CN101596101B - Method for determining fatigue state according to electroencephalogram - Google Patents
Method for determining fatigue state according to electroencephalogram Download PDFInfo
- Publication number
- CN101596101B CN101596101B CN2009100889147A CN200910088914A CN101596101B CN 101596101 B CN101596101 B CN 101596101B CN 2009100889147 A CN2009100889147 A CN 2009100889147A CN 200910088914 A CN200910088914 A CN 200910088914A CN 101596101 B CN101596101 B CN 101596101B
- Authority
- CN
- China
- Prior art keywords
- eeg signals
- data
- brain
- fatigue state
- signal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Landscapes
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention provides a method for determining fatigue state according to electroencephalogram (EEG) which adopts a plurality of electroencephalographs and connecting electrodes for realizing the real time acquisition of electroencephalogram. The method comprises the following steps: running interface programs of a PC and the electroencephalographs; realizing the synchronous acquisition of data by using a VC++ to compile visual interface program of the electroencephalographs under the Windows platform, and displaying EEG waveforms acquired in real-time; pre-processing the acquired data; carrying out the low-pass filtering at 0Hz to 30Hz to the data by an FIR (Finite Impulse Response) filter, so as to eliminate the power frequency noise and external interference; decomposing the filtered EEG waveforms by the blind-source separation method, so as to acquire each component of the mixed signal comprising electro-oculogram (EOG) and left and right brain EEGs; carrying out the fast Fourier transform (FFT) on the left and right brain EEGs, and converting the time-domain signals to the frequency-domain signals; working out the energy of alpha, beta, theta and delta waves in the EEGs and classifying the BP (back propagation) neural network of the multi-layer perceptron. The invention has the characteristics of directness and rapidness.
Description
Technical field
The present invention relates to brain-computer interface (brain-computer interface, BCI) E.E.G α, β, θ, δ divide the extracting method of energy, feature extraction and sorting technique that particularly a kind of independent component analysis method combines with the BP neutral net in the device.
Background technology
Brain-computer interface (BCI) is the direct connecting path of setting up between human or animal's brain (the perhaps culture of brain cell) and external equipment.Under the situation of unidirectional brain-computer interface, computer or accept the order that brain transmits perhaps transmits a signal to brain (for example video reconstruction), but can not send simultaneously and received signal.And two-way brain-computer interface allows the bi-directional exchanges of information between brain and external equipment.
Brain-computer interface is divided into intrusive mood brain-computer interface, part intrusive mood brain-computer interface and non-intrusion type brain-computer interface again.
The intrusive mood brain-computer interface is mainly used in the motor function of rebuilding the special sense (for example vision) and paralytic.This type of brain-computer interface directly is implanted to the grey matter of brain usually, thereby the mass ratio of the nerve signal that is obtained is higher.But its shortcoming is to cause immunoreation and callus (scar) easily, and then causes the decline even the disappearance of signal quality.
Part intrusive mood brain-computer interface generally is implanted in the cranial cavity, but is positioned at outside the grey matter.Its spatial resolution still is better than the non-intrusion type brain-computer interface not as the intrusive mood brain-computer interface.Its another advantage is that the probability of initiation immunoreation and callus is less.The technical foundation of electrocorticogram (ECoG) and electroencephalogram similar, but its electrode directly is implanted on the cerebral cortex subdural zone.
The non-intrusion type brain-computer interface be by be attached to electrode record EEG signals on the scalp surface (Electroencephalogram, EEG).Though the device of this non-intrusion type conveniently is worn on human body, because skull is to the attenuation of signal with to electromagnetic dispersion and blurring effect that neuron sends, the resolution that records signal is not high.In general, the EEG EEG signals has following characteristics:
(1) amplitude is faint, but ambient noise signal is strong.Though the EEG signal can be detected, be difficult to determine to send the brain district of signal or relevant single neuronic discharge.And the wave amplitude of the brain potential that occurs on scalp fluctuation is very low, generally about 50 μ V, and maximum 100 μ V.Need to amplify 1,000,000 times and just can record, be easy to be subjected to interferential influences such as electrocardio (ECG), myoelectricity (EMG), eyespot (EOG), perspiration and external environment.
