CN114983412A - Non-contact type brain cognitive load objective detection method based on distributed radar - Google Patents
Non-contact type brain cognitive load objective detection method based on distributed radar Download PDFInfo
- Publication number
- CN114983412A CN114983412A CN202210512666.XA CN202210512666A CN114983412A CN 114983412 A CN114983412 A CN 114983412A CN 202210512666 A CN202210512666 A CN 202210512666A CN 114983412 A CN114983412 A CN 114983412A
- Authority
- CN
- China
- Prior art keywords
- signal
- channel
- radar
- time domain
- brain
- 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.)
- Granted
Links
- 230000001149 cognitive effect Effects 0.000 title claims abstract description 47
- 210000004556 brain Anatomy 0.000 title claims abstract description 42
- 238000001514 detection method Methods 0.000 title claims abstract description 20
- 230000000241 respiratory effect Effects 0.000 claims abstract description 53
- 238000012706 support-vector machine Methods 0.000 claims abstract description 29
- 238000000034 method Methods 0.000 claims abstract description 25
- 230000036992 cognitive tasks Effects 0.000 claims abstract description 22
- 238000001914 filtration Methods 0.000 claims abstract description 22
- 210000000115 thoracic cavity Anatomy 0.000 claims abstract description 6
- 230000029058 respiratory gaseous exchange Effects 0.000 claims description 30
- 239000000523 sample Substances 0.000 claims description 27
- 238000004364 calculation method Methods 0.000 claims description 13
- 230000008569 process Effects 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 11
- 238000010606 normalization Methods 0.000 claims description 9
- 238000001228 spectrum Methods 0.000 claims description 8
- 238000012549 training Methods 0.000 claims description 7
- ADEORFBTPGKHRP-UHFFFAOYSA-N 1-[7-(dimethylamino)-4-methyl-2-oxochromen-3-yl]pyrrole-2,5-dione Chemical compound O=C1OC2=CC(N(C)C)=CC=C2C(C)=C1N1C(=O)C=CC1=O ADEORFBTPGKHRP-UHFFFAOYSA-N 0.000 claims description 5
- 210000000038 chest Anatomy 0.000 claims description 5
- 238000005516 engineering process Methods 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 239000013598 vector Substances 0.000 claims description 4
- 230000000737 periodic effect Effects 0.000 claims description 3
- 230000000276 sedentary effect Effects 0.000 claims description 3
- 230000001360 synchronised effect Effects 0.000 claims description 3
- 238000009423 ventilation Methods 0.000 claims description 3
- 230000003340 mental effect Effects 0.000 description 18
- 230000006870 function Effects 0.000 description 12
- 238000005259 measurement Methods 0.000 description 8
- 238000012360 testing method Methods 0.000 description 5
- 238000013145 classification model Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000000691 measurement method Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000033001 locomotion Effects 0.000 description 3
- 239000000284 extract Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000036544 posture Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 206010063659 Aversion Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000008451 emotion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000006461 physiological response Effects 0.000 description 1
- 230000003938 response to stress Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
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
- A61B5/165—Evaluating the state of mind, e.g. depression, anxiety
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/05—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves
- A61B5/0507—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves using microwaves or terahertz waves
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/08—Detecting, measuring or recording devices for evaluating the respiratory organs
-
- 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
- A61B5/72—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
- A61B5/7225—Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation
-
- 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
- A61B5/725—Details of waveform analysis using specific filters therefor, e.g. Kalman or adaptive filters
-
- 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
- A61B5/7253—Details of waveform analysis characterised by using transforms
- A61B5/7257—Details of waveform analysis characterised by using transforms using Fourier transforms
-
- 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
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Surgery (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- General Health & Medical Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Psychiatry (AREA)
- Signal Processing (AREA)
- Artificial Intelligence (AREA)
- Physiology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Developmental Disabilities (AREA)
- Hospice & Palliative Care (AREA)
- Social Psychology (AREA)
- Child & Adolescent Psychology (AREA)
- Educational Technology (AREA)
- Mathematical Physics (AREA)
- Psychology (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Power Engineering (AREA)
- Radiology & Medical Imaging (AREA)
- Pulmonology (AREA)
- Evolutionary Computation (AREA)
- Fuzzy Systems (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention discloses a non-contact type brain cognitive load objective detection method based on a distributed radar. The method comprises the following steps: aiming a distributed radar at the thoracic cavity of a human body, and respectively collecting multichannel radar echo signals during sitting and executing cognitive tasks; acquiring a channel with the highest signal-to-noise ratio of a respiratory signal from a multi-channel radar echo signal, and then filtering the wave to acquire the respiratory signal; extracting time domain characteristics and frequency domain characteristics from the obtained respiratory signals, and removing individual differences; and selecting the features based on the time domain features and the frequency domain features after the individual differences are removed, and establishing a support vector machine model to detect the cognitive load level of the brain. The invention provides a non-contact, strong-robustness and objective detection method for brain cognitive load detection.
Description
Technical Field
The invention relates to the technical field of signal processing, in particular to a non-contact type brain cognitive load objective detection method based on a distributed radar.
Background
With the advance of the society to informatization, the amount of information and the complexity of information to be processed by people in work become larger and larger, and mental labor gradually becomes the main labor form of most workers. The improvement of the mental demand of the work inevitably leads to the increase of the mental load, and the higher mental load can cause rapid fatigue, reduced flexibility, stress response and frustrated emotion, and cause errors in information acquisition and analysis and decision errors; too low mental load may result in wasting resources such as manpower, causing aversion and decreasing performance. Therefore, measuring the mental load of a person is of increasing practical importance.
At present, there are three main methods for measuring mental load, namely task performance measurement, subjective assessment measurement and physiological measurement. Assessment of task performance measurement is too dependent on task types, and when tasks are changed, assessment needs to be re-established, so that universality is not achieved. Subjective assessment measures, although simple and effective, can only be reviewed by testers, and besides lack of real-time performance, the obtained conclusions also lack of objectivity. Both of these measurement methods are qualitative in nature.
The physiological measurement method is an indirect objective measurement mode, physiological signals for evaluating cognitive load comprise electrocardio, skin electrocardio, electroencephalogram, pulse, respiration and other physiological signals, but the acquisition mode of the method requires a tester to wear a plurality of complex devices, and the contact type sensor is inconvenient and uncomfortable and is not beneficial to monitoring the cognitive load for a long time. The radar collects human physiological signals in a non-contact mode, and the problems of contact type measuring equipment can be effectively solved.
In the radar non-contact type physiological signal measurement, a single radar can only detect a certain area due to a single angle. When the posture of a tester is changed, the signal-to-noise ratio of a radar echo signal is reduced, the waveform quality of a physiological signal is greatly influenced, and a test result is misjudged. Therefore, a device with strong robustness is needed to improve the non-contact measurement effect by acquiring a signal with a high signal-to-noise ratio.
Disclosure of Invention
The invention aims to solve the defects in the prior art and provides a non-contact type brain cognitive load objective detection method based on a distributed radar.
The purpose of the invention can be achieved by adopting the following technical scheme:
a non-contact type brain cognitive load objective detection method based on a distributed radar comprises the following steps:
s1, aiming the distributed radar at the thoracic cavity of the human body, and respectively collecting multichannel radar echo signals during sitting and executing cognitive tasks;
s2, obtaining a channel with the highest signal-to-noise ratio of a respiratory signal from the multi-channel radar echo signal, and then filtering the wave to obtain the respiratory signal;
s3, extracting time domain features and frequency domain features from the obtained respiratory signals, and removing individual differences;
and S4, selecting features based on the time domain features and the frequency domain features after the differences are removed, and establishing a support vector machine model to detect the cognitive load level of the brain.
Further, in step S1, the process of directing the distributed radar to the chest of the human body and respectively acquiring the multichannel radar echo signals during the sitting and the cognitive tasks is as follows:
s101, a distributed radar comprising P radar probes is used, the P radar probes are respectively subjected to frequency division processing, are distributed into sectors around a tester, and respectively acquire radar echo signals of the tester in a sitting state, a low-load cognitive task executing state and a high-load cognitive task executing state, wherein the sitting state is used as a baseline state, a label for executing the low-load cognitive task is set to be-1, and a label for executing the high-load cognitive task is set to be +1, so that the radar echo signals can be used in subsequent classification. The step uses a fan-shaped distributed radar, so that when the orientation of a human body changes, the front of a radar node can be ensured to be aligned with the thoracic cavity of the human body, and a respiratory signal with a high signal-to-noise ratio can be acquired;
and S102, performing multi-channel synchronous acquisition by using a multi-bit high-precision ADC (analog to digital converter) of the MCU processor to obtain a multi-channel radar echo signal. Converting a multi-channel radar echo analog signal into a multi-channel radar echo digital signal by using a multi-bit high-precision ADC (analog to digital converter) and then collecting the multi-channel radar echo digital signal so as to facilitate computer processing;
s103, wirelessly transmitting the multi-channel radar echo data to a cloud end by using a WiFi technology. In order to solve the problem of inconvenient movement caused by wired transmission, the wireless transmission is carried out by using a WiFi technology.
Further, in step S2, a channel with the highest signal-to-noise ratio of the respiratory signal is obtained from the multi-channel radar echo signal, and then the process of obtaining the respiratory signal by filtering the wave is as follows:
s201, low-pass filtering is carried out on the multi-channel radar echo signals respectively. In the step, the influence of high-frequency noise such as power frequency noise, internal noise of a radar probe and the like on the radar echo signal is considered, so that I/Q signal calibration is influenced, and the high-frequency noise is removed by using low-pass filtering;
s202, I/Q signal calibration is carried out on the radar echo signals of each channel after low-pass filtering, and then arc tangent demodulation is carried out by using an extended DACM algorithm to obtain demodulation signals of each channel. The method comprises the following steps of obtaining an I/Q signal in a radar echo signal, and solving the problems of amplitude imbalance, phase imbalance, direct current offset and phase ambiguity of the I/Q signal in the radar echo signal. The extended DACM algorithm solves the phase ambiguity problem by keeping the demodulated signal as a continuous single-valued function, leaving its amplitude range unrestricted.
S203, removing baseline drift of the obtained demodulation signals of each channel, then calculating autocorrelation coefficients, and rejecting the channels with the autocorrelation coefficients lower than a threshold value according to the threshold value. The method comprises the steps of firstly removing baseline drift influence caused by human body micromotion and hardware circuits, and then primarily removing channels with low periodicity by utilizing autocorrelation coefficients based on the characteristic that radar echo signals are quasi-periodic;
s204, based on the maximum signal-to-noise ratio principle, selecting a channel signal with the highest signal-to-noise ratio of the respiratory signal from the rest channels as a final output signal, wherein the signal-to-noise ratio calculation formula is as follows:
SNR＝10log(|X is | 2 /E(|X in | 2 ) Equation (1)
Wherein, X is And X in Respectively representing the amplitudes of the respiration signal and the noise of the ith channel in the frequency domain, and the SNR is the signal-to-noise ratio of the respiration signal in the channel. In the step, the signal-to-noise ratio represents the ratio of the power of the respiratory signal to the power of the noise, and the higher the signal-to-noise ratio is, the smaller the noise is, so that the signal-to-noise ratio of the respiratory signal is calculated from a frequency domain, and a channel signal with the highest signal-to-noise ratio of the respiratory signal is selected as a final output signal, so that the influence of the noise is reduced to the maximum extent;
and S205, performing band-pass filtering on the final output signal to obtain a respiratory signal. In the step, the final output signal is considered to contain two physiological signals of heartbeat and respiration, so that the heartbeat is removed by using band-pass filtering, and a pure respiration signal is obtained.
Further, in step S3, the time domain feature and the frequency domain feature are extracted from the obtained respiratory signal, and the process of removing the individual difference is as follows:
s301, searching a peak point and a valley point of the respiratory signal, and extracting a respiratory signal time domain characteristic sequence according to the peak point and the valley point; wherein the time domain characteristic sequence comprises 8 characteristic sequences of inspiration time, expiration amplitude, respiration interval, periodic ventilation volume, average respiration rate, average tidal volume and minute inspiration volume which are sequentially and respectively marked as X i I ∈ {1,2,3,4,5,6,7,8}, where the average respiration rate, average tidal volume, and minute inspiration are calculated, respectively, over a sliding window of window length M and sliding step size N. The purpose of this step is to fully extract the abundant breathing patterns contained in the breathing signal for subsequent classification.
S302, respectively calculating the mean value, the median value, the standard deviation, the root mean square, the maximum value, the minimum value, the variation coefficient and the quartile range of the characteristic sequence as time domain characteristics. The purpose of this step is to fully mine the statistical features of the feature sequence.
Wherein the mean value u of the ith feature sequence i Standard deviation σ i And root mean square rms i The calculation formulas of (A) are respectively as follows:
wherein, X i [n]Is the N-th element of the i-th signature sequence, N i Represents the sequence length of the ith characteristic sequence;
coefficient of variation CV of ith signature sequence i The calculation formula is as follows:
s303, carrying out fast Fourier transform on the obtained respiratory signal to obtain a frequency spectrum X [ k ], extracting a frequency mean value and a frequency standard deviation as frequency domain characteristics, wherein the calculation formulas are respectively as follows:
wherein f _ mean is a frequency mean value, f _ std is a frequency standard deviation, X [ K ] is a fast Fourier transform spectrum, and K is a signal length. In the step, the frequency spectrum reflects the distribution form of the signal in the frequency domain, and various frequency components contained in the signal are displayed, so that the respiratory signal frequency spectrum is obtained through Fourier transform, and further, the frequency mean value and the frequency standard deviation are obtained and are used as the frequency domain characteristics;
s304, respectively subtracting the time domain characteristic and the frequency domain characteristic of the respiration signal in the sedentary state from the time domain characteristic and the frequency domain characteristic of the respiration signal in the cognitive task, so as to remove the individual difference and obtain the time domain characteristic and the frequency domain characteristic after the individual difference is removed. This step is to avoid the problem that the breathing pattern of each person is different, so that the breathing pattern of some individuals in a sedentary state is the same as the breathing pattern of other individuals in a cognitive task, and the classification accuracy is finally reduced.
Further, in the step S4, based on the time domain features and the frequency domain features after the individual differences are removed, feature selection is performed, and a process of establishing a support vector machine model to detect the cognitive load level of the brain is as follows:
s401, respectively carrying out normalization processing on the time domain characteristics and the frequency domain characteristics after the individual differences are removed, and unifying the numerical values of the characteristics to 0-1 to obtain the normalized time domain characteristics and frequency domain characteristics. In the step, the normalization processing can keep the dimension of the sample consistent, and the influence of the extreme point on the precision of the classification model is avoided, so that the calculation precision and the convergence speed of the classification model are improved;
s402, performing mutual information screening on the time domain characteristics and the frequency domain characteristics obtained after normalization in the step S401, and selecting the first m characteristics as final brain cognitive load characteristics. In the step, considering that dimension disaster is easily caused when the feature dimension is too high, so that the classification accuracy of the classification model is reduced, the feature subset with the best performance can be screened from the original feature set by carrying out mutual information screening, and the classification accuracy and the model learning speed are improved;
s403, sending the first m features in the step S402 into a support vector machine model for training, and setting a training sample set as { x } m ,y m }，y m E { -1, +1}, m { -1,. the L, L is sample oneIn order to solve the problem of nonlinear classification, a nonlinear kernel support vector machine is adopted, and a decision function f (x) and an RBF kernel function formula of the nonlinear kernel support vector machine are as follows:
wherein,
is a Lagrange multiplier, b * As classification threshold, x m Is the m-th input variable, x n Is the nth input variable, y m Is the m-th output value, k (x) m ,x n ) Is a kernel function, and sigma is an RBF kernel penalty coefficient. In the step, the data in an actual scene are mostly linear and inseparable, and the original data can be mapped into a higher-dimensional space by using a nonlinear kernel support vector machine and then are subjected to linear classification again;
s404, after feature normalization processing is carried out on a brain cognitive load feature vector sample of an unknown type, the decision function is calculated, brain cognitive load level detection is carried out according to the values of f (x), and if f (x) -1 is a label-1, the brain cognitive load level is represented; if f (x) >1 is the label +1, it represents a high cognitive load level in the brain. In the step, the trained model is used for detecting the unknown sample, and a final detection result is obtained according to the value of the decision function.
Compared with the prior art, the invention has the following advantages and effects:
(1) the invention detects the cognitive load through the respiratory signal, and the method belongs to a physiological measurement method. Because the physiological response is difficult to control and adjust by subjective consciousness, the method has the advantages of objectivity, accuracy and strong real-time performance, and solves the problems of large subjectivity and large error of a subjective measurement method.
(2) The invention uses the radar to transmit the electromagnetic wave to the human thorax, thereby collecting the respiratory signal of the tester, the signal acquisition process does not need to contact with the tester, the non-inductive, non-invasive and non-contact measurement is realized, and the problem that the contact type collection method brings discomfort to the tester is solved.
(3) According to the invention, a plurality of radar probes are distributed into a sector to form a distributed radar, so that a human body breathing signal can be acquired from multiple directions and angles, and when the orientation of a human body changes, a radar node can be ensured to be aligned to the thoracic cavity of the human body in the front, so that a breathing signal with a high signal-to-noise ratio can be acquired; the method selects the channel signal with the highest signal-to-noise ratio of the respiratory signal from the multi-channel signals by using an autocorrelation coefficient method and based on the principle of the highest signal-to-noise ratio, and ensures the accuracy of the time domain characteristic and the frequency domain characteristic of the respiratory signal.
(4) The invention extracts a plurality of characteristics of the respiratory signals, removes individual differences and improves the classification accuracy. Meanwhile, the extracted features are screened, so that dimension disasters are avoided. And moreover, a nonlinear kernel support vector machine classification model is used for establishing a mapping relation between the provided features and the cognitive load level, so that the objective detection of the cognitive load level is realized.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention and do not constitute a limitation of the invention. In the drawings:
FIG. 1 is a flow chart of a non-contact type brain cognitive load objective detection method based on distributed radar disclosed in the embodiment of the invention;
FIG. 2 is a schematic structural diagram of a distributed radar-based physiological signal acquisition in an embodiment of the present invention;
FIG. 3 is a flow chart of acquiring a respiratory signal from a multi-channel radar echo signal in an embodiment of the present invention;
FIG. 4 is a flowchart illustrating the detection of cognitive load in the brain using a support vector machine model according to an embodiment of the present invention;
FIG. 5 is a waveform diagram of demodulation of each channel in an embodiment of the present invention;
FIG. 6 is a waveform diagram of a respiratory signal acquired in an embodiment of the present invention;
FIG. 7 is a ROC plot of a support vector machine model showing a characteristic curve of a subject operation in an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
Fig. 1 is a flowchart of a non-contact type brain cognitive load objective detection method based on a distributed radar according to an embodiment of the present invention.
S1, aiming the distributed radar at the thoracic cavity of the human body, and respectively collecting multichannel radar echo signals during sitting and executing cognitive tasks;
fig. 2 is a schematic structural diagram of a distributed radar for acquiring physiological signals.
S101, in order to cope with different postures possibly appearing in an actual scene, a distributed radar comprising P radar probes is used, the P radar probes are respectively subjected to frequency division processing and are distributed into a sector around a tester. In this example P is chosen to be 6, i.e. 6 radar probes are used. Respectively collecting radar echo signals of a tester in a sitting state, a low-load cognitive task execution state and a high-load cognitive task execution state, wherein the sitting state is used as a baseline state, a label for executing the low-load cognitive task execution state is set to be-1, and a label for executing the high-load cognitive task execution state is set to be 1, so that the radar echo signals are used for subsequent classification;
and S102, performing multi-channel synchronous acquisition by using a multi-bit high-precision ADC of the MCU processor to obtain multi-channel radar echo signals. In the present embodiment, an STM32F429 micro control processor is used as the MCU;
and S103, wirelessly transmitting the multi-channel radar echo data to a cloud terminal by using a WiFi technology. In the present embodiment, the ESP32 WiFi chip is used to implement WiFi wireless transmission.
S2, obtaining a channel with the highest signal-to-noise ratio of a respiratory signal from the multi-channel radar echo signal, and then filtering the wave to obtain the respiratory signal;
fig. 3 is a flowchart of acquiring a channel with the highest signal-to-noise ratio of a respiratory signal from a multi-channel radar echo signal, and then filtering the wave to acquire the respiratory signal.
S201, because the existence of the high-frequency noise may cause unsuccessful demodulation, low-pass filtering is respectively carried out on the multi-channel radar echo signals, and the high-frequency noise is removed. In the embodiment, a Butterworth low-pass filter with the cutoff frequency of 10 Hz and the order of 4 is adopted to remove high-frequency noise;
s202, Doppler frequency shift generated by human chest motion belongs to phase modulation, and direct current offset, amplitude and phase imbalance are generated due to static background clutter in the testing process, micro-motion interference of a tester and errors on hardware. And respectively carrying out I/Q signal calibration on radar echo signals of each channel after low-pass filtering in order to obtain a displacement curve of actual chest movement, and then carrying out arc tangent demodulation by using an extended DACM algorithm to obtain demodulation signals of each channel.
S203, removing the baseline drift of the obtained demodulation signals of each channel, wherein the demodulation signals with good quality generally show good periodicity, calculating the autocorrelation coefficient of the demodulation signals of each channel, and rejecting the channels with the autocorrelation coefficients lower than the threshold value according to the threshold value. In the present embodiment, the threshold is selected to be 0.8.
S204, based on the maximum signal-to-noise ratio principle, selecting a channel signal with the highest signal-to-noise ratio of the respiratory signal from the residual channels obtained in the step S203 as a final output signal, wherein the signal-to-noise ratio calculation formula is as follows:
SNR＝10log(|X s | 2 /E(|X n | 2 ) Equation (1)
In the formula, X S And X n Respectively representing the amplitudes of the respiratory signal and the noise in the frequency domain, and the SNR is the signal-to-noise ratio of the respiratory signal in the channel.
And S205, performing band-pass filtering on the final output signal to obtain a respiratory signal. Since the frequency of the respiration signal is generally between 0.1 and 0.5 Hz, the filtering process is performed by using a Butterworth band-pass filter with a cut-off frequency of 0.1 to 0.5 Hz and an order of 4.
S3, extracting time domain features and frequency domain features from the obtained respiratory signal, and removing the individual difference as follows:
s301, searching a peak point and a valley point of the respiratory signal, and extracting a time domain characteristic sequence of the respiratory signal according to the peak point and the valley point.
Wherein the time domain feature sequence comprises 8 feature sequences of inspiration time, expiration amplitude, breathing interval, periodic ventilation volume, average respiration rate, average tidal volume and minute inspiration volume which are sequentially and respectively marked as X i I belongs to {1,2,3,4,5,6,7,8}, wherein the average respiration rate, the average tidal volume and the minute inspiration volume are respectively calculated by using a sliding window with the window length of M and the sliding step length of N;
s302, respectively calculating the mean value, the median value, the standard deviation, the root mean square, the maximum value, the minimum value, the variation coefficient and the four-quadrant spacing of the obtained characteristic sequences as time domain characteristics;
wherein the mean value u of the ith feature sequence i Standard deviation σ i And root mean square rms i The calculation formulas of (A) are respectively as follows:
wherein X i [n]Is the N-th element of the i-th signature sequence, N i Indicates the sequence length of the ith signature sequence.
Coefficient of variation CV of ith signature sequence i The calculation formula is as follows:
s303, carrying out fast Fourier transform on the obtained respiratory signal to obtain a frequency spectrum X [ k ], extracting a frequency mean value and a frequency standard deviation as frequency domain characteristics, wherein the formulas are respectively as follows:
wherein f _ mean is a frequency mean, f _ std is a frequency standard deviation, X [ K ] is a fast Fourier transform spectrum, and K is a signal length.
S304, because the breathing modes of each person are different, time domain and frequency domain features in the sitting state are respectively subtracted from time domain and frequency domain features of the breathing signals under the cognitive task, and therefore the time domain and frequency domain features with the individual differences removed are obtained.
S4, selecting features based on the time domain features and the frequency domain features after the individual differences are removed, and establishing a support vector machine model to detect the cognitive load level of the brain, wherein the process comprises the following steps:
fig. 4 is a flowchart of detecting cognitive load of the brain using a support vector machine model.
S401, because the dimension of each acquired feature is not uniform, the accuracy of each acquired feature is reduced when the dimension is sent to a trainer for training, the time domain feature and the frequency domain feature which are subjected to the removal of the individual difference and are obtained in the step S304 are normalized, the numerical value of the feature is unified to 0-1, and the normalized time domain feature and the normalized frequency domain feature are obtained;
s402, because the problem of information redundancy exists in the provided characteristics, dimension disasters are easily caused, and the generalization capability of the classifier is reduced. And (4) performing mutual information screening on the time domain characteristics and the frequency domain characteristics obtained after normalization in the step (S401), and selecting the first m characteristics as final brain cognitive load characteristics. In the embodiment, the first 10 characteristics are selected as the final brain cognitive load characteristics;
and S403, sending the top m characteristics in the step S402 into a support vector machine model for training. Let training sample set be { x m ,y m }，y m E { -1, +1}, m { -1,.. and L, L is the number of samples, in order to solve the nonlinear classification problem, a nonlinear kernel support vector machine is adopted, and a decision function f (x) and an RBF kernel function formula of the nonlinear kernel support vector machine are as follows:
wherein,
is a Lagrange multiplier, b * To classify threshold, x m Is the m-th input variable, x n Is the nth input variable, y m Is the m-th output value, k (x) m ,x n ) Is a kernel function, and sigma is an RBF kernel penalty coefficient. In the present embodiment, a gaussian kernel function is adopted as the nonlinear kernel function.
S404, after feature normalization processing is carried out on a brain cognitive load feature vector sample of an unknown type, the decision function is calculated, and brain cognitive load level detection is carried out according to values f (x). If f (x) is less than or equal to-1, the label is-1 and represents the low cognitive load level of the brain, and if f (x) is more than or equal to 1, the label is +1 and represents the high cognitive load level of the brain.
Example 2
Based on the non-contact type brain cognitive load objective detection method based on the distributed radar disclosed in embodiment 1, the embodiment continues to provide an implementation process of the non-contact type brain cognitive load objective detection method based on the distributed radar, and the specific process is as follows:
s1, refer to the corresponding steps in embodiment 1, which are not described herein again;
s2, obtaining a channel with the highest signal-to-noise ratio of a respiratory signal from the multi-channel radar echo signal, and then filtering the wave to obtain the respiratory signal;
s201, referring to the corresponding steps in embodiment 1, which are not described herein again;
s202, IQ signal calibration is carried out on the radar echo signals of each channel after low-pass filtering, and then arctangent demodulation is carried out by using an extended DACM algorithm to obtain demodulation signals of each channel;
the waveform demodulated by each channel is shown in fig. 5.
S203, removing baseline drift of the obtained demodulation signals of all channels, calculating autocorrelation coefficients, and rejecting the channels with lower autocorrelation coefficients according to a threshold value. The autocorrelation coefficients of each channel are shown in table 1 below:
TABLE 1 result table of autocorrelation coefficients of each channel in EXAMPLE 2
| | Channel | 1 | |
|
Channel 4 | Channel 5 | Channel 6 |
| Coefficient of autocorrelation | 0.7667 | 0.9205 | 0.9252 | 0.4854 | 0.9243 | 0.7376 |
The threshold value set according to the corresponding step in embodiment 1 is 0.8, so that the channel 1, the channel 4, and the channel 6 are rejected.
And S204, selecting a channel signal with the highest signal-to-noise ratio of the respiratory signal from the channel 2, the channel 3 and the channel 5 as a final output signal based on the maximum signal-to-noise ratio principle, wherein the signal-to-noise ratio calculation formula is not described herein again with reference to the corresponding steps in the embodiment 1. The signal to noise ratios for channel 2, channel 3 and channel 5 are calculated as shown in table 2 below:
TABLE 2 Signal-to-noise ratio results table in EXAMPLE 2
| | Channel | 2 | |
Channel 5 |
| Signal-to-noise ratio (dB) | 22.7026 | 22.1262 | 22.3434 |
Thus, channel 5 is selected as the final output channel.
And S205, performing band-pass filtering on the final output signal to obtain a respiratory signal. As shown in fig. 6.
S3, refer to the corresponding steps in embodiment 1, which are not described herein again;
and S4, selecting features based on the time domain features and the frequency domain features after the individual differences are removed, and establishing a support vector machine model to detect the cognitive load level of the brain.
S401, refer to the corresponding steps in embodiment 1, which are not described herein again;
s402, referring to the corresponding steps in embodiment 1, which are not described herein again;
s403, refer to the corresponding steps in embodiment 1, which are not described herein again;
s404, the identification of a sample of cognitive load feature vectors for a brain of unknown type can yield 4 results:
(1) the input sample is high mental load, and the input sample is identified and judged as the high mental load and is marked as true high load (TP); (2) the input sample is low mental load, and the input sample is identified and judged as high load and is marked as pseudo high load (FP); (3) the input sample is a low mental load, identified and judged as a low mental load and recorded as a true low load (TN) (4), the input sample is a high mental load, identified and judged as a low mental load and recorded as a false low load (FN).
Based on the results, the following three indexes can be constructed to evaluate the recognition effect, namely:
the TPR reflects the identification accuracy rate of the support vector machine model to the high mental load; the TNR reflects the identification accuracy rate of the support vector machine model to the low mental load; the ACC reflects the overall recognition accuracy of the support vector machine model for all samples.
In the present embodiment, the test results of the support vector machine model are as follows:
TABLE 3 test results of the support vector machine model in EXAMPLE 2
| Test index | ACC | TPR | TNR |
| Results | 84.51％ | 83.12％ | 85.71％ |
Therefore, the support vector machine model has better recognition accuracy for high mental load, low mental load and the whole sample.
FIG. 7 is a ROC plot of a support vector machine model. The ROC curve refers to a receiver operating characteristic curve (receiver operating characteristic curve), is a comprehensive index reflecting sensitivity and specificity continuous variables, the real positive rate of the ROC curve is a vertical coordinate, the false positive rate is a horizontal coordinate, the larger the area under the curve is, the higher the classification accuracy and stability are. The ROC curve graