Fatigue intervention system and device based on electroencephalogram

Abstract

The application discloses a fatigue intervention system and device based on electroencephalogram, wherein the intervention system comprises an electroencephalogram acquisition module, at least comprising a first electrode, for acquiring electroencephalogram data of a target through the first electrode; a transcranial electrical stimulation module, at least comprising a second electrode, for implementing weak current electrical stimulation signals on the target through the second electrode according to a stimulation instruction, wherein the electrical stimulation signals comprise direct current stimulation signals or alternating current stimulation signals or random noise signals; and a processing module, for acquiring the electroencephalogram data fed back by the electroencephalogram acquisition module, and matching the stimulation instruction according to the electroencephalogram data. The application provides a system and device which are mild in stimulation, not easy to be interfered, obvious in intervention effect and effective in recovery of fatigue.

Description

TECHNICAL FIELD

[0001] The present application relates to an electroencephalogram-based fatigue intervention device, in particular to an electroencephalogram-based fatigue intervention system and device. BACKGROUND

[0002] Modern technology society relies on 24-hour operation or shift work of industry, such as transportation, nursing, manufacturing, military, aviation, hospitals and many public service industries. People working in shifts mainly disrupt sleep and physiological cycle rhythm. The disruption of physiological cycle and sleep deprivation can also lead to a decrease in alertness, impaired performance and fatigue. Secondly, some high-stress occupations (such as nuclear power plant operators) can also cause psychological fatigue. When people are in a state of physical or psychological fatigue, their cognitive ability will decrease, which has a great negative impact on safe production or safe work (such as fatigue driving). Therefore, it is necessary to detect fatigue and implement specific intervention programs to prevent accidents caused by working in a state of fatigue.

[0003] Among the prior art, "A voice fatigue degree detection method for mental fatigue" provides a judgment and detection of fatigue degree according to the way of voice. "A fatigue degree detection method and device" provides a detection of fatigue degree based on electroencephalogram signals and electrocardiogram signals. "An electroencephalogram detection method for fatigue driving based on comprehensive judgment of electrooculogram and electroencephalogram" provides a judgment and detection of fatigue degree based on electroencephalogram signals and electrooculogram signals. "Fatigue driving early warning system based on image processing" provides a detection and judgment of fatigue degree based on facial image recognition, and displays on a display module. "Electroencephalogram fatigue and drowsiness early warning and brain-awakening device based on displacement electrodes and control method thereof" also detects and judges fatigue degree by collecting electroencephalogram of human brain, and plays brain-awakening music through sound and performs intervention on fatigue through earlobe vibration massage.

[0004] The above prior art provides different technical means or combinations of different technologies (multimodal) to detect fatigue degree. Most of the patents mainly collect electroencephalogram signals and auxiliary collect other signals of human body (such as electromyogram, electrooculogram and electrocardiogram, etc.), adopt respective algorithms to realize detection of fatigue degree, and display the detected fatigue degree result on a host computer or a display module (such as a display screen), without intervention on fatigue.

[0005] Current main fatigue intervention methods include rest, drugs (modafinil), auditory reminders, visual reminders, etc.

[0006] The prior art displays fatigue degree through the background, which is a visual intervention mode for fatigue, but this mode can only intervene manually, which has the disadvantage that someone needs to monitor the background all the time, and a third party needs to intervene in the fatigue person. In the scenarios of fatigue driving and nuclear power plants, manual intervention may cause the fatigue person to be startled, resulting in misoperation and accidents.

[0007] Another prior art "EEG fatigue and drowsiness warning and brain-awakening device based on displacement electrode and control method thereof" uses brain-awakening music and earlobe vibration to stimulate and intervene the fatigue person. First, relevant literature research shows that sound stimulation to intervene fatigue is effective, but the duration is relatively short (less than 10 seconds). That is, for the fatigue person, suddenly playing brain-awakening music can make the fatigue person's spirit rise for a short time, but this cannot change the nature of fatigue (that is, it cannot restore fatigue). Second, sound stimulation is effective for mild fatigue, but for severe fatigue, the effect of sound stimulation is weak. Some severe fatigue persons can maintain a drowsy state in a noisy environment. Third, sudden sound stimulation can also cause the fatigue person to be startled. Finally, the implementation of earlobe vibration in this patent needs to use a clip to clip on the ear, which is not a good experience. In the stage before fatigue, long-term use of a clip on the ear will cause pain, and the audio hanging on the neck has high power consumption, which is not conducive to the long-term operation of embedded devices.

