Fatigue suppression and emotion regulation system and feedback method based on multiple physiological parameters

Through the fatigue suppression and emotion regulation system with multiple physiological parameters, using pattern recognition of EEG signals and pulse signals, accurate judgment and real-time regulation of user physiological status are achieved, and the problem that cannot be effectively judged and regulated in the existing technology is solved. Improves the portability and practicality of the system.

CN117653114BActive Publication Date: 2024-11-22XI AN JIAOTONG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202311786928.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-23
Publication Date
2024-11-22
Estimated Expiration
2043-12-23

AI Technical Summary

Technical Problem

The existing fatigue and emotion detection methods cannot effectively determine physiological status, and cannot independently regulate fatigue and emotional status in real time according to changes in physiological indicators. The traditional device is poor in portability and cannot monitor and modulate physiological status for a long time.

Method used

Adopt a fatigue suppression and emotion regulation system based on multiple physiological parameters, collecting EEG signals and pulse signals, using the cascade structure of convolutional neural network and circulating neural network for pattern recognition, generate current signals for electrical stimulation, and combine with the ring ear-like brain. Electric electrode and earlobe pulse electrode acquisition device, a portable head-mounted mount is designed to achieve real-time regulation.

Benefits of technology

It realizes accurate judgment and real-time regulation of the user's physiological state, improves the portability and practicality of the system, can automatically or manually regulate fatigue and emotions, avoid misjudgment and misjudgment in traditional methods, and enhances physiological state A comprehensive understanding of the

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117653114B_ABST
    Figure CN117653114B_ABST
Patent Text Reader

Abstract

The present invention relates to a fatigue and emotion detection system capable of feedback intervention. The existing fatigue and emotion detection and related intervention methods have the technical problem that they cannot effectively determine the physiological state and cannot perform autonomous regulation in real time according to changes in physiological indicators. A fatigue suppression and emotion regulation system and a feedback method based on multiple physiological parameters are provided. The acquisition unit is used to acquire electroencephalogram signals and pulse signals. The first signal processing unit is used to simulate and amplify the electroencephalogram signals and pulse signals to obtain a first signal. The second signal processing unit is used to filter the first signal to obtain a second signal. The recognition unit is used to perform pattern recognition on the second signal to obtain data representing the fatigue and emotional state. The stimulation unit is used to generate a corresponding current signal based on the data representing the fatigue and emotional state to apply electrical stimulation to the user.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to a fatigue and emotion detection system capable of feedback intervention, and specifically relates to a fatigue suppression and emotion regulation system and a feedback method based on multiple physiological parameters. Background Art

[0002] In the process of economic and social development, technological progress, globalization, and economic prosperity are constantly evolving. However, with the accelerated pace of life and increased work pressure, many people are often troubled by fatigue and negative emotions. In addition, long-term busy and high-intensity work can also lead to physical and mental fatigue, lack of concentration, and depression, which may further affect people's work efficiency, physical function, and overall life satisfaction.

[0003] At present, the existing fatigue and emotion detection methods are: determining the individual's emotions and fatigue status through face recognition; detecting fatigue and emotional changes based on electrocardiographic respiratory indicators; and determining fatigue and emotional status through electroencephalogram monitoring. The intervention methods after detection mainly include: buzzer alarm, voice prompt, transcranial direct current stimulation, etc. The above detection methods are all aimed at a single indicator and cannot effectively determine the physiological state, and cannot be used as fatigue and emotional indicators for most people. In addition, the relevant devices for implementing the above detection and intervention methods are less portable, and cannot monitor and modulate the physiological state for a long time, nor can they autonomously regulate fatigue and emotional states in real time according to changes in physiological indicators. Summary of the invention

[0004] The present invention aims to address the technical problems that existing fatigue and emotion detection and related intervention methods cannot effectively determine the physiological state and cannot perform autonomous regulation in real time according to changes in physiological indicators. The present invention provides a fatigue suppression and emotion regulation system and feedback method based on multiple physiological parameters.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] In a first aspect, the present invention provides a fatigue suppression and emotion regulation system based on multiple physiological parameters, comprising:

[0007] A collection unit, used for collecting brain electrical signals and pulse signals;

[0008] A first signal processing unit, used for analog amplifying the EEG signal and the pulse signal to obtain a first signal;

[0009] A second signal processing unit, configured to filter the first signal to obtain a second signal;

[0010] an identification unit, configured to perform pattern recognition on the second signal to obtain data representing fatigue and emotional state;

[0011] The stimulation unit is used to generate corresponding current signals according to the data representing the fatigue and emotional state, and apply electrical stimulation to the user.

