CN108652650A - Alertness real-time detection method based on pulse wave signal and system - Google Patents
Alertness real-time detection method based on pulse wave signal and system Download PDFInfo
- Publication number
- CN108652650A CN108652650A CN201810505524.4A CN201810505524A CN108652650A CN 108652650 A CN108652650 A CN 108652650A CN 201810505524 A CN201810505524 A CN 201810505524A CN 108652650 A CN108652650 A CN 108652650A
- Authority
- CN
- China
- Prior art keywords
- pulse wave
- alertness
- wave signal
- module
- real
- 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.)
- Pending
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/16—Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state
- A61B5/18—Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state for vehicle drivers or machine operators
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/16—Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state
- A61B5/168—Evaluating attention deficit, hyperactivity
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
- A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
- A61B5/6802—Sensor mounted on worn items
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
- A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
- A61B5/6813—Specially adapted to be attached to a specific body part
- A61B5/6814—Head
- A61B5/6815—Ear
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
- A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
- A61B5/6813—Specially adapted to be attached to a specific body part
- A61B5/6825—Hand
- A61B5/6826—Finger
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B2503/00—Evaluating a particular growth phase or type of persons or animals
- A61B2503/20—Workers
- A61B2503/22—Motor vehicles operators, e.g. drivers, pilots, captains
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- Surgery (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Psychiatry (AREA)
- Artificial Intelligence (AREA)
- Signal Processing (AREA)
- Physiology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Developmental Disabilities (AREA)
- Hospice & Palliative Care (AREA)
- Educational Technology (AREA)
- Child & Adolescent Psychology (AREA)
- Psychology (AREA)
- Social Psychology (AREA)
- Evolutionary Computation (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Otolaryngology (AREA)
- Cardiology (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
Abstract
The invention discloses a kind of alertness real-time detection method and system based on pulse wave signal, it is related to technical field of life science.Its detecting step is：Pulse wave data is obtained by multi-physiological-parameter acquisition instrument, the pulse wave data is pre-processed, noise reduction simultaneously filters out all kinds of physiology interference, physiological characteristic is extracted from the pulse wave data after the completion of the pretreatment, the physiological characteristic is screened, to reject the redundancy in the pulse wave characteristic, classified to the pulse wave characteristic according to grader, to carry out alertness condition discrimination according to classification results.The present invention can directly detect operator's alertness state, easy to operate, be not easy to be disturbed, and have stronger real-time and higher accuracy, huge potentiality are presented in operator's alertness status monitoring field.
Description
Technical field
The present invention relates to technical field of life science, and in particular to the alertness based on pulse wave signal detects in real time
Method and system.
Background technology
Since the nineties in last century, the research of alertness is increasingly favored by vast scientific research institution and company, currently, alert
The detection means of Juedu mainly has subjective questionnaire detection method, behaviouristics Data Detection method and physio-parameter detection method at this stage.
In correlation detection technology, subjective questionnaire detection means is intended to be affected by subjects subjective, and behaviouristics data are to behavior
A kind of trend prediction, therefore, accuracy and real-time are not very high, and bio-signal acquisition method is by feat of objective, reliable, energy
The physiological change of enough directly reflection subjects has obtained the approval of numerous researchers, is examined for alertness in all physiological signals
In the research of survey.The acquisition of pulse wave signal is more convenient, and data capacity smaller detects the angle of realization from alertness in real time
On from the point of view of, pulse wave is with the obvious advantage, more application prospect.Based on this, in order to overcome prior art alertness detection of complex and easily
It is particularly necessary to design a kind of alertness real-time detection method and system based on pulse wave signal for the defect being disturbed.
Invention content
In view of the shortcomings of the prior art, purpose of the present invention is to be to provide a kind of vigilance based on pulse wave signal
Real-time detection method and system are spent, operator's alertness state can be accurately detected, it is easy to operate, it is not easy to be disturbed, it is practical
Strong, the use easy to spread of property.
To achieve the goals above, the present invention is to realize by the following technical solutions：Police based on pulse wave signal
Juedu real-time detection method, step are：
(1) pulse wave data is obtained by multi-physiological-parameter acquisition instrument；
(2) pulse wave data is pre-processed, noise reduction simultaneously filters out all kinds of physiology interference, improves pulse wave signal
Reliability；
(3) physiological characteristic is extracted from the pulse wave data after the completion of the pretreatment；
(4) physiological characteristic is screened, to reject the redundancy in the pulse wave characteristic；
(5) classified to the pulse wave characteristic according to grader, sentenced with carrying out alertness state according to classification results
Not.