[0008] Therefore, it can be known that most of the current existing technologies are based on the detection of fatigue degree, and no intervention scheme for fatigue degree is proposed. A small part of the technical implementation adopts sound stimulation or visual monitoring as the intervention scheme for fatigue.

[0009] Research shows that an effective fatigue intervention scheme can alleviate fatigue and repair the physiological and psychological trauma caused by fatigue. Current fatigue intervention schemes include rest, drugs (modafinil, etc.), sound reminders, visual reminders, etc.

[0010] In the prior art, visual and sound reminders are mainly used to intervene fatigue. The visual reminder is mainly to display the fatigue result on the host computer or the display in the background. This visual stimulation method is not obvious and needs a third party to monitor the fatigue state and intervene manually. The sound reminder is more effective than the visual reminder, but the effective duration of intervention is short (less than 10 seconds), which has a certain effect on mild fatigue but is not obvious for severe fatigue. Visual and sound reminders are currently the two main fatigue intervention modes. These two modes can cause the fatigue person to be startled and misoperate, and these two methods only intervene fatigue and cannot recover fatigue.

[0011] Rest is the most effective way to intervene fatigue and recover fatigue, but in some specific situations, such as driving on a highway, real-time parking for rest cannot be realized.

[0012] In addition, drugs have a certain effect on fatigue intervention, but the side effects of drug intervention on fatigue need further research.

[0013] The information disclosed in this section is only intended to increase the understanding of the overall background of the present application, and should not be regarded as acknowledging or implying in any form that the information constitutes prior art known to those skilled in the art. SUMMARY

[0014] The purpose of the present application is to provide a fatigue intervention system and device based on electroencephalogram, which has high recognition and judgment accuracy, fast response, mild stimulation, is not easy to be disturbed, has obvious intervention effect and can effectively recover fatigue.

[0015] To achieve the above purpose, the embodiment of the present application provides a fatigue intervention system based on electroencephalogram, which comprises an electroencephalogram acquisition module, at least comprising a first electrode, for acquiring electroencephalogram data of a target through the first electrode; a transcranial electrical stimulation module, at least comprising a second electrode, for implementing weak current electrical stimulation signals on the target through the second electrode according to a stimulation instruction, the electrical stimulation signals comprising direct current stimulation signals or alternating current stimulation signals or random noise signals; a processing module, for acquiring the electroencephalogram data fed back by the electroencephalogram acquisition module, and matching the stimulation instruction according to the electroencephalogram data.

[0016] In one or more embodiments of the present application, the first electrode and the second electrode are common electrodes (a common electrode is an electrode that functions as both a first electrode to acquire electroencephalogram data and a second electrode to implement electrical stimulation).

[0017] In one or more embodiments of the present application, the processing module further comprises an SVM classifier to obtain a stimulation instruction library matched with the electroencephalogram data and the fatigue degree. The SVM is a binary classification algorithm which can classify according to the electroencephalogram features to determine whether the match is fatigue or not fatigue.

[0018] In one or more embodiments of the present application, the processing module comprises a cloud system, at least for collecting electroencephalogram data and / or output processing and / or instruction matching in the cloud.

[0019] In one or more embodiments of the present application, the fatigue intervention system further comprises a communication module, at least for transmitting electroencephalogram data or stimulation instructions.

[0020] In one or more embodiments of the present application, the communication module is a wired communication module or a wireless communication module.

[0021] In one or more embodiments of the present application, the wireless communication module is at least any one of a mobile communication module, a wifi communication module, a near field communication module, and an optical communication module.

[0022] In one or more embodiments of the present application, the near field communication module comprises an NFC communication module or a Bluetooth communication module or an infrared communication module.

[0023] In one or more embodiments of the present application, the fatigue intervention device comprises the electroencephalogram-based fatigue intervention system as described above.

[0024] In one or more embodiments of the present application, the fatigue intervention system device comprises a controller, an electroencephalogram cap provided with a plurality of electrodes, the electrodes comprising a first electrode and / or a second electrode; a signal acquisition device communicatively connected to the first electrode, for acquiring electroencephalogram data fed back by the first electrode; and an electrical stimulation device comprising at least a stimulation source, the stimulation source releasing a weak current to the second electrode according to stimulation instructions of the controller. Preferably, the fatigue intervention system device can further be provided with a terminal, which is a device capable of displaying fatigue state, intervention state, device running state, etc. in real time by running software, such as a mobile phone loaded with an APP and a computer, etc. At this time, the terminal can be communicatively connected to the processing module through the communication module to obtain information in real time. The stimulation source can be a current source, etc. to meet the release of stimulation electrical signals according to instructions. The stimulation signal in the present application is continuous, which can continuously stimulate for 8-30 minutes. The fatigue intervention function can be realized through a portable device such as an electroencephalogram cap, so as to meet the fatigue detection of long-term continuous workers such as drivers without going to a fixed place for detection.

[0025] Compared with the prior art, the electroencephalogram-based fatigue intervention system and device according to the embodiment of the application judges the fatigue degree by collecting the electroencephalogram data of the human brain, and intervenes the fatigue degree in the mode of TES (transcranial electrical stimulation). The sample entropy is used as the characteristic value of the fatigue degree, and the machine learning algorithm SVM is used for binary classification to judge the fatigue degree, and the quantitative calculation formula for the accuracy of judging the fatigue degree is provided. Secondly, the application provides a new fatigue intervention mode, that is, TES (transcranial electrical stimulation). The specific TES stimulation mode has three kinds, which are tDCS (transcranial direct current stimulation), tACS (transcranial alternating current stimulation) and tRNS (transcranial random noise stimulation). The transcranial electrical stimulation has a good intervention effect on different degrees of fatigue, and because the transcranial electrical stimulation can intervene and change the brain waves of the human brain, the fatigue can be recovered to a certain extent.

[0026] The application collects the electroencephalogram data of the human brain, processes the electroencephalogram data by an algorithm, judges the fatigue degree of the human, and determines to use the transcranial electrical stimulation to intervene the fatigue according to whether the human is fatigued and the fatigue degree grade. The purpose of the application is to provide a new fatigue intervention scheme, which has the advantages of not being frightened, obvious intervention effect on fatigue and effective recovery of fatigue compared with the traditional sound prompting or visual prompting.

[0027] In the implementation process of the application, a large amount of computing resources are required for calculation and processing due to the removal of motion artifacts, filtering, notch pre-processing, sample entropy characteristic value extraction and SVM machine learning binary classification, so the application proposes that the fatigue degree detection process can be mainly processed by a cloud server, and of course this process can also be implemented by distributed operation with idle computing resources in the terminal such as a mobile phone in the work area.

[0028] The prior art scheme intervenes the fatigue based on vision and hearing, and the intervention effect is not ideal. The intervention system based on the transcranial electrical stimulation can not only intervene the fatigue, but also affect and improve the brain waves of the brain, so that the brain waves of the brain are recovered from the abnormal state to the normal state, and the brain waves of the brain have a good recovery effect on the fatigue. In the application, the first electrode and the second electrode are non-invasive and are attached to the scalp skin, directly stimulating the brain and the central nervous system, and the transcranial electrical stimulation realizes the fatigue intervention. The application detects the fatigue degree based on the electroencephalogram, classifies the algorithm based on the sample entropy as the characteristic value, judges whether the fatigue is fatigued, and does not need to use the multi-modal detection scheme (simultaneously detecting the electroencephalogram, detecting the electrocardiogram, detecting the electrooculogram, etc.). BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1 is a schematic view of the structure according to an embodiment of the application;

[0030] Figure 2is a workflow schematic according to an embodiment of the present application;

[0031] Figure 3 is a transcranial electrical stimulation signal model illustration according to an embodiment of the present application. DETAILED DESCRIPTION

[0032] The specific embodiments of the present application will now be described in detail below with reference to the drawings, but the scope of protection of the present application is not limited by the specific embodiments.

[0033] Unless otherwise defined, the terms "including", "comprising", "consisting" or "having" and the like in the specification and claims are used in their open-ended sense, and are not limited to the elements or components listed.

[0034] TMS (Transcranial Magnetic Stimulation)

[0035] TES (Transcranial Electrical Stimulation)

[0036] tDCS (transcranial Direct Current Stimulation)

[0037] tACS (transcranial Alternating Current Stimulation)

[0038] tRNS (transcranial Random Noise Stimulation)

[0039] SE (sample entropy)

[0040] FFS (false fatigue State)

[0041] FNS (false Normal State)

[0042] TFS (true fatigue State)

[0043] TNS (true Normal State)

[0044] SVM (support vector machines)

[0045] Transcranial magnetic stimulation (TMS) can make the brain locally excited by stimulating frequency, but TMS has some inherent defects, that is, the head or body has relevant metal inside, which cannot do TMS transcranial magnetic stimulation. Such as heart stent, cardiac pacemaker, cochlea implant, hearing aid, medical pump, etc. These metal devices may cause damage risk in the magnetic field environment of transcranial magnetic stimulation, and may also cause displacement risk.