[0012] Further, the acquisition unit includes an ear ring-shaped EEG electrode acquisition device and an earlobe pulse electrode acquisition device;

[0013] The ear-ring-shaped EEG electrode collection device comprises a base, and a reference electrode, a ground electrode and an EEG collection electrode mounted on the base; the base is arc-shaped; the reference electrode, the ground electrode and the EEG collection electrode are all fixed with suction cups on their peripheries for adsorbing to the user's ears;

[0014] The earlobe pulse electrode collection device comprises a clamping body and two pulse collection electrodes mounted on the clamping body; the two pulse collection electrodes are arranged opposite to each other.

[0015] Furthermore, the reference electrode, the ground electrode and the EEG acquisition electrode are all Ag / Agcl electrodes;

[0016] The pulse collecting electrode is embedded with a photodiode and a photosensitive detector.

[0017] Further, it also includes a head-mounted mounting frame;

[0018] The head-mounted mounting frame is arc-shaped and matches the contour of the user's head between the two ears;

[0019] The stimulation unit is mounted on one end of the head-mounted mounting frame, and the earlobe pulse electrode collection device and the ear-ring EEG electrode collection device are mounted on the other end of the head-mounted mounting frame;

[0020] The first signal processing unit, the second signal processing unit and the identification unit are all installed on a head-mounted mounting frame.

[0021] In a second aspect, the present application proposes a feedback method based on multiple physiological parameters, comprising:

[0022] Performing analog amplification on the collected EEG signal and pulse signal to obtain a first signal;

[0023] Performing filtering on the first signal to obtain a second signal;

[0024] Performing pattern recognition on the second signal to obtain data representing fatigue and emotional state;

[0025] Based on the data representing the fatigue and emotional state, a corresponding current signal is generated for application to the user.

[0026] Further, performing pattern recognition on the second signal includes:

[0027] Converting the EEG signal portion of the second signal into an EEG time series in chronological order;

[0028] The convolutional neural network and recurrent neural network cascade structure are used to perform pattern recognition on the EEG time series and the pulse signal part of the second signal.

[0029] Furthermore, the convolutional neural network and recurrent neural network cascade structure includes a convolutional neural network, a first fully connected layer, a recurrent neural network, a second fully connected layer and a softmax layer connected in sequence;

[0030] The convolutional neural network is used to extract and process the spatial information of the EEG time series and capture the spatial information of the pulse signal from the pulse signal part in the second signal;

[0031] The first fully connected layer is used to downsample and feature-convert the spatial information of the EEG time series and the spatial information of the pulse signal to obtain first spatial information;

[0032] The recurrent neural network is used to extract and process the time information of the EEG time series, and receive and process the time series information of the pulse signal part of the second signal;

[0033] The second fully connected layer is used to downsample and feature-convert the time information of the EEG time series and the time series information of the pulse signal part to obtain the first time information;

[0034] The Softmax layer is used to combine the first spatial information and the first temporal information to obtain data representing fatigue and emotional state.

[0035] Furthermore, the recurrent neural network is LSTM;

[0036] The convolutional neural network includes Block1, Block2 and Block3;

[0037] A residual structure is added to Block2 and Block3.

[0038] Furthermore, the connection mode of the residual structure is indirect connection.

[0039] Furthermore, the residual structure is defined as follows:

[0040] Y'=F(X,{W i})+W s X

[0041] Among them, Y' represents the output of the residual network, F(X,{W i}) represents the residual mapping, X represents the physiological signal data, W irepresents the weight matrix, W s Represents a 1*1 convolution kernel.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. The present invention proposes a fatigue suppression and emotion regulation system based on multiple physiological parameters. The stimulation unit generates a corresponding current signal according to the processing results of the collected EEG signals and pulse signals, which is used to apply corresponding electrical stimulation to the user. Compared with the traditional auricular vagus nerve stimulation device, which requires manual control and cannot simultaneously perceive the changes in physiological state indicators in real time, and cannot measure the regulation effect of stimulation on fatigue and emotions, the regulation system of the present invention can apply stimulation while collecting, and can be manually intervened or automatically executed. In addition, the traditional regulation system only targets a single physiological indicator and cannot measure it in an all-round way. It may miss or misjudge the situation. The present invention combines EEG signals and pulse signals, which can more comprehensively understand the user's physiological state, provide strong support for accurate judgment, and further improve the practicality and portability of the system.