(2) non-stationary strong with randomness.Mechanism of production and rule thereof for EEG signals still do not have clear and definite understanding so far.A lot of preliminary rules can only be analyzed from the statistics angle.The factor that influences EEG signals is numerous, and because the activity of brain is very active and be subjected to the influence that surrounding even experimenter self thought change very easily, and it is very strong non-stationary that this makes that EEG signals shows.Studies show that the length of EEG signals increases to 10s from 1s, its stationarity reduces to 10% by 90%, and therefore non-stationary signal analysis method is significant in EEG Processing.
(3) non-linear.The self regulation of biological tissue and adaptation mechanism make electro-physiological signals have nonlinear feature.Because traditional signal processing technology is to be based upon on the analysis foundation of lineary system theory basically.How handling the error that nonlinear properties bring, also is the problem that must note in the EEG Processing.
(4) frequency domain character of EEG signals is more outstanding.Therefore the frequency-domain analysis signal processing method based on spectrum occupies prior status in the electroencephalogramsignal signal analyzing method.
(5) lead signal more.General EEG signals all adopts multilead electrode to measure, certain interactive information is arranged between each lead signals, and how effectively to utilize these to lie in the key character of leading between the EEG signals is a major criterion setting up and estimate brain-electrical signal processing method more.
At present, multiple brain-electrical signal processing method has been arranged,, roughly can be divided into following a few class according to the difference of the angle of setting out:
(1) time-domain analysis method.Early stage EEG signals adopts the time-domain analysis method mostly, obtains some temporal signatures by the analysis to the EEG signals time domain waveform, as variance analysis, correlation analysis, peak value detection, zero passage detection etc.
(2) frequency-domain analysis method.Mainly refer to power spectrum analysis method,, obtain the frequency domain character of EEG signals by EEG signals is carried out power spectrumanalysis.Estimate (periodogram), modern spectrum estimation (parameter model estimation) etc. as Classical Spectrum.
(3) Time-Frequency Analysis Method.With the wavelet transformation is the new signal analysis method of main representative.In the non-stationary signal analysis field, occupy critical role.Because EEG signals is non-stationary very strong, Time-Frequency Analysis Method such as wavelet transformation have been subjected in the EEG Processing field paying attention to widely.
(4) nonlinear analysis method.From nonlinear motion mechanics angle, extract the non-linear dynamic mathematic(al) parameter of EEG signals.As analysis of complexity, correlation dimension, Lorenz scatterplot, Kolmogorov entropy, Li Ya spectrum promise husband index etc.
(5) multidimensional statistics analytical method.Because the multidimensional statistics analysis method characteristic is can handle simultaneously to lead EEG signals more, therefore more helps disclosing the hidden feature in the EEG signals, and have unique effect at aspects such as brain electricity de-noising and feature extractions.As principal component analysis (Principal ComponentAnalysis, PCA), independent component analysis (IndependentComponentAnalysis, ICA), factorial analysis (FactorAnalysis), the common space filtering method (Common SpecialPattern, CSP) etc.
Wherein the basic theories of independent component analysis (ICA) method was set up on higher order statistical analysis theories and the information theory basis in modern age.Its object function is that foundation and Higher Order Cumulants, the very big transmission theory of information, central limit theorem, maximal possibility estimation, Mutual Information Theory, maximum entropy theorem etc. have confidential relation.ICA solves the isolating a kind of effective method of blind source signal.
Therefore, demand developing a kind of method of judging fatigue state according to EEG signals urgently, with existing initiation immunoreation of the method that changes present judgement fatigue state and callus, not high, the jitter of resolution that records signal and the problem of poor anti jamming capability, make that it has directly, stablizes, characteristics fast.