[0046] Transcranial electrical stimulation (TES) and transcranial magnetic stimulation (TMS) are similar, which is a non-invasive technology. But compared with TMS, TES has no requirement for whether there is metal in the head or other parts of the body. TES is a kind of neuromodulation technology, which uses low-intensity current (0-2 mA) to deliver current to the specified brain area through electrodes. The current passing through the intracranial will increase or decrease the excitability of neuron cells, which is reflected in the electroencephalogram that the α brain wave increases and the δ brain wave decreases. The α brain wave represents the degree of relaxation of a person, the more relaxed and relaxed a person is, the more obvious the α brain wave is. The δ brain wave represents the state of a person's light sleep and drowsiness and fatigue. When a person is tired and drowsy, the δ brain wave will be significantly enhanced.

[0047] Transcranial electrical stimulation can adjust the experience of stimulation by adjusting the size of the stimulating current (0-2 mA), and can also adjust the way of current stimulation. Among them, tDCS (transcranial direct current stimulation) uses direct current to stimulate brain area, tACS (transcranial alternating current stimulation) uses alternating current of a certain frequency to stimulate brain area, and the specific alternating current frequency can be set to adjust in the α brain wave frequency band or the δ brain wave frequency band. tRNS (transcranial random noise stimulation) stimulates brain area by random noise, which contains a wide range of frequencies and can cover the α, β, γ, δ and θ brain wave frequency bands of human brain.

[0048] Transcranial electrical stimulation transmits special frequency current to different brain areas of the skull, thereby affecting and improving abnormal brain waves of the brain, making it return from an abnormal state to a normal state, significantly enhancing the power of α brain wave (brain relaxation state) and reducing the power of δ brain wave (brain drowsy and tired state). That is, transcranial electrical stimulation can significantly intervene in the fatigue state and effectively recover the fatigue state (specifically reflected in the increase of α brain wave and the decrease of δ brain wave in electroencephalogram).

[0049] For example, Figures 1 to 2As shown, the electroencephalogram-based fatigue intervention system according to the preferred embodiment of the present application can include an electroencephalogram acquisition module for acquiring the electroencephalogram signal of a target subject (such as a person with fatigue symptoms, etc., which can be a driver, a worker, a student, etc.). In order to obtain the electroencephalogram signal, the module should include a sensitive electrode (i.e., a first electrode) for acquiring the electroencephalogram, which, in the working state, can be close to the target's scalp to respond to the electroencephalogram signal generated by the current brain physiological activity to obtain the electroencephalogram data of the target. Such an electroencephalogram acquisition module can be the prior art in the field, such as using an electroencephalogram cap, etc., as long as it can obtain good response accuracy.

[0050] In order to realize the process of electrostimulation fatigue intervention, a transcranial electrostimulation module can also be included, which should include a stimulation electrode (i.e., a second electrode) for implementing electrostimulation on the corresponding region of the head, which is used for the system to respond to the electroencephalogram data obtained by the electroencephalogram acquisition module, and the feedback stimulation instruction is used to implement appropriate transcranial electrostimulation on the target to achieve the purpose of fatigue intervention.

[0051] A processing module for analyzing and processing the electroencephalogram data acquired by the electroencephalogram acquisition module can also be included. The processing module can perform processing on the relevant electroencephalogram data, including but not limited to: motion artifact removal, filtering, notch preprocessing, sample entropy feature value extraction, and SVM machine learning.

[0052] As an embodiment, the data processing process can be completely performed by a cloud server, at which time the local only completes the process of electroencephalogram information acquisition and electroencephalogram stimulation, i.e., in the working process, the electroencephalogram signal of the target is first acquired by the electroencephalogram acquisition module, and the electroencephalogram data is uploaded to the cloud server through a wireless local area network or the Internet. The cloud server remotely implements operation and processing such as motion artifact removal, filtering, notch preprocessing, sample entropy feature value extraction, and SVM machine learning, and matches the database to obtain and download the stimulation instruction. The stimulation instruction generally includes regional coding, stimulation region site, stimulation current type, current size strength, stimulation time, etc. in execution.