[0044] 2. The acquisition unit in the present invention can be designed as an ear-ring EEG electrode acquisition device and an earlobe pulse electrode acquisition device, which are used to collect EEG signals and pulse signals respectively, and can also be installed on a head-mounted mounting frame accordingly. Traditional EEG signal acquisition equipment is generally an EEG cap. The wearing of the EEG cap is relatively complicated and requires the cooperation of multiple experimenters. Each electrode also needs to be injected with conductive paste and then correctly worn with the help of the experimenter. The impedance and position of the EEG cap must be constantly calibrated. The structural design of the present invention solves the problem of the complicated wearing and use of traditional EEG caps, and designs an ear EEG acquisition device with the characteristics of portable acquisition. Special operations personnel can wear it by themselves without injecting conductive paste, and it also has the advantage of long-term acquisition.

[0045] 3. The present invention also proposes a feedback method based on multiple physiological parameters, which has the advantages of the above-mentioned control system. In addition, the convolutional neural network and the recurrent neural network cascade structure are used to perform pattern recognition of EEG signals and pulse signals, and a residual structure can be added to the structure of the convolutional neural network, so that the present invention can recognize more accurately and process faster, thereby achieving more accurate feedback. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0047] Figure 1 This is a connection diagram of Embodiment 1;

[0048] Figure 2 The figure is a connection diagram of Block 1, Block 2 and Block 3 in the second embodiment;

[0049] Figure 3 This is a schematic diagram of the structure of the ring ear-shaped EEG electrode acquisition device in Example 3;

[0050] Figure 4 This is a schematic diagram of the structure of the earlobe pulse electrode acquisition device in Example 3;

[0051] Figure 5 for Figure 4 Schematic diagram of the structure inside the middle earlobe pulse electrode collection device;

[0052] Figure 6 Schematic diagram of the installation structure of various components on the head-mounted mounting frame in Example 3;

[0053] Figure 7 This is a connection diagram of Embodiment 4;

[0054] Wherein: 1-base, 2-reference electrode, 3-ground electrode, 4-EEG collection electrode, 5-suction cup, 6-clamping body, 7-pulse collection electrode, 8-photodiode, 9-photosensitive detector, 10-head-mounted mounting frame, 11-stimulation unit, 12-earlobe pulse electrode collection device, 13-ear-shaped EEG electrode collection device, 14-collection unit, 15-power supply unit, 16-power supply circuit system, 17-dry battery, 18-first signal processing unit, 19-microcontroller, 20-second signal processing unit, 21-fatigue detection algorithm unit, 22-emotion detection algorithm unit, 23-fatigue and emotion discrimination unit, 24-reference and bias circuit, 25-debugging circuit, 26-communication circuit, 27-terminal equipment, 28-boosting module, 29-constant current module, 30-electric stimulation signal generating module, 31-stimulation electrode. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0056] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0057] In the description of the embodiments of the present invention, it should be noted that if the terms "upper", "lower", "horizontal", "inner", etc. indicate an orientation or positional relationship based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of the invention is usually placed when in use, it is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0058] In addition, if the term "horizontal" appears, it does not mean that the component must be absolutely horizontal, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", which does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0059] In the description of the embodiments of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal connection of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0060] In view of the existing fatigue and emotion detection methods, as well as the intervention methods after detection, there are problems such as single indicators and inability to implement autonomous regulation. The present invention proposes a fatigue suppression and emotion regulation system and feedback method based on multiple physiological parameters, which can collect individual physiological indicators in a portable manner, including EEG signals and pulse signals, and can decode fatigue and emotional states based on a microcontroller 19, and autonomously regulate the auricular vagus nerve stimulator to implement electrical stimulation. After regulation, individuals can get rid of fatigue and negative emotional states, ensure normal work and life, and can be widely used in related research experiments to efficiently collect related test data.

[0061] The present invention is further described in detail below in conjunction with the embodiments and drawings:

[0062] Embodiment 1

[0063] A fatigue suppression and emotion regulation system based on multiple physiological parameters includes a collection unit 14, a first signal processing unit 18, a second signal processing unit 20, a recognition unit and a stimulation unit 11.

[0064] Among them, the acquisition unit 14 is used to collect EEG signals and pulse signals. The first signal processing unit 18 is used to perform analog amplification on the EEG signals and pulse signals to obtain a first signal. The second signal processing unit 20 is used to filter the first signal to obtain a second signal. The recognition unit is used to perform pattern recognition on the second signal to obtain data representing fatigue and emotional states. The stimulation unit 11 is used to generate a corresponding current signal based on the data representing fatigue and emotional states, and apply electrical stimulation to the user. The magnitude of the current is determined based on the result of pattern recognition. Therefore, the present invention can perform precise real-time regulation based on the collected EEG signals and pulse signals.

[0065] Embodiment 2

[0066] Based on the first embodiment, when the recognition unit of the present invention performs pattern recognition, the following preferred method can be adopted:

[0067] In the second embodiment, based on the cascade structure of convolutional neural network and recurrent neural network, deep learning method is used for pattern recognition, and convolutional neural network (CNN) is used to extract and process the spatial information of EEG signal and pulse signal. Recurrent neural network (RNN) network is used to extract and process the time information of EEG signal and pulse signal, and the two neural network algorithms are cascaded to convert the EEG signal part in the second signal into EEG time series in time sequence, and the EEG time series and the pulse signal part in the second signal are used as input, and the user's emotion and fatigue state are judged by calculating the spatiotemporal feature information.

[0068] The specific model implementation is as follows:

[0069] The overall network structure consists of 5 parts: Block1, Block2 and Block3, 2 layers of full connection, 1 layer of LSTM and 1 layer of Softmax. Block1, Block2 and Block3 form a convolutional neural network. LSTM is a special recurrent neural network. As a preferred solution, the recurrent neural network adopts LSTM. The connection relationship is: the convolutional neural network, the first fully connected layer, LSTM, the second fully connected layer and the softmax layer are connected in sequence.

[0070] The EEG signal time series and the pulse signal part of the second signal are input into the neural network model. The convolutional neural network model is used to extract data features, which are processed by the LSTM network layer and output by the fully connected layer and the softmax layer to finally obtain the prediction result.

[0071] Convolutional Neural Network (CNN): In the cascade structure, CNN is responsible for extracting and processing the spatial information of the EEG time series, and also includes the processing of the pulse signal to capture the spatial characteristics of the pulse signal.

[0072] First fully connected layer: This layer is used to downsample and transform the spatial information, while also considering the characteristics of the pulse signal at this stage.

[0073] Recurrent Neural Network (RNN): RNN is used in this structure to extract and process the time information of the EEG time series, and also receives and processes the time series information of the pulse signal.

[0074] Second fully connected layer: Similar to the first fully connected layer, this layer continues to downsample and transform the temporal information, while comprehensively considering the features from the EEG and pulse signals.

[0075] Softmax layer: The final Softmax layer is used to combine the spatial information and temporal information after feature conversion, including the pulse signal, to calculate the data representing fatigue and emotional state. This enables the system to comprehensively consider EEG and pulse signals to achieve accurate judgment of the user's state.

[0076] like Figure 2 As shown in the figure, it is a preferred structure of Block1, Block2 and Block3, where (a) is a schematic diagram of the structure of Block1, (b) is a schematic diagram of the structure of Block2, and (c) is a schematic diagram of the structure of Block3. Block1 includes three 1×3 convolutional layers, Block2 and Block3 both include three 1×3 convolutional layers and an ADD&ReLU block, and a 1×1 convolutional layer as a residual structure. Among them, the ADD operation performs a weighted summation on the input data through the weights, and then the ReLU function performs a nonlinear transformation on the result of the weighted summation. This combined operation can help the neural network better learn and represent complex input data.

[0077] The residual structure generally includes direct mapping and indirect mapping. The residual structure is defined as follows:

[0078] X l+1 =X l +F(X l ,{W i}) (1)

[0079] x l and x l +1 are input and output data, respectively, F(X,{W i}) is the residual mapping, defined as:

[0080] y=F(X, {Wi}) (2)

[0081] Formula (1) reflects the properties of the residual network: the l+1 layer can be expressed as the sum of the lth layer and the residual structure. According to the chain rule of multi-layer neural networks, the loss function ε is l The gradient value of can be expressed by the following formula (3):

[0082]

[0083] Among them, X L is the Lth element in the input vector. L is a parameter representing the dimension of the input vector, that is, the length of the input vector X. i is the i-th element in the input vector, where i is greater than or equal to 1 and less than or equal to L-1.

[0084] Because during the training process, It will not always be -1, so the gradient disappearance phenomenon will not occur during the reverse calculation of the residual network. There are two connection methods for the residual structure, direct connection and indirect connection. In the present invention, because the input and output data dimensions are inconsistent, indirect connection is used. In formula (4), W s X can be used to match dimensions, where W s The convolution kernel is 1*1. The residual structure is constructed by indirect connection to avoid network degradation:

[0085] Y'=F(X,{W i})+W s X (4)

[0086] Among them, Y' represents the output of the residual network, F(X,{W i}) represents the residual mapping, X represents the physiological signal data, W i represents the weight matrix, W s Represents a 1*1 convolution kernel.

