral density in different frequency bands using the fast Fourier transform (FFT) to obtain the power spectral densities P of the δ-wave, θ-wave, α-wave, and β-wave frequency bands δ (c) , P θ (c) , P α (c) , P β (c) .

4. The method for relieving motion sickness by controlling an intelligent voice system based on electroencephalogram signals according to claim 3, characterized in that, The specific content of step S3 includes: S3.1: Calculate the variance of the waveform change of the time-domain EEG signal of each channel in each time window; Among them, VAR (c) EEG is the waveform change variance of the time-domain EEG signal of a time window of channel c, is the mean of X (c) and X i (c) represents the data of the i-th sampling point in X (c) ; S3.2: Calculate the fusion feature CVEI of each channel in each time window; Among them, CVEI (c) is the CVEI value of a time window of channel c, and ε is a constant used to prevent the denominator from being zero S3.3: Statistically fuse the CVEI values of all channels corresponding to the same time window to obtain the overall CVEI index of the whole brain in different time windows: CVEI total Represents the whole-brain global CVEI metric for a time window.

5. The method for relieving motion sickness by controlling an intelligent sound generation system based on electroencephalogram signals according to claim 1, characterized in that, The training of the motion sickness level recognition model is specifically as follows: Design a motion sickness induction experiment and collect the EEG signals during motion sickness; preprocess the EEG signals and extract the EEG features including the frequency bands of delta wave, theta wave, alpha wave and beta wave; Calculate the comprehensive variation index of electroencephalogram (CVEI) based on the extracted EEG features as an objective evaluation index of motion sickness level, and combine the subjective questionnaire score to divide the motion sickness level, and construct a training set to train the motion sickness level classification model.

6. The method for relieving motion sickness by controlling an intelligent voice system based on electroencephalogram signals according to claim 5, wherein ​ ​ 7. The method for relieving motion sickness by controlling an intelligent voice system based on electroencephalogram signals according to claim 1, wherein ​ ​ ​ 8. The method for relieving motion sickness by controlling an intelligent voice system based on electroencephalogram signals according to claim 7, characterized in that ​ where m B and m A are the subjective scores before and after motion sickness relief, respectively.

9. The method for relieving motion sickness by controlling an intelligent sound generation system based on electroencephalogram signals according to claim 1, characterized in that, ​ 10. A device for controlling an intelligent voice system based on electroencephalogram signals to relieve motion sickness, characterized in that, The described device is implemented by using the method for relieving motion sickness by controlling an intelligent sound generation system based on electroencephalogram signals as described in any one of claims 1-9, and includes the following modules: An electroencephalogram acquisition module: to acquire the electroencephalogram signals of the occupant during the ride in real time; A signal processing module: to preprocess the electroencephalogram signals and extract the electroencephalogram features including the frequency bands of delta waves, theta waves, alpha waves and beta waves; An evaluation module: to calculate the comprehensive variation index of electroencephalogram CVEI based on the extracted electroencephalogram features; A motion sickness recognition module: to classify the motion sickness level according to the comprehensive variation index of electroencephalogram CVEI of the occupant by using a motion sickness level recognition model; An audio control module: to call the audio samples of the matching level according to the classification result of the occupant's motion sickness level; An intelligent sound generation module: to output the audio samples to intervene and relieve the discomfort of motion sickness.

Images