[0053] Of course, as an embodiment, the data processing process can also be distributed operation by the local terminal such as mobile phone, local computer host and the like cooperating with the cloud server. At this time, the local completes the process of electroencephalogram information collection, part of operation and electroencephalogram stimulation, that is, in the working process, the electroencephalogram signal of the target is collected by the electroencephalogram collection module, the local terminal distributes the task according to the preset, part of which is operated locally, and the part which needs to be operated by the cloud server is uploaded to the cloud server through the wireless local area network or the Internet, and the cloud server is remotely implemented. The operation content includes but is not limited to: secret key or identity allocation, motion artifact removal, filtering, notch wave preprocessing, sample entropy feature value extraction and SVM machine learning operation processing, and matching the database to obtain and download the stimulation instruction. The stimulation instruction generally includes secret key or identity code for realizing two-part data pairing, stimulation area site, stimulation current type, current size, stimulation time and the like in execution.

[0054] Of course, as an embodiment, the cloud server can also perform SVM classifier, thereby establishing the database of electroencephalogram data-fatigue type-stimulation type matching, and improving the response ability of the system.

[0055] Of course, in the above scheme, the first electrode and the second electrode can be independent of each other, or can be combined into a common electrode.

[0056] As an embodiment, since the communication and transmission of data are required in the implementation process, a communication module for implementing data transmission can also be provided, and the transmission content includes but is not limited to electroencephalogram data or stimulation instruction or operation data and other data information meeting the system operation requirements and technical scheme implementation.

[0057] As an embodiment, the communication model can be a wired communication module such as an analog communication network such as a telephone communication network, a digital communication network such as an optical fiber communication system and the like. It can also be a wireless communication network such as a mobile communication network taking a mobile phone as a carrier, a wifi communication module, an NFC communication module, a Bluetooth communication module, an infrared communication module, a laser communication module and the like.

[0058] In an embodiment of the present application, specifically, the fatigue intervention system device can include four modules.

[0059] The first is an electroencephalogram acquisition module. The electroencephalogram acquisition module and the transcranial electrical stimulation module share electrodes of an electroencephalogram cap. Because the electrodes are shared, in the electroencephalogram acquisition mode, the transcranial electrical stimulation module does not work (because the electroencephalogram data of the human brain needs to be collected through the electrodes, if the transcranial electrical stimulation module works at this time, current will be introduced through the electrodes, that is, noise will be introduced, which will affect the collection of the electroencephalogram signal, and the collected electroencephalogram signal will be submerged in the signal of the stimulation current), and in the transcranial electrical stimulation mode, the electroencephalogram acquisition module does not work. The electroencephalogram cap supports configurations of 2, 4, 8, 32 leads, etc. In addition to being responsible for the collection of electroencephalogram data, the electroencephalogram acquisition module can also be integrated with a nine-axis motion sensor, which is mainly used to collect artifacts generated by the movement of the human body in a natural environment, for subsequent removal of motion artifact noise.

[0060] The second module is a mobile phone APP module. The electroencephalogram acquisition module is responsible for transmitting the collected electroencephalogram data and nine-axis sensor data to the mobile phone APP through Bluetooth wireless mode, and the mobile phone APP can be responsible for uploading the original electroencephalogram data and nine-axis sensor data to the cloud server through 4G or wifi local area network, etc. At this time, the mobile phone participates in the information transfer of the communication module. And the cloud server is responsible for algorithm processing of the original electroencephalogram data, and the mobile phone can obtain the processed information, so as to display the current state such as fatigue state, intervention state, device running state, etc. in real time.

[0061] The third module is a cloud server module. The cloud server module mainly performs algorithm processing on the electroencephalogram data to obtain the fatigue degree of the human. First, the original electroencephalogram data is removed from the motion artifact through the nine-axis sensor data, and then filtering and notch preprocessing are performed. Then, the sample entropy is extracted as the characteristic value of the fatigue degree. Finally, the fatigue degree of the human is obtained through the SVM machine learning algorithm binary classification. And the fatigue degree is fed back to the mobile phone APP through the network. The mobile phone APP module can display the fatigue degree, etc.

[0062] The fourth module is a transcranial electrical stimulation module. After the mobile phone APP module receives the fatigue degree state returned by the cloud server, the fatigue degree state is returned to the electroencephalogram acquisition module through Bluetooth, and the electroencephalogram acquisition module decides whether to start the transcranial electrical stimulation according to the fatigue degree. In the present scheme, three transcranial electrical stimulation schemes are provided, which are transcranial direct current stimulation (tDCS), transcranial alternating current stimulation (tACS), and transcranial random noise stimulation (tRNS). At the same time, the transcranial electrical stimulation module can also adjust the size of the stimulation current, and when the stimulation mode is tACS, the specific alternating current stimulation frequency can be set.