[0087] After passing through Block1, Block2 and Block3, the data is downsampled and converted through the first fully connected layer. LSTM is an important part of the network, which is used to extract data features in the time dimension to reduce network parameters. In the LSTM structure, Hidden_size is 128, and the number of LSTM stacked layers is two. The data processed by LSTM is input into the second fully connected layer, and finally the predicted value is calculated by the function of the softmax layer. Different predicted values ​​are obtained to represent different fatigue and emotional conditions. The network model parameters can also be dynamically updated through feedback.

[0088] Embodiment 3

[0089] Based on the first embodiment, the third embodiment proposes a preferred specific structure of the acquisition unit 14, wherein the acquisition unit 14 includes an auricular EEG electrode acquisition device 13 and an earlobe pulse electrode acquisition device 12. The auricular EEG electrode acquisition device 13 is used to acquire EEG signals, and the earlobe pulse electrode acquisition device 12 is used to acquire pulse signals.

[0090] like Figure 3 As shown, the ring-ear-shaped EEG electrode collection device 13 includes a base 1, and a reference electrode 2, a ground electrode 3, and an EEG collection electrode 4 installed on the base 1. The base 1 is arc-shaped and basically matches the curvature of the user's auricle. The reference electrode 2, the ground electrode 3, and the EEG collection electrode 4 are arranged in sequence along the arc of the base 1, and a suction cup 5 is fixed on the periphery to allow the ring-ear-shaped EEG electrode collection device 13 to be adsorbed on the user's ear. Figure 3 A preferred solution is shown in FIG. 8 , where the number of EEG acquisition electrodes 4 is 8, and one reference electrode 2 and one ground electrode 3 are provided. Figure 3 The fifth electrode is the reference electrode 2, the sixth electrode is the ground electrode 3, and the rest are EEG acquisition electrodes 4. In some embodiments of the present invention, the reference electrode 2, the ground electrode 3 and the EEG acquisition electrode 4 can all be Ag / Agcl electrodes.

[0091] In the ring-ear-shaped EEG electrode acquisition device 13, the reference electrode 2 provides a relatively stable potential as a signal comparison benchmark, the ground electrode 3 is connected to the current loop to ensure the normal flow of the signal, and the EEG acquisition electrode 4 records the potential changes caused by the EEG activity on the scalp. In signal processing, the signal of the reference electrode 2 is usually subtracted to obtain the EEG signal relative to the reference electrode 2, and the signal of the ground electrode 3 is used to represent the current loop of the EEG signal. The signal recorded by the EEG acquisition electrode 4 is amplified and filtered, and by performing a differential operation with the reference electrode 2, the EEG signal for studying the timing and spectral characteristics of EEG activity is finally obtained.

[0092] like Figure 4 As shown, the earlobe pulse electrode collection device 12 includes a clamping body 6 and two pulse collection electrodes 7 mounted on the clamping body 6. The two pulse collection electrodes 7 are arranged opposite to each other. The structure of the clamping body 6 is similar to the shape of a conventional clip, which enables the two pulse collection electrodes 7 to be attached to the earlobe. Figure 5 As shown, a photodiode 8 and a photosensitive detector 9 are embedded in the pulse collecting electrode 7 .

[0093] The photodiode 8 converts light into current through the photoelectric effect, and the photosensitive detector 9 converts the photoelectric signal generated by the photodiode 8 into a voltage signal. This combination allows the device to achieve non-invasive heart rate and pulse waveform monitoring by monitoring tiny blood flow changes on the earlobe, especially blood flow changes related to the heart pulse.

[0094] The traditional EEG signal acquisition device is an EEG cap, which is complicated to wear and requires the cooperation of multiple experimenters. First, each electrode needs to be injected with conductive paste, and then the EEG cap is worn correctly with the help of the experimenter, and the impedance and position of the EEG cap are constantly calibrated. Figure 6 As shown, in order to solve the problem of the complexity of wearing a traditional EEG cap and further improve portability, each component is installed on a head-mounted mounting frame 10, especially for special operations personnel, who can wear it by themselves without injecting conductive paste, and have the advantage of long-term collection, specifically:

[0095] The overall shape of the head-mounted mounting frame 10 is similar to that of a headset, and is basically adapted to the contour of the user's head between the two ears. The stimulation unit 11 is mounted at one end of the head-mounted mounting frame 10, and the earlobe pulse electrode acquisition device 12 and the ear-shaped EEG electrode acquisition device 13 are mounted at the other end of the head-mounted mounting frame 10. The first signal processing unit 18, the second signal processing unit 20 and the recognition unit are all mounted on the head-mounted mounting frame 10. In other embodiments of the present invention, the specific installation position of each component can also be adjusted accordingly according to usage habits and design requirements.

[0096] Embodiment 4

[0097] like Figure 7 As shown, based on the first embodiment, a specific implementation method is proposed in the fourth embodiment, including a collection unit 14, a first signal processing unit 18, a second signal processing unit 20, a recognition unit, a stimulation unit 11, a power supply unit 15 and a terminal device 27. The collection unit 14 includes an ear-shaped EEG electrode collection device 13 and an earlobe pulse electrode collection device 12, and the recognition unit includes a fatigue detection algorithm unit 21, an emotion detection algorithm unit 22 and a fatigue and emotion discrimination unit 23.

[0098] Fatigue and emotion detection algorithm unit 22: In the cascade neural network structure, the fatigue and emotion detection unit corresponds to the entire deep learning model, which may use a convolutional neural network (CNN) to process the spatial information of EEG signals and a recurrent neural network (RNN) to process the temporal information of EEG signals. These two functions work together in the same neural network structure to comprehensively consider the temporal and spatial characteristics of EEG signals.

[0099] Fatigue and emotion discrimination unit 23: Corresponding cascade neural network structure: Fatigue and emotion discrimination unit 23 is the output layer of the entire cascade neural network, integrating the outputs of fatigue detection algorithm unit 21 and emotion detection algorithm unit 22. The goal of this unit is to calculate spatiotemporal feature information, comprehensively consider the spatial and temporal information of the EEG signal, and ultimately discriminate the user's emotion and fatigue state.

[0100] The first signal processing unit 18 adopts an analog circuit, the second signal processing unit 20 adopts a digital filtering algorithm unit, and the stimulation unit 11 is a vagus nerve stimulation module, including a boost module 28, a constant current module 29, an electrical stimulation signal generation module 30 and a stimulation electrode 31. The terminal device 27 is connected to the fatigue and emotion discrimination unit 23 in the recognition unit through a communication circuit 26. The second signal processing unit 20 and the recognition unit together constitute a microcontroller 19, and the microcontroller 19 is also configured with a corresponding reference and bias circuit 24 and a debugging circuit 25.

[0101] The role of the reference and bias circuit 24 in this application is mainly to ensure the stability and accuracy of the system. These circuits include a stable reference voltage provided to the analog circuit and a required bias current for each circuit. A stable reference voltage is essential for the accuracy of the analog circuit, while the bias current ensures that each part is within the appropriate operating range. The debug circuit 25 is used to diagnose and adjust the performance of the system, making it easier to troubleshoot during development, testing, and maintenance. These circuits may include an interface for communicating with the microcontroller 19, through which the status of each part of the system can be monitored in real time and adjusted. The debug circuit 25 provides a connection to other parts of the system, allowing developers to monitor and intervene in the system in real time through debugging tools or interfaces.

[0102] The power supply unit 15 generally includes a power supply circuit system 16 and a dry battery 17 to supply power to the first signal processing unit 18 , the microcontroller 19 and the stimulation unit 11 .

[0103] The specific data flow and working principle are as follows:

[0104] Step 1: Turn on the power supply unit 15 to transmit power to the first signal processing unit 18, the microcontroller 19 and the stimulation unit 11 to supply power and perform energy supply.

[0105] Step 2: The user's pulse signal and EEG signal are collected as multiple physiological parameters through the earlobe pulse electrode collection device 12 and the ear ring EEG electrode collection device 13.

[0106] Step 3: The acquisition unit 14 transmits the multi-physiological parameter data including the pulse signal and the EEG signal to the first signal processing unit 18, and the first signal processing unit 18 performs analog front-end amplification on the EEG signal and the pulse signal.

[0107] Step 4: The amplified signal is transmitted to the microcontroller 19, and the second signal processing unit 20 in the microcontroller 19 performs notch filtering and low-pass filtering on the amplified multi-physiological parameter data. Then, the signal is transmitted to the recognition unit to perform pattern recognition on the fatigue and emotional data of the user. The pattern recognition can be performed by the method in the second embodiment, or by a three-layer back propagation neural network structure.