Preferably, multi-physiological-parameter acquisition instrument uses photoelectric sensor in the step (1), photoelectric sensing is used
Device acquires experimenter's pulse wave signal, and is voltage signal by opto-electronic conversion；The photoelectric sensor is wearable in experimenter
Left and right ear-lobe or little finger of toe fingertip location, collecting device is light portable, and data transfer mode is transmitted for wireless blue tooth.
Preferably, described step (2) the pulse wave data pre-treatment step is：To the pulse wave data by 500Hz
It is downsampled to 100Hz；Down-sampled rear pulse wave signal is filtered and wavelet filtering is handled.
Preferably, feature extraction includes temporal signatures and frequency domain character in the step (3), temporal signatures have trough
Phase, wave crest delitescence, secondary peaks between phase, secondary peaks between phase, wave crest between amplitude, wave crest amplitude, secondary peaks amplitude, trough
This 9 features of interval time between incubation period and secondary peaks and wave crest；
The frequency domain character is calculated by WAVELET PACKET DECOMPOSITION mode：To the signal after denoising using db3 wavelet basis into
6 layers of WAVELET PACKET DECOMPOSITION of row, signal are decomposed into 64 sub-bands by corresponding, and each sub-band width is 50Hz/64=0.78Hz；It is fixed
The a certain frequency band energy of justice is the quadratic sum of wavelet coefficient under the frequency range, i.e.,
In formula (1), I is the WAVELET PACKET DECOMPOSITION number of plies, and N is every layer of wavelet coefficient number, Ci(k) it is k-th of i-th layer of small echo
The gross energy of coefficient, entire signal is defined as
The ratio that then frequency band i accounts for gross energy is
Pi=Ei/Etotal (3)
After WAVELET PACKET DECOMPOSITION, the energy probability P of 64 sub-bandsiFrequency domain character as pulse wave.
Preferably, it includes carrying out feature using Relief algorithms that the step (4) carries out screening to pulse wave characteristic
Weight computing simultaneously screens：When evaluating the weight of feature in some sample, one is found out in all generic samples first
Closest sample Y then finds a different classes of nearest samples Z again；If some characteristic point again between X and Y very
It is close, but distance is again especially remote between X and Z, then the characteristic point for meeting such condition, which can be regarded as, is to discriminate between inhomogeneity
Not more important point can assign a higher weight to these points.If sample data is discrete, sample characteristics point
The distance between pass through formula (4) calculate：
If sample data is continuous, the distance between sample characteristics point is carved by the Euclidean distance between vector
It draws, such as formula (5).
diff(Si,Yi)=| Si-Yi| (5)
Preferably, using support vector machine classifier to alertness distinguished number in the step (5).
Alertness real-time detecting system based on pulse wave signal, including acquisition module, preprocessing module, extraction module,
Screening module and discrimination module, acquisition module, preprocessing module, extraction module, screening module, discrimination module are sequentially connected, institute
It states acquisition module and is used to filter out all kinds of physiology to pulse wave signal and interfere for obtaining pulse wave data, preprocessing module, extraction
Module is used to extract each category feature of pulse wave signal, and it is complicated to simplify modeling for being screened to carried feature for screening module
Degree；Discrimination module is for classifying to the brain electrical feature according to grader, to carry out alertness state according to classification results
Identification.
Beneficial effects of the present invention：This method can directly and accurately detect operator's alertness state, have stronger
Real-time and higher accuracy, it is easy to operate, it is not easy to be disturbed, be presented in operator's alertness status monitoring field huge
Potentiality；For this system by wireless blue tooth mode interactive information, equipment wearing mode is lightly convenient simultaneously, is easy to use.
Description of the drawings
The following describes the present invention in detail with reference to the accompanying drawings and specific embodiments；
Fig. 1 is the detection method flow chart of the present invention；
Fig. 2 is the detecting system structure diagram of the present invention；
Fig. 3 is the pulse wave temporal signatures figure of the present invention.
Specific implementation mode
To make the technical means, the creative features, the aims and the efficiencies achieved by the present invention be easy to understand, with reference to
Specific implementation mode, the present invention is further explained.