[0063] The implementation process of one embodiment of the present application is as follows, as shown in Figure 1 and 2 ​

[0064] Step 1: The electroencephalogram acquisition module and the transcranial electrical stimulation module are connected to the electroencephalogram cap

[0065] Step 2: The electroencephalogram cap is installed with sampling electrodes according to the requirements. It can be configured with 2, 4, 8, or 32 leads according to actual needs. The sampling electrodes can be used as positive and negative electrodes for transcranial electrical stimulation. At the same time, reference and ground electrodes are additionally installed.

[0066] Step 3: Set the stimulation mode of the transcranial electrical stimulation module according to the requirements: tDCS, tACS, or tRNS. Set the stimulation current size of the transcranial electrical stimulation (0-2mA range). When the stimulation mode is tACS, the specific alternating current stimulation frequency can be set, otherwise the default alternating current frequency is used.

[0067] Step 4: Wear the electroencephalogram cap and check the impedance of each lead. Ensure that each electrode is in good contact with the scalp. Turn on the power and collect electroencephalogram data. The transcranial electrical stimulation module is in the default idle mode.

[0068] Step 5: Open the mobile phone APP and turn on the Bluetooth to establish a Bluetooth connection with the electroencephalogram acquisition module. After the Bluetooth connection is successfully established, start receiving electroencephalogram data and nine-axis sensor data. The electroencephalogram acquisition module transmits the electroencephalogram data and nine-axis sensor data to the mobile phone APP through Bluetooth.

[0069] Step 6: The mobile phone APP uploads the raw electroencephalogram data and nine-axis sensor data to the cloud server for algorithm processing through 4G or Wi-Fi. The cloud server removes motion artifacts, filters, and notch filters for preprocessing, then extracts sample entropy as a fatigue degree feature value, and finally uses machine learning algorithm to classify and obtain fatigue degree, and then transmits the fatigue degree to the mobile phone APP through the public network.

[0070] Step 7: The mobile phone APP receives the fatigue degree transmitted by the cloud server, dynamically displays it on the mobile phone APP, and transmits the fatigue degree value to the electroencephalogram acquisition module through Bluetooth. After receiving the fatigue degree value, the electroencephalogram acquisition module communicates with the transcranial electrical stimulation module to determine whether the transcranial electrical stimulation module works or not.

[0071] Step 8: According to the situation of Step 7, if the transcranial electrical stimulation module is working, the electroencephalogram acquisition module does not work temporarily, does not collect data, and does not upload electroencephalogram data and nine-axis sensor data to the mobile phone APP module. After the transcranial electrical stimulation module stimulates for a period of time, it will improve the brain alpha and delta waves, at which point it stops working and switches to the electroencephalogram acquisition module to continue working.

[0072] Step 9: After transcranial electrical stimulation, repeat the process of Step 5-Step 9 of electroencephalogram data collection, upload, and algorithm processing.

[0073] Figure 3 is a transcranial electrical stimulation signal model, which reflects various stimulation paradigms of transcranial electrical stimulation, including signal frequency, amplitude, etc., reflecting changes over time and affecting the waveband characteristics of the brain:

[0074] wherein Figure 3 In a, the tDCS transcranial direct current stimulation signal model is selected as 1mA. The current intensity can be selected in the range of 0-2mA. The current has no fluctuation. The stimulation does not change over time.

[0075] Figure 3 In b, the tACS transcranial alternating current stimulation signal model is 3HZ (mainly stimulating the delta wave band of the human brain, and the stimulation frequency range can be adjusted in the range of 1-4HZ). The current intensity is 0-2mA.

[0076] Figure 3 In c, the tACS transcranial alternating current stimulation signal model is 10HZ (mainly stimulating the alpha wave band of the human brain, and the stimulation frequency range can be adjusted in the range of 8-12HZ). The current intensity is 0-2mA.

[0077] Figure 3 In d, the tRNS transcranial random noise stimulation signal model has no fixed frequency, and the frequency range covers a wide range (1-100HZ can be covered, i.e. δ wave, θ wave, α wave, β wave and γ wave can be covered), the current intensity is in the range of 0-2mA, and the stimulation current amplitude changes over time (randomly).

[0078] One innovation of the present application is the process of obtaining the fatigue degree characteristic value (i.e. the determination of sample entropy as the fatigue degree characteristic value) and the process of binary classification judgment of fatigue degree accuracy in machine learning.