[0108] The three-layer back-propagation neural network structure usually includes an input layer, a hidden layer, and an output layer, which is trained through a back-propagation algorithm to learn and recognize different patterns. In this case, the processed multi-physiological parameter data may be used as input, and the weights and biases learned by the neural network are used for pattern classification, and finally the recognition results of the user's fatigue and emotional state are output.

[0109] The output results of the fatigue detection algorithm unit 21 and the emotion detection algorithm unit 22 are transmitted to the fatigue and emotion determination unit 23 . The fatigue and emotion determination unit 23 determines the fatigue and emotion state of the individual according to the algorithm results and sends instructions to the stimulation unit 11 .

[0110] Step 5: The instruction sent by the fatigue and emotion judgment unit turns on the vagus nerve stimulation module. The boost module 28 converts the input low voltage signal into a high voltage output signal through circuit design and conversion technology. The constant current module 29 controls the current value output by the electrical stimulation signal generating module 30 to remain constant. The electrical stimulation signal generating module 30 transmits the electrical signal to the stimulation electrode 31, and performs electrical stimulation on the user through the stimulation electrode 31.

[0111] At the same time, the fatigue and emotion determination unit 23 can send the result to the terminal device 27 through the communication circuit 26 for viewing or interaction.

[0112] When traditional auricular vagus nerve stimulation devices regulate fatigue and emotional states, they do not take into account individual physiological indicators in real time. They only regulate physiological states by simply manually controlling the vagus nerve stimulator, which is too violent. After manual regulation, changes in physiological state indicators cannot be perceived in real time, and the regulatory function of the stimulator on fatigue and emotions cannot be measured. The control system designed by the present invention is triggered by the user's negative emotions and fatigue state, and acts on the vagus nerve in an electrical stimulation manner, which can improve the regulatory function of the human nervous system. It is compatible with both automatic and manual control modes to achieve individualized and customized effects. It has broad application prospects in experimental research and teaching related to regulation, and can also provide certain guidance for related treatments in the future. It is a non-drug research method with the advantages of safety, painlessness and non-invasiveness, and can realize intelligent regulation of fatigue and emotional states of special operations personnel.

[0113] In addition, traditional portable physiological indicator acquisition devices usually only collect a single physiological indicator. For example, only EEG signals or ECG signals are collected. The collection of a single physiological indicator is relatively simple and cannot comprehensively measure the physiological indicators of special operations personnel. Missed or misjudgment may occur. Compared with traditional physiological signal acquisition equipment, the present invention integrates the collection of ear EEG signals and pulse signals to achieve the collection of multiple physiological parameters. By collecting multiple physiological parameters of the ear, the user's physiological state can be more comprehensively understood, providing more powerful support for more accurate fatigue and emotional regulation. In addition, the integrated design of ear collection and electrical stimulation will further improve the practicality and portability of the product.

[0114] Corresponding to the above fatigue suppression and emotion regulation system based on multiple physiological parameters, the present invention also proposes a feedback method based on multiple physiological parameters, including:

[0115] S101, analog amplifying the collected EEG signal and pulse signal to obtain a first signal;

[0116] S102, filtering the first signal to obtain a second signal;

[0117] S103, performing pattern recognition on the second signal to obtain data representing fatigue and emotional state;

[0118] S104, generating a corresponding current signal according to the data representing the fatigue and emotional state, for applying to the user.

[0119] The specific method for implementing each step can adopt the specific scheme in each embodiment of the control system.

[0120] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A fatigue suppression and emotion regulation system based on multiple physiological parameters, characterized in that: include: A collection unit (14), used for collecting brain electrical signals and pulse signals; A first signal processing unit (18) is used to perform analog amplification on the brain electrical signal and the pulse signal to obtain a first signal; A second signal processing unit (20) is used to filter the first signal to obtain a second signal; The recognition unit is used to perform pattern recognition on the second signal to obtain data representing fatigue and emotional state; the pattern recognition includes: converting the EEG signal part in the second signal into an EEG time series in chronological order; using a convolutional neural network and a recurrent neural network cascade structure to perform pattern recognition on the EEG time series and the pulse signal part in the second signal; the convolutional neural network and the recurrent neural network cascade structure include a convolutional neural network, a first fully connected layer, a recurrent neural network, a second fully connected layer and a softmax layer connected in sequence; the convolutional neural network is used to extract and process the spatial information of the EEG time series, and capture the spatial information of the pulse signal from the pulse signal part in the second signal; the first fully connected layer is used to The spatial information of the EEG time series and the pulse signal is downsampled and feature transformed to obtain the first spatial information; the recurrent neural network is used to extract and process the time information of the EEG time series, and receive and process the time series information of the pulse signal part in the second signal; the second fully connected layer is used to downsample and feature transform the time information of the EEG time series and the time series information of the pulse signal part to obtain the first time information; the Softmax layer is used to combine the first spatial information and the first time information to obtain data representing fatigue and emotional state; the recurrent neural network is LSTM; the convolutional neural network includes Block1, Block2 and Block3; residual structures are added to Block2 and Block3; The stimulation unit (11) is used to generate corresponding current signals based on data representing fatigue and emotional states, and to apply electrical stimulation to the user.