Summary of the invention
The objective of the invention is to, by a kind of method that can obtain left and right sides brain EEG E.E.G is provided, and the active degree of α, β, θ, δ ripple wherein differentiated, and then accurately judge tired degree.
The present invention realizes by the following technical solutions:
Signal collecting device adopts 16 to lead electroencephalograph, and international 10-20 standard is adopted in the placement of electrode, and sample frequency is 1KHz.EEG signals is imported computer with the voltage magnitude of EEG signals via the USB serial port by after amplification, the A/D conversion.Adopt independent component analysis method that the data that collect are carried out the classification of blind source, extract corresponding characteristic signal, re-use the BP neutral net corresponding signal is carried out Classification and Identification,, and export result of determination at last by the self study ordering parameter.
These method concrete steps are as follows:
1. adopt 16 to lead electroencephalograph and carry out EEG signals and gather in real time.
Use 16 to lead electroencephalograph, connect 18 electrodes, carry out EEG signals and gather in real time by Fig. 2 mode.
2. design PC and electroencephalograph interface routine.
Utilize VC++ to write under the windows platform and electroencephalograph visualization interface program, realize the data synchronization collection.And can show the electroencephalogram waveform of real-time collection.
Program adopts MFC single document application framework to make up, and this application program has the general all basic elements of application program such as document window, program frame, title bar, menu bar, toolbar, status bar.
3. the data that collect are carried out pretreatment.
Because low-frequency range mainly appears in the signal of being concerned about, therefore designed 512 sampled points, the FIR wave filter on 48 rank carries out the 0-35Hz low-pass filtering to data, to remove industrial frequency noise and external disturbance.
4. adopt the ICA method that the E.E.G through filtering is decomposed.
ICA, full name Independent ComponentAnalysis is a kind of blind source separation method based on independence.The EEG data are made up of the current potential of the electrode record of one group of many places diverse location that is placed on scalp, and these current potentials can think that then the signal of telecommunication that the signal of telecommunication that produced by cerebral activity and some musculation produce mixes generation.Utilize the ICA method, can obtain each composition of these mixed signals.Wherein include electro-oculogram (EOG), left and right sides brain electroencephalogram (EEG).
5. to resulting left and right sides brain electroencephalogram EEG L, EEGR carries out fast Fourier transform (FFT).Be transformed into frequency-region signal from time-domain signal.Be that discrete-time series is converted into discrete frequency domain sequence X (n) → X (k).
6. obtain the energy of α, β in the EEG ripple, θ, δ ripple.
Wherein α ripple frequency band is 8-13Hz, when the α ripple is the advantage E.E.G, and people's Consciousness, but health but loosens.In this state, the energy that body and mind expends is all fewer, and the energy that relative brain obtains is more, so brain is more flexible, reaction is also relatively quicker, action is more smooth and easy.
β ripple frequency band 13-30Hz, when the β ripple was the advantage E.E.G, brain presented a kind of tense situation, and health is prepared external environment condition is made a response at any time.In this state, people's body and mind energy charge is huge, and is very fast easily tired.
θ ripple frequency band is 4-8Hz, and when the θ ripple was the advantage E.E.G, people's confusion, health entered degree of depth relaxation state.
δ ripple frequency band is 2-4Hz, when the δ ripple is the advantage E.E.G, is deep sleep, automatism.
So, can be according to the property determination people's of prevailing ripple in the E.E.G state of consciousness, and then judge whether fatigue.
7. design BP neutral net is classified.The BP network is actually a multilayer perceptron, is made of input layer, hidden layer (intermediate layer), output layer.The BP network has feedback and feed forward mechanism simultaneously, and in a cycle of training of network, the output of network feeds back to the outside input of the input neuron of network as network simultaneously, and promptly error is back propagated study mechanism (Back-propagation Learning).Through repetition learning, revise the bond strength value between the neuronic deviation value and neuron in the BP network repeatedly, finally make its performance function be tending towards minimum, when error during less than allowed band, training finishes.
A kind of method according to EEG signals judgement fatigue state of the present invention compared with prior art, has following remarkable advantages and beneficial effect:
The present invention judges the method for fatigue state according to EEG signals, and by the training of good clear-headed, the tired training data set pair network of prior collection, the BP network is revised each weights, threshold value automatically, finishes up to training.Method and fast Fourier transform with the FastICA independent component analysis, and ask the relative energy of α, β in the decerebrate ripple, θ, δ ripple E.E.G to be analyzed as the input data of BP neutral net input neuron, and then judge the method for brain fag state, have directly, characteristics fast.Compare with the method that traditional images is handled, have significant creativeness and practicality.
Description of drawings
Fig. 1 is a BCI system structure sketch map;
Fig. 2 is the electrode for encephalograms distribution schematic diagram;
Fig. 3 is that the ICA algorithm separates 4 dimension source signal sketch maps by 16 dimension mixed signals;
Fig. 4 transfers the frequency-region signal sketch map to for the EEG time-domain signal after the FFT conversion;
Fig. 5 is a BP neural network structure sketch map;
Fig. 6 carries out fatigue state decision-making system flow chart for the present invention is based on EEG signals.
The specific embodiment
Below in conjunction with accompanying drawing specific embodiments of the invention are described further.
See also shown in Figure 1ly, be BCI system structure sketch map.As can be seen from the figure, the EEG signals of brain is connected with external equipment by control signal after passing through signals collecting, signal processing, pattern recognition.
See also shown in Figure 2ly, be the electrode for encephalograms distribution schematic diagram.At first lead electroencephalograph with 16 and connect by mode among the figure, EEG signals is gathered in real time, sample frequency is 1KHz.Signal through amplifier amplify, after the A/D conversion, by the USB oral instructions to PC.
The data of electroencephalograph being sent by the PC end windows platform application of being write by Visual C++6.0 receive, handle.This program is at first carried out initialization to electroencephalograph, corresponding button or menucommand are waited in the demonstration man machine interface on computer screen then, after receiving the button or menucommand that begins to gather, every 200 milliseconds to electroencephalograph in the data buffer zone read, be saved in corresponding array and the file.
The signal of gathering is carried out pretreatment, i.e. low-pass filtering, because the EEG signals frequency of being concerned about in the 0-35Hz scope, designs 48 rank, the FIR low pass filter of 512 sampled points is to remove external disturbance such as industrial frequency noise.Concrete wave filter M file can be designed to: sample frequency FS is set to 512, and exponent number N is set to 48, and passband initial frequency Fpass is set to 1, and cut-off frequency Fstop is set to 35.Damping is set to 80 decibels.
Separate carry out the blind source of FastICA algorithm through pretreated signal.Fig. 3 is that the ICA algorithm separates 4 dimension source signal sketch maps by 16 dimension mixed signals; Fig. 4 transfers the frequency-region signal sketch map to for the EEG time-domain signal after the FFT conversion.This algorithm is a kind of quick, well behaved independent component analysis algorithm based on the negentropy criterion of improved.The FastICA algorithm at first goes average and whitening pretreatment to signal, going average to handle makes observation signal become the zero-mean variable, this pretreatment is the step for shortcut calculation, after estimating separation matrix W, can on the estimated value y of source signal s, adding this average, so go the average pretreatment and do not mean that average can't estimate.And albefaction is handled, and will make that the observation signal through after the albefaction is uncorrelated, and has unit variance.Through after the pretreatment, the FastICA algorithm uses the fixing point iteration theory to seek W
TNon-Gauss's maximum of X is promptly obtained y
i=W
TThe projecting direction of X makes its Gauss's minimum.By using Newton iterative that a large amount of sampled points of observational variable X are carried out batch processing, and then from observation signal X, divide the luxuriant isolated component y of going out at every turn
i
Through the iterative cycles several times, just can obtain whole separation matrix W, and isolated each y
i(y
iBe an independent source signal s
1Estimated value, i ≠ j generally speaking, this is caused by source signal S and the equal the unknown of former hybrid matrix A)
After FastICA algorithm separation, can get 4 groups of source signal components to 16 dimensional signals.Select EEG signal wherein, it is carried out fast Fourier transform (FFT).Fast Fourier transform (FFT) is method very commonly used in Digital Signal Analysis and Processing as a kind of fast algorithm of discrete Fourier transform (DFT).
Calculate the energy of each wavelet in the E.E.G.Try to achieve special frequency band pairing serial number k value by the y (k) that time domain y (t) is transformed to frequency domain by Δ 1 through EEG signal sequence after the FFT conversion.Utilize formula k * Δ f=f, try to achieve 2~4 frequency bands (corresponding δ ripple), 4~8 frequency bands (corresponding θ ripple), 8~13 frequency bands (corresponding α ripple), the corresponding k value of 13~30 frequency bands (corresponding β ripple).Again y (k) in each frequency range is asked quadratic sum, the relative energy that has so just obtained α, β, θ, δ ripple is (respectively as x
1, x
2, x
3, x
4).
Design BP neural network classifier.The BP neutral net has good non-linear approximation capability, can approach any non-linear continuous function with arbitrary accuracy.And the BP network also has characteristics such as parallel distributed processing mode, self study adaptive ability, multi-variable system, can well be as the grader of fatigue state.
X
1, x
2, x
3, x
4As the input of BP neutral net, be output as a neuron and represent fatigue state, Fig. 5 is a BP neural network structure figure sketch map.Hidden layer is made as 1 layer, and the hidden neuron number designed according to following 2 steps:
n=√(n
i+n
o)+a
In the formula: n is a number of hidden nodes, n
iBe input number of nodes, n
oBe the output node number, a is the constant between 1~10.
Change n (being that n changes to 13 from 3),, therefrom determine the number of hidden nodes of hour correspondence of network error with same sample set training.
Fig. 6 carries out fatigue state decision-making system flow chart for the present invention is based on EEG signals.As can be seen from the figure, the signal of the described several steps of the foregoing description.
Claims (6)
1. the method according to EEG signals judgement fatigue state adopts electroencephalograph to carry out EEG signals and gathers in real time, uses and leads electroencephalograph more, and connection electrode is carried out EEG signals and gathered in real time; It is characterized in that: may further comprise the steps:
1.1. operation PC and electroencephalograph interface routine;
Utilize VC++ to write under the windows platform and electroencephalograph visualization interface program, realize the synchronous acquisition of EEG signals potential data, and show the electroencephalogram waveform of gathering in real time;
1.2. the data that collect are carried out pretreatment;
Adopt the FIR wave filter that data are carried out the 0-30Hz low-pass filtering, to remove industrial frequency noise and external disturbance;
1.3 adopt the independent component analysis blind source separation method that the E.E.G data through filtering are decomposed, obtain each composition of mixed signal, include electro-oculogram, left and right sides brain electroencephalogram;
1.4. resulting left and right sides brain electroencephalogram is carried out fast Fourier transform, is transformed into frequency-region signal from time-domain signal;
1.5. obtain the energy of α, β, θ, δ ripple in the E.E.G data that frequency domain passed through fast Fourier transform;
Wherein α ripple frequency band is the more 8-13Hz of energy that brain obtains;
β ripple frequency band is the 13-30Hz that brain presents a kind of tense situation;
The 4-8Hz of the confusion that θ ripple frequency band is behaved;
δ ripple frequency band is the 2-4Hz of deep sleep, automatism;
1.6. multilayer perceptron BP neutral net is classified; Described BP neutral net comprises: input layer, hidden layer, output layer; The BP network has feedback and feed forward mechanism simultaneously, and in a cycle of training of network, the output of network feeds back to the outside input of the input neuron of network as network simultaneously.
2. according to claim 1 judge according to EEG signals it is characterized in that the method for fatigue state: the described electroencephalograph numbers of leading of leading adopt 16 more.
3. the method according to EEG signals judgement fatigue state according to claim 1, it is characterized in that: described electrode is 18.
4. according to claim 1 judge according to EEG signals it is characterized in that the method for fatigue state: described the data that collect are carried out pretreatment, adopt 512 sampled points in low-frequency range, the FIR wave filter on 48 rank carries out filtering.
5. the method according to EEG signals judgement fatigue state according to claim 1 is characterized in that: described program adopts MFC single document application framework to make up.
6. according to claim 1 judge according to EEG signals it is characterized in that the method for fatigue state: described E.E.G data are made up of the current potential of the electrode record of one group of many places diverse location that is placed on scalp.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2009100889147A CN101596101B (en) | 2009-07-13 | 2009-07-13 | Method for determining fatigue state according to electroencephalogram |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2009100889147A CN101596101B (en) | 2009-07-13 | 2009-07-13 | Method for determining fatigue state according to electroencephalogram |
Publications (2)
Publication Number | Publication Date |
---|---|
CN101596101A CN101596101A (en) | 2009-12-09 |
CN101596101B true CN101596101B (en) | 2011-03-23 |
Family
ID=41417832
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2009100889147A Expired - Fee Related CN101596101B (en) | 2009-07-13 | 2009-07-13 | Method for determining fatigue state according to electroencephalogram |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN101596101B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106175752A (en) * | 2015-04-30 | 2016-12-07 | 深圳市前海览岳科技有限公司 | Eeg signal obtains Apparatus and method for, status assessing system and method |
Families Citing this family (36)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102096798B (en) * | 2011-01-25 | 2013-08-21 | 上海交通大学 | Signal processing method for electromagnetic induction state recognition |
CN102274032A (en) * | 2011-05-10 | 2011-12-14 | 北京师范大学 | Driver fatigue detection system based on electroencephalographic (EEG) signals |
CN102184415B (en) * | 2011-05-17 | 2013-01-02 | 重庆大学 | Electroencephalographic-signal-based fatigue state recognizing method |
CN103431859B (en) * | 2012-01-10 | 2014-12-10 | 西安交通大学 | Experimental method for determining brain load in multitask visual cognition |
CN103297118A (en) * | 2012-02-29 | 2013-09-11 | 深圳光启创新技术有限公司 | Visible light signal testing method and system |
CN102657526B (en) * | 2012-04-23 | 2013-12-04 | 中国医学科学院生物医学工程研究所 | Method for evaluating R value and power spectrums of electroencephalogram signals causing discomfort when people watch three-dimensional (3D) images |
TWI471120B (en) * | 2012-05-02 | 2015-02-01 | Tzu Chi University | Detecting apparatus for head physiological parameters and method for the same and application equipment for the same |
CN102920453A (en) * | 2012-10-29 | 2013-02-13 | 泰好康电子科技(福建)有限公司 | Electroencephalogram signal processing method and device |
CN103815900B (en) * | 2013-11-22 | 2015-08-05 | 刘志勇 | A kind of method of the measurement Vigilance based on brain electricity frequency domain character indexing algorithm |
CN103815901B (en) * | 2013-11-22 | 2015-09-09 | 刘志勇 | A kind of frequency domain character extracting method being applied to the portable brain electric equipment that singly leads |
CN103876734B (en) * | 2014-03-24 | 2015-09-02 | 北京工业大学 | A kind of EEG signals feature selection approach based on decision tree |
CN104473635B (en) * | 2014-12-16 | 2016-07-06 | 重庆邮电大学 | Right-hand man's Mental imagery EEG feature extraction method of hybrid wavelet and common space pattern |
CN104461017A (en) * | 2014-12-31 | 2015-03-25 | 北京智谷睿拓技术服务有限公司 | Interactive method, interactive device and user equipment |
CN106175799A (en) * | 2015-04-30 | 2016-12-07 | 深圳市前海览岳科技有限公司 | Based on brain wave assessment human body emotion and the method and system of fatigue state |
US20170185149A1 (en) * | 2015-12-26 | 2017-06-29 | Intel Corporation | Single sensor brain wave monitor |
CN106037724A (en) * | 2016-06-24 | 2016-10-26 | 深圳市宇通联发科技有限公司 | Bone conduction feedback device based on brain wave anti-fatigue monitoring |
CN106504475B (en) * | 2016-10-15 | 2019-05-10 | 北海益生源农贸有限责任公司 | Method for detecting fatigue driving based on EEG signals |
CN106510700A (en) * | 2016-11-28 | 2017-03-22 | 邹敏伟 | Brain wave super-slow encephalofluctuo monitoring device and method thereof |
CN106691440A (en) * | 2016-12-07 | 2017-05-24 | 中国民用航空总局第二研究所 | Controller fatigue detection method and system based on BP neural network |
CN108078563A (en) * | 2017-01-11 | 2018-05-29 | 浙江师范大学 | A kind of EEG signal analysis method of integrated classifier |
CN107595302B (en) * | 2017-09-06 | 2021-01-19 | 西安交通大学 | Device and method for monitoring mental state of user by electroencephalogram signals |
CN107729676B (en) * | 2017-10-31 | 2020-10-30 | 中国地质大学(武汉) | Correlation analysis method for blast furnace operation parameters, state variables and CO utilization rate |
CN108415559B (en) * | 2018-02-11 | 2021-08-31 | Oppo广东移动通信有限公司 | Interaction method and related product |
CN108681392A (en) * | 2018-03-30 | 2018-10-19 | 广东欧珀移动通信有限公司 | Information query method and Related product |
CN108520239B (en) * | 2018-04-10 | 2021-05-07 | 哈尔滨理工大学 | Electroencephalogram signal classification method and system |
CN108836321A (en) * | 2018-05-03 | 2018-11-20 | 江苏师范大学 | A kind of EEG signals preprocess method based on adaptive noise cancel- ation system |
CN109124625B (en) * | 2018-09-04 | 2021-07-20 | 大连理工大学 | Driver fatigue state level grading method |
CN109472224A (en) * | 2018-10-26 | 2019-03-15 | 蓝色传感(北京)科技有限公司 | The fatigue driving detecting system merged based on EEG with EOG |
CN109567789B (en) * | 2018-12-03 | 2021-08-31 | 东软集团股份有限公司 | Electrocardiogram data segmentation processing method and device and readable storage medium |
CN110367967B (en) * | 2019-07-19 | 2021-11-12 | 南京邮电大学 | Portable lightweight human brain state detection method based on data fusion |
CN113059480B (en) * | 2019-12-31 | 2022-07-29 | 清华大学 | Chemical mechanical polishing device and control method thereof |
CN111528839B (en) * | 2020-05-29 | 2023-06-23 | 北京京东方健康科技有限公司 | Sleep detection method and device, sleep assisting equipment and method |
CN112622277A (en) * | 2020-12-16 | 2021-04-09 | 深圳市好智力教育科技有限公司 | Brain-controlled printing system |
CN113180705A (en) * | 2021-04-07 | 2021-07-30 | 北京脑陆科技有限公司 | Fatigue detection method and system based on EEG brain waves |
CN113674756B (en) * | 2021-10-22 | 2022-01-25 | 青岛科技大学 | Frequency domain blind source separation method based on short-time Fourier transform and BP neural network |
CN114288520A (en) * | 2021-12-31 | 2022-04-08 | 广州酷狗计算机科技有限公司 | Sleep assisting method, device, equipment and storage medium based on brain waves |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1613425A (en) * | 2004-09-15 | 2005-05-11 | 南京大学 | Method and system for drivers' fatigue prealarming biological identification |
CN101243973A (en) * | 2008-01-31 | 2008-08-20 | 杨杰 | Method and apparatus for monitoring and awakening fatigue doze |
-
2009
- 2009-07-13 CN CN2009100889147A patent/CN101596101B/en not_active Expired - Fee Related
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1613425A (en) * | 2004-09-15 | 2005-05-11 | 南京大学 | Method and system for drivers' fatigue prealarming biological identification |
CN101243973A (en) * | 2008-01-31 | 2008-08-20 | 杨杰 | Method and apparatus for monitoring and awakening fatigue doze |
Non-Patent Citations (2)
Title |
---|
张崇,郑崇勋等.基于多导脑电特征的生理性精神疲劳分析.《西安交通大学学报》.2007,第41卷(第2期),250-254. * |
王黎,于涛等.基于脑电a波的非线性参数人体疲劳状态判定.《东北大学学报》.2005,第26卷(第12期),1174-1177. * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106175752A (en) * | 2015-04-30 | 2016-12-07 | 深圳市前海览岳科技有限公司 | Eeg signal obtains Apparatus and method for, status assessing system and method |
CN106175752B (en) * | 2015-04-30 | 2020-12-01 | 浙江美迪克医疗科技有限公司 | Brain wave signal acquisition device and method, and state evaluation system and method |
Also Published As
Publication number | Publication date |
---|---|
CN101596101A (en) | 2009-12-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN101596101B (en) | Method for determining fatigue state according to electroencephalogram | |
Mishra et al. | An efficient method for analysis of EMG signals using improved empirical mode decomposition | |
CN102940490B (en) | Method for extracting motor imagery electroencephalogram signal feature based on non-linear dynamics | |
CN101515200B (en) | Target selecting method based on transient visual evoked electroencephalogram | |
CN111067514B (en) | Multi-channel electroencephalogram coupling analysis method based on multi-scale multivariable transfer entropy | |
CN102835955A (en) | Method of automatically removing ocular artifacts from electroencephalogram signal without setting threshold value | |
CN114533086B (en) | Motor imagery brain electrolysis code method based on airspace characteristic time-frequency transformation | |
CN111820876B (en) | Dynamic construction method of electroencephalogram spatial filter | |
CN112137616B (en) | Consciousness detection device for multi-sense brain-body combined stimulation | |
CN113536882B (en) | Multi-class motor imagery electroencephalogram signal feature extraction and classification method | |
Sarin et al. | Automated ocular artifacts identification and removal from EEG data using hybrid machine learning methods | |
Islam et al. | Probability mapping based artifact detection and wavelet denoising based artifact removal from scalp EEG for BCI applications | |
Dzitac et al. | Identification of ERD using fuzzy inference systems for brain-computer interface | |
Tian et al. | Latency estimation of auditory brainstem response by neural networks | |
Ahmed et al. | Effective hybrid method for the detection and rejection of electrooculogram (EOG) and power line noise artefacts from electroencephalogram (EEG) mixtures | |
CN117064405A (en) | Single-channel electroencephalogram signal artifact removal method, equipment and medium | |
Xu et al. | Feature extraction and classification of EEG for imaging left-right hands movement | |
CN115299960A (en) | Electric signal decomposition method and electroencephalogram signal decomposition device based on short-time varying separate modal decomposition | |
Li et al. | sEMG signal filtering study using synchrosqueezing wavelet transform with differential evolution optimized threshold | |
Naik et al. | Performance comparison of ICA algorithms for isometric hand gesture identification using surface EMG | |
Michahial et al. | Hand rotate EEG signal feature extraction by second order Daubechies wavelet transform (DWT) | |
Naik et al. | Towards classification of low-level finger movements using forearm muscle activation: a comparative study based on ICA and Fractal theory | |
Yong et al. | Automatic artefact detection in a self-paced brain-computer interface system | |
Kim et al. | Analysis of brain function and classification of sleep stage EEG using Daubechies wavelet | |
Chandra et al. | Classification of Electroencephalography for Neurobiological Spectrum Disorder Diagnosis |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20110323 Termination date: 20170713 |
|
CF01 | Termination of patent right due to non-payment of annual fee |