[0079] The characteristic value (sample entropy) of fatigue degree can be obtained by the following steps:

[0080] Step A: According to the obtained electroencephalogram data, a time series data Se with a length of L is constructed, as follows:

[0081] Se(t)=Se(1),Se(2),Se(3)......,Se(L)

[0082] StepB: Select a window size N (in the actual algorithm calculation process, N can be set to 2 or 3), according to the electroencephalogram time series data Se(t) constructed in Step A and the window size N, two vectors X and Y are constructed as follows. The row dimension of the two vectors is based on the window size N, and the column dimension is based on the maximum time series length minus the window size, i.e. (L-N):

[0083]

[0084]

[0085] StepC: According to the following formula, calculate the number of vector pairs in the X vector and the Y vector in Step B whose distance is less than f, where f is the filter parameter and N is the window size. Where is the distance of the vector pair in the X vector and the Y vector, then the distance obtained is added to the absolute value, to determine whether the distance of the obtained vector pair is less than the filter parameter f, if it is less than, count plus 1, indicating that a vector pair that meets the requirements is obtained. This operation is to count and add 1 to the vector pairs that meet the condition of distance less than f. Each element pair in the X vector and the Y vector is counted, where the subscript range is from 1 to (L-N), that is, the column dimension length of the vector is counted.

[0086]

[0087] StepD: Add 1 to N in StepC (i.e. the column range is 1 to (L-N+1)), repeat the steps of StepB-StepC. That is, according to the row and column dimension calculation formula of the vector X in StepB, new vectors X and Y are constructed according to the window size (N+1) and the electroencephalogram time series in StepA, and the number of vector pairs in the new vectors X and Y whose distance is less than f is calculated.

[0088]

[0089] StepE: Sample entropy SampleEntropy can be obtained by the number of vector pairs Size1 obtained in StepC and the number of vector pairs Size2 obtained in StepD, and the inverse of lg is obtained, as shown in the following formula. Size2 and Size1 represent the probability of similarity between two sequences. Sample entropy is affected by window size N and filter parameter f.

[0090]

[0091] Specifically, the SVM classifier can be:

[0092] The best choice for testing fatigue is to choose people who work night shifts in hospitals or factories, but because the test time span is relatively long (usually more than 8 hours, even 12 hours), the test cost is relatively high, so the simulation driving is chosen as the fatigue test paradigm. According to the regulations of the traffic department, people will feel tired after driving for 3 hours, at which time they need to choose to rest.

[0093] Before the test, the EEG data of the subjects is collected for a period of time (about 5-10 minutes), and then the EEG data is marked as true normal state (TNS: true Normal State). That is, during this period, the brain of the person is in a normal state, not a fatigue state. This state is true, not false.

[0094] Then let the subjects drive for 3 hours in a simulated normal driving test. After 3 hours, the EEG data of the subjects is collected for 5-10 minutes, and this EEG data is marked as true fatigue state (TFS: true fatigue State). That is, after 3 hours of driving, the brain of the person is in a fatigue state, and this fatigue state is true, not false.

[0095] Let multiple subjects perform the above simulated driving test and collect the corresponding pre-test and post-test EEG data.

[0096] Extract the sample entropy from the EEG data of these subjects, and use the data of these subjects as the input of the SVM support vector machine to train a trained model with parameters. In the obtained trained model, the SVM support vector machine may have some misjudgments. For example, before the test, the brain data of the person is in a normal state, and if the SVM misjudges it as a fatigue state, it is recorded as a false fatigue state (FFS: false fatigue State); Similarly, after the simulated driving test, the brain data of the person is in a fatigue state, and if the SVM classification determines it as a normal state, it is recorded as a false normal state (FNS: false Normal State).

[0097] When using support vector machine SVM to judge fatigue, the correct rate of detecting fatigue can be calculated by the following formula.

[0098]

[0099] In actual testing, the correct rate of SVM detecting fatigue can reach 75%, indicating that the detection effect is very good. If the actual detection accuracy is too low, the parameters of the SVM training model need to be adjusted again for classification and judgment. The subsequent actual test data can be combined with the previous test data as new input to train the SVM and correct the parameters of the SVM training model.

[0100] The foregoing description of specific exemplary embodiments of the application has been presented for the purposes of illustration and description. It is not intended to be a limitation on the broad concepts of the application. Obviously, many modifications and variations of the specific exemplary embodiments described herein are possible in light of this disclosure, and it is intended that the scope of the application be limited only by the claims appended hereto and their equivalents. Examples of selected embodiments were chosen and described in order to explain the principles of the application and its practical application. Those skilled in the art will readily appreciate that various modifications and changes can be made to the specific exemplary embodiments without departing from the scope of the application. The scope of the application is defined only and exactly by the claims.

Claims

1. A brainwave-based fatigue intervention system, characterized by, Comprising an electroencephalogram acquisition module comprising at least a first electrode for acquiring electroencephalogram data of a target through the first electrode; a transcranial electrical stimulation module comprising at least a second electrode for implementing a weak current electrical stimulation signal on the target through the second electrode according to a stimulation instruction, the electrical stimulation signal comprising a direct current stimulation signal or an alternating current stimulation signal or a random noise signal; a processing module for acquiring the electroencephalogram data fed back by the electroencephalogram acquisition module and matching the stimulation instruction according to the electroencephalogram data, the processing module further comprising an SVM classifier to obtain a stimulation instruction library matched with the electroencephalogram data and the fatigue degree, wherein a feature value of the fatigue degree is obtained through the following steps: Step A: constructing a time series data Se with a length of L according to the acquired electroencephalogram data, as shown below: ； Step B: selecting a window size N, setting N as 2 or 3, constructing two vectors X and Y according to the electroencephalogram time series data Se(t) constructed in Step A and the window size N, wherein the row dimension of the two vectors is based on the window size N and the column dimension is based on the maximum time series length minus the window size, i.e. (L-N): ； Step C: Calculate the number of vector pairs in X vector and Y vector whose distance is less than f according to the following formula, where f is the filter parameter and N is the window size. is to calculate the distance of vector pairs in X vector and Y vector, and then add the absolute value to the distance obtained to determine whether the distance of the obtained vector pair is less than the filter parameter f. If it is less than, count it by 1, indicating that a vector pair that meets the requirements is obtained; This operation is to count and add 1 to the vector pairs that meet the condition of distance less than f. Each element pair in X vector and Y vector is counted, where the subscript range is from 1 to (L-N), that is, the dimension length of the column in the vector is counted: ； Step D: adding 1 to N in Step C and repeating the steps of Step B-Step C; i.e. according to the row and column dimension calculation formula of the vector X in Step B, constructing new vectors X and Y according to the window size (N+1) and the electroencephalogram time series in Step A, and then calculating the number of vector pairs with a distance less than f in the new vectors X and Y: ； Step E: calculating lg and taking the inverse to obtain the sample entropy SampleEntropy according to the number of vector pairs Size1 obtained in Step C and the number of vector pairs Size2 obtained in Step D, as shown in the following formula; wherein Size2 and Size1 represent the probability of similarity between two sequences; the sample entropy is affected by the window size N and the filter parameter f: ； The sample entropy is taken as the input of the SVM support vector machine to obtain a training model with parameters; when the support vector machine SVM is used to judge the fatigue degree, the following formula is used to calculate the accuracy of detecting the fatigue degree: ； Wherein, before the test, the brain data of the person is in a normal state, if the SVM misjudges as a fatigue state, it is recorded as a false fatigue state FFS; similarly, after the simulation driving test, the brain data of the person is in a fatigue state, if the SVM classification determines as a normal state, it is recorded as a false normal state FNS.

2. The electroencephalogram-based fatigue intervention system of claim 1, wherein, The first electrode and the second electrode are common electrodes.

3. The electroencephalogram-based fatigue intervention system of claim 1, wherein, The processing module comprises a cloud system, at least for collecting electroencephalogram data and / or outputting processing and / or instruction matching in the cloud.

4. The electroencephalographic-based fatigue intervention system of any one of claims 1-3, wherein, The fatigue intervention system further comprises a communication module, the communication module is at least used for transmitting the electroencephalogram data or the stimulation instruction.

5. The electroencephalogram-based fatigue intervention system of claim 4, wherein, The communication module is a wired communication module or a wireless communication module.

6. The electroencephalogram-based fatigue intervention system of claim 5, wherein, The wireless communication module is at least any one of the following: a mobile communication module, a wifi communication module, a near field communication module, and an optical communication module.

7. The electroencephalogram-based fatigue intervention system of claim 6, wherein, The near field communication module comprises an NFC communication module or a Bluetooth communication module or an infrared communication module.

8. A fatigue intervention device comprising the electroencephalogram-based fatigue intervention system according to any one of claims 1-7.

9. The fatigue intervention system apparatus of claim 8, wherein, comprising a controller; an electroencephalogram cap provided with a plurality of electrodes, the electrodes comprising a first electrode and / or a second electrode; a signal acquisition device communicatively connected to at least the first electrode, for acquiring electroencephalogram data fed back by the first electrode; an electrical stimulation device comprising at least a stimulation source configured to release a weak current to the second electrode according to stimulation instructions of the controller.

Images