2. The fatigue suppression and emotion regulation system based on multiple physiological parameters according to claim 1 is characterized in that: The collection unit (14) comprises an ear-ring-shaped EEG electrode collection device (13) and an earlobe pulse electrode collection device (12); The ear-ring-shaped EEG electrode collection device (13) comprises a base (1), and a reference electrode (2), a ground electrode (3), and an EEG collection electrode (4) mounted on the base (1); the base (1) is arc-shaped; suction cups (5) are fixed to the peripheries of the reference electrode (2), the ground electrode (3), and the EEG collection electrode (4) for adsorbing to the user's ear; The earlobe pulse electrode collection device (12) comprises a clamping body (6) and two pulse collection electrodes (7) mounted on the clamping body (6); the two pulse collection electrodes (7) are arranged opposite to each other.

3. The fatigue suppression and emotion regulation system based on multiple physiological parameters according to claim 2 is characterized in that: The reference electrode (2), the ground electrode (3) and the EEG acquisition electrode (4) are all Ag / Agcl electrodes; The pulse collecting electrode (7) is embedded with a photodiode (8) and a photosensitive detector (9).

4. The fatigue suppression and emotion regulation system based on multiple physiological parameters according to claim 2 or 3, characterized in that: Also includes a head-mounted mounting frame (10); The head-mounted mounting frame (10) is arc-shaped and matches the contour of the user's head between the two ears; The stimulation unit (11) is mounted on one end of the head-mounted mounting frame (10), and the earlobe pulse electrode collection device (12) and the ear-ring EEG electrode collection device (13) are mounted on the other end of the head-mounted mounting frame (10); The first signal processing unit (18), the second signal processing unit (20) and the identification unit are all mounted on a head-mounted mounting frame (10).

5. A feedback method based on multiple physiological parameters, characterized in that: include: Performing analog amplification on the collected EEG signal and pulse signal to obtain a first signal; Performing filtering on the first signal to obtain a second signal; The second signal is subjected to pattern recognition to obtain data characterizing fatigue and emotional state; the pattern recognition includes: converting the EEG signal part in the second signal into an EEG time series in chronological order; using a convolutional neural network and a recurrent neural network cascade structure to perform pattern recognition on the EEG time series and the pulse signal part in the second signal; the convolutional neural network and the recurrent neural network cascade structure include a convolutional neural network, a first fully connected layer, a recurrent neural network, a second fully connected layer and a softmax layer connected in sequence; the convolutional neural network is used to extract and process the spatial information of the EEG time series, and capture the spatial information of the pulse signal from the pulse signal part in the second signal; the first fully connected layer is used to The spatial information of the pulse signal is downsampled and feature transformed to obtain the first spatial information; the recurrent neural network is used to extract and process the time information of the EEG time series, and receive and process the time series information of the pulse signal part in the second signal; the second fully connected layer is used to downsample and feature transform the time information of the EEG time series and the time series information of the pulse signal part to obtain the first time information; the Softmax layer is used to combine the first spatial information and the first time information to obtain data representing fatigue and emotional state; the recurrent neural network is LSTM; the convolutional neural network includes Block1, Block2 and Block3; residual structures are added to Block2 and Block3; Based on the data representing the fatigue and emotional state, a corresponding current signal is generated for application to the user.

6. The feedback method based on multiple physiological parameters according to claim 5, characterized in that: The connection mode of the residual structure is indirect connection.

7. The feedback method based on multiple physiological parameters according to claim 6, characterized in that: The residual structure is defined as follows: Y'=F(X,{W i })+W s X Among them, Y' represents the output of the residual network, F(X,{W i }) represents the residual mapping, X represents the physiological signal data, W i represents the weight matrix, W s Represents a 1*1 convolution kernel.