Referring to Fig.1-3, present embodiment uses following technical scheme：Alertness based on pulse wave signal is examined in real time
Survey method, step are：
(1) it uses photoelectric sensor to acquire experimenter's pulse wave signal, and is voltage signal by opto-electronic conversion；
(2) pulse wave data is pre-processed, noise reduction simultaneously filters out all kinds of physiology interference, improves pulse wave signal
Reliability；Pre-treatment step is：100Hz is downsampled to by 500Hz to the pulse wave data, to down-sampled rear pulse wave signal
It is filtered and wavelet filtering processing；
(3) physiological characteristic is extracted from the pulse wave data after the completion of the pretreatment；Feature extraction includes temporal signatures
And frequency domain character, temporal signatures have between trough amplitude, wave crest amplitude, secondary peaks amplitude, trough phase, secondary wave between phase, wave crest
This 9 features of interval time between peak-to-peak phase, wave crest delitescence, secondary peaks incubation period and secondary peaks and wave crest；
(4) physiological characteristic is screened, to reject the redundancy in the pulse wave characteristic；
(5) classified to the pulse wave characteristic according to grader, sentenced with carrying out alertness state according to classification results
Not, alertness state is labeled by experiment, and then determined between the pulse wave signal and driver's alertness state
Relationship, estimation is made to driver's alertness state.
It is worth noting that, step (1) photoelectric sensor is wearable in experimenter or so ear-lobe or little finger of toe finger tip position
It sets, photoelectric sensor is light portable, and data transfer mode is transmitted for wireless blue tooth.
Step (2) is easy since pulse wave signal is fainter by subject's light exercise and industrial frequency noise etc.
Interference, therefore need first to pre-process pulse wave signal when extracting pulse wave correlated characteristic, filter out various noises.Arteries and veins
Wave number of fighting Data preprocess is as follows：First, source signal sample rate is reduced to 100Hz by original 512Hz, reduced
Follow-up data processing procedure is also convenient for while data volume；Secondly, it is replaced with the average value of 5, each of signal point periphery point
The point carries out signal smooth；The bandpass filtering of 0.5Hz-40Hz is carried out to signal again；Finally, according to the position of goal stimulus
It sets data sectional, is -2s to 8s (being reference with goal stimulus start time) per a bit of data time span.
Feature extraction includes temporal signatures and frequency domain character in step (3), temporal signatures have trough amplitude, wave crest amplitude,
Phase, wave crest delitescence, secondary peaks incubation period and secondary between phase, secondary peaks between phase, wave crest between secondary peaks amplitude, trough
This 9 features of interval time between wave crest and wave crest；With reference to trough, wave crest and the corresponding waveform of secondary peaks in Fig. 3, two phases
Lead time between the adjacent trough phase between trough, between wave crest between phase and secondary peaks the phase define it is similar, from trough to adjacent
Time difference between wave crest is defined as wave crest delitescence；Trough is defined as secondary peaks to the time difference between stimulation wave crest and hides
Phase.Whole features of the interior data of each goal stimulus data segment span -2s to 8s (being reference with goal stimulus start time) are flat
Mean value is as a sample.
The frequency domain character is calculated by WAVELET PACKET DECOMPOSITION mode：The sample frequency of pulse wave signal is in the present invention
100Hz, corresponding pulse wave signal frequency are 0-50Hz.6 layers of wavelet packet point are carried out using db3 wavelet basis to the signal after denoising
Solution, signal are decomposed into 64 sub-bands by corresponding, and each sub-band width is 50Hz/64=0.78Hz；Define a certain frequency band energy
Amount is the quadratic sum of wavelet coefficient under the frequency range, i.e.,
In formula (1), I is the WAVELET PACKET DECOMPOSITION number of plies, and N is every layer of wavelet coefficient number, Ci(k) it is k-th of i-th layer of small echo
The gross energy of coefficient, entire signal is defined as
The ratio that then frequency band i accounts for gross energy is
Pi=Ei/Etotal (3)
After WAVELET PACKET DECOMPOSITION, the energy probability P of 64 sub-bandsiFrequency domain character as pulse wave.
The feature extracted in step (4) includes time domain 9, and frequency domain 64, feature sum has reached 73, by all features
It is introduced into model, one side model time complexity is too high, and computation rate is slow；On the other hand, model is also easily caused
Over-fitting.Therefore, it before establishing alertness detection model, needs to screen feature, reject unrelated with alertness state change
Feature.It includes carrying out feature weight calculating using Relief algorithms and screening to carry out screening to pulse wave characteristic：When evaluate some
In sample when the weight of feature, a closest sample Y is found out in all generic samples first, then finds one again
A different classes of nearest samples Z；So it is contemplated that if some characteristic point is very close between X and Y again, but X and Z
Between distance it is again especially remote, can be with then the characteristic point for meeting such condition, which is regarded as, is to discriminate between different classes of more important point
A higher weight is assigned to these points.Relief algorithms are implemented as follows description：
Input parameter：Training sample space X, number of samples M, each sample characteristics number are N,
Iterative calculation thorn number T (in general, T takes M).
Output parameter：For each characteristic point FiWeight Wi。
1. initializing：Each FiThe initial weight of feature is 0, Wi=0；
2. for i=1to T do
3. selecting an example sample from X at random is denoted as S；
4. finding Ss, k different classes of lower nearest samples Sd of sample closest under generic with S in Xk；
5. for j=1to N do
⑥pkWhat is indicated is kth class number
According to the ratio of number and total data number；
⑦end for
⑧end for
⑨return(W)。
If sample data is discrete, the distance between sample characteristics point is calculated by formula (4)：
When sample data is continuous, then the distance between sample characteristics point is carved by the Euclidean distance between vector
It draws, such as formula (5)：
diff(Si,Yi)=| Si-Yi| (5)
To alertness distinguished number using support vector machines (support vector machine, SVM) point in step (5)
Class device.
Alertness real-time detecting system based on pulse wave signal, including acquisition module 210, preprocessing module 220, extraction
Module 230, screening module 240 and discrimination module 250, acquisition module 210, preprocessing module 220, extraction module 230, screening mould
Block 240, discrimination module 250 are sequentially connected, and for obtaining pulse wave data, preprocessing module 220 is used for the acquisition module 210
All kinds of physiology interference are filtered out to pulse wave signal, extraction module 230 is used to extract each category feature of pulse wave signal, screening module
240, for being screened to carried feature, simplify modeling complexity；Discrimination module 250 is used for according to grader to brain electricity
Feature is classified, to carry out alertness state recognition according to classification results.
Present embodiment is established before the class models of alertness, is needed first to collected physiological parameter into pedestrian
For calibration, the priori conditions as sorting algorithm.The foundation of data mark of the present invention has two parts：The subjective scale of subject is commented
Divide and behaviouristics data (goal response time and goal response accuracy).In data annotation process, subject is required
Experiment beginning the previous day will normally rest, and experimental day cannot carry out intense physical or mental labour.When entire Therapy lasted
Between for 95min or so, test once later, must not exit halfway.Without clock and watch in experimental situation, ensures that subject is not known and work as
The preceding time.It tests the 5min most started to be used for acquiring subject's quiescent condition signal, requires subject to sit quietly in this stage, as possible
Reduce mental activity.And then continuous 3 periods subject is carried out identical Mackworth clocks experiment (MCT).MCT is
One experiment dedicated for research alertness variation, in this experiment, a pointer is around a dial plate without reference point with solid
Determine angle and frequency jitter, at the indefinite moment, pointer can jump larger angle, which is referred to as object event.In experiment
It is required that subject identifies object event, and given reaction is made to it.After data mark is completed, it is suitable also to select
Grader establish the detection model of alertness.Using the SVM suitable for Small Sample Database as learning method in the present invention.
Present embodiment can directly and accurately detect operator's alertness state, have stronger real-time, operation
It is convenient, it is not easy to be disturbed, presents huge potentiality in operator's alertness status monitoring field, have a vast market application
Foreground.
The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention.The technology of the industry
Personnel are it should be appreciated that the present invention is not limited to the above embodiments, and the above embodiments and description only describe this
The principle of invention, without departing from the spirit and scope of the present invention, various changes and improvements may be made to the invention, these changes
Change and improvement all fall within the protetion scope of the claimed invention.The claimed scope of the invention by appended claims and its
Equivalent thereof.
Claims (7)
1. the alertness real-time detection method based on pulse wave signal, which is characterized in that its step is：
(1) pulse wave data is obtained by multi-physiological-parameter acquisition instrument；
(2) pulse wave data is pre-processed, noise reduction simultaneously filters out all kinds of physiology interference, improves the reliable of pulse wave signal
Property；
(3) physiological characteristic is extracted from the pulse wave data after the completion of the pretreatment；
(4) physiological characteristic is screened, to reject the redundancy in the pulse wave characteristic；
(5) classified to the pulse wave characteristic according to grader, to carry out alertness condition discrimination according to classification results.
2. the alertness real-time detection method according to claim 1 based on pulse wave signal, which is characterized in that described
Multi-physiological-parameter acquisition instrument uses photoelectric sensor in step (1), and experimenter's pulse wave signal is acquired using photoelectric sensor,
And by opto-electronic conversion be voltage signal；The photoelectric sensor is wearable in experimenter or so ear-lobe or little finger of toe fingertip location.
3. the alertness real-time detection method according to claim 1 based on pulse wave signal, which is characterized in that described
Step (2) pulse wave data pre-treatment step is：100Hz is downsampled to by 500Hz to the pulse wave data；After down-sampled
Pulse wave signal is filtered and wavelet filtering processing.
4. the alertness real-time detection method according to claim 1 based on pulse wave signal, which is characterized in that described
Feature extraction includes temporal signatures and frequency domain character in step (3), and temporal signatures have trough amplitude, wave crest amplitude, secondary peaks
Phase, wave crest delitescence, secondary peaks incubation period and secondary peaks and wave between phase, secondary peaks between phase, wave crest between amplitude, trough
This 9 features of interval time between peak；
The frequency domain character is calculated by WAVELET PACKET DECOMPOSITION mode：6 layers are carried out using db3 wavelet basis to the signal after denoising
WAVELET PACKET DECOMPOSITION, signal are decomposed into 64 sub-bands by corresponding, and each sub-band width is 50Hz/64=0.78Hz；Define certain
One frequency band energy is the quadratic sum of wavelet coefficient under the frequency range, i.e.,
In formula (1), I is the WAVELET PACKET DECOMPOSITION number of plies, and N is every layer of wavelet coefficient number, Ci(k) it is k-th of coefficient of i-th layer of small echo,
The gross energy of entire signal is defined as
The ratio that then frequency band i accounts for gross energy is
Pi=Ei/Etotal (3)
After WAVELET PACKET DECOMPOSITION, the energy probability P of 64 sub-bandsiFrequency domain character as pulse wave.
5. the alertness real-time detection method according to claim 1 based on pulse wave signal, which is characterized in that described
It includes carrying out feature weight calculating using Relief algorithms and screening that step (4) carries out screening to pulse wave characteristic：When evaluate certain
In a sample when the weight of feature, a closest sample Y is found out in all generic samples first, is then found again
One different classes of nearest samples Z；When some characteristic point again between X and Y it is very close, but distance is again special between X and Z
Far, it then the characteristic point for meeting such condition, which is regarded as, is to discriminate between different classes of more important point, is assigned to these points one higher
Weight；When sample data is discrete, then the distance between sample characteristics point is calculated by formula (4)：
When sample data is continuous, then the distance between sample characteristics point is portrayed by the Euclidean distance between vector, such as
Formula (5).
diff(Si,Yi)=| Si-Yi| (5)。
6. the alertness real-time detection method according to claim 1 based on pulse wave signal, which is characterized in that described
Support vector machine classifier is used to alertness distinguished number in step (5).
7. the alertness real-time detecting system based on pulse wave signal, which is characterized in that including acquisition module (210), pretreatment
Module (220), extraction module (230), screening module (240) and discrimination module (250), acquisition module (210), preprocessing module
(220), extraction module (230), screening module (240), discrimination module (250) are sequentially connected, and the acquisition module (210) is used for
Pulse wave data is obtained, preprocessing module (220) is used to filter out all kinds of physiology to pulse wave signal and interfere, extraction module (230)
It is complicated to simplify modeling for being screened to carried feature for each category feature for extracting pulse wave signal, screening module (240)
Degree；Discrimination module (250) is for classifying to the brain electrical feature according to grader, to carry out alertness according to classification results
State recognition.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810505524.4A CN108652650A (en) | 2018-05-24 | 2018-05-24 | Alertness real-time detection method based on pulse wave signal and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810505524.4A CN108652650A (en) | 2018-05-24 | 2018-05-24 | Alertness real-time detection method based on pulse wave signal and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108652650A true CN108652650A (en) | 2018-10-16 |
Family
ID=63776477
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810505524.4A Pending CN108652650A (en) | 2018-05-24 | 2018-05-24 | Alertness real-time detection method based on pulse wave signal and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108652650A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110811573A (en) * | 2019-10-29 | 2020-02-21 | 依脉人工智能医疗科技(天津)有限公司 | Device and method for regulating and controlling brain alertness based on photoelectric pulse feedback |
| CN113100727A (en) * | 2021-05-12 | 2021-07-13 | 深圳市通久电子有限公司 | Method for analyzing and identifying pulse wave crest in real time |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105960196A (en) * | 2014-02-24 | 2016-09-21 | 索尼公司 | Smart wearable devices and methods for acquisition of sensorial information from wearable devices to activate functions in other devices |
| CN107072541A (en) * | 2014-09-09 | 2017-08-18 | 托维克公司 | For utilizing wearable device monitoring individual alertness and the method and apparatus that provides notice |
-
2018
- 2018-05-24 CN CN201810505524.4A patent/CN108652650A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105960196A (en) * | 2014-02-24 | 2016-09-21 | 索尼公司 | Smart wearable devices and methods for acquisition of sensorial information from wearable devices to activate functions in other devices |
| CN107072541A (en) * | 2014-09-09 | 2017-08-18 | 托维克公司 | For utilizing wearable device monitoring individual alertness and the method and apparatus that provides notice |
Non-Patent Citations (2)
| Title |
|---|
| 周谭琪等: "基于Relief算法的心血管疾病辅助诊断研究", 《生物医学工程学杂志》 * |
| 曹勇等: "基于脉搏波的警觉度检测研究", 《生物医学工程学杂志》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110811573A (en) * | 2019-10-29 | 2020-02-21 | 依脉人工智能医疗科技(天津)有限公司 | Device and method for regulating and controlling brain alertness based on photoelectric pulse feedback |
| CN113100727A (en) * | 2021-05-12 | 2021-07-13 | 深圳市通久电子有限公司 | Method for analyzing and identifying pulse wave crest in real time |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Singh et al. | Classification of unsegmented heart sound recording using KNN classifier | |
| CN108304917A (en) | A kind of P300 signal detecting methods based on LSTM networks | |
| Tripathy et al. | Automated detection of congestive heart failure from electrocardiogram signal using Stockwell transform and hybrid classification scheme | |
| Hsieh et al. | Feature selection framework for XGBoost based on electrodermal activity in stress detection | |
| CN110432898A (en) | A kind of epileptic attack eeg signal classification system based on Nonlinear Dynamical Characteristics | |
| CN107095669B (en) | A kind of processing method and system of epileptic's EEG signals | |
| Soh et al. | A computational intelligence tool for the detection of hypertension using empirical mode decomposition | |
| CN109934089A (en) | Multistage epileptic EEG Signal automatic identifying method based on supervision gradient lifter | |
| CN103034837B (en) | Characteristic parameter is associated with pulse condition key element | |
| WO2010115939A2 (en) | A method for the real-time identification of seizures in an electroencephalogram (eeg) signal | |
| CN111248859A (en) | Automatic sleep apnea detection method based on convolutional neural network | |
| Nascimento et al. | Heart arrhythmia classification based on statistical moments and structural co-occurrence | |
| Hamza et al. | Svm for human identification using the ECG signal | |
| CN108256453A (en) | A kind of method based on one-dimensional ECG signal extraction two dimension CNN features | |
| CN112674782B (en) | Device and method for detecting epileptic-like electrical activity of epileptic during inter-seizure period | |
| Al-Assaf | Surface myoelectric signal analysis: dynamic approaches for change detection and classification | |
| CN116400800B (en) | ALS patient human-computer interaction system and method based on brain-computer interface and artificial intelligence algorithm | |
| CN108652650A (en) | Alertness real-time detection method based on pulse wave signal and system | |
| Lee et al. | Real-time seizure detection using EEG: a comprehensive comparison of recent approaches under a realistic setting | |
| Zhang et al. | Hybrid feature fusion for classification optimization of short ECG segment in IoT based intelligent healthcare system | |
| CN113768515A (en) | Electrocardiosignal classification method based on deep convolutional neural network | |
| Liu et al. | Motion artifact detection in ppg signals based on gramian angular field and 2-d-cnn | |
| Mateo et al. | Analysis of atrial and ventricular premature contractions using the short time fourier transform with the window size fixed in the frequency domain | |
| CN115089139A (en) | Personalized physiological parameter measuring method combining biological characteristic identification | |
| Jiang et al. | Attention-based multi-scale features fusion for unobtrusive atrial fibrillation detection using ballistocardiogram signal |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181016 |
|
| RJ01 | Rejection of invention patent application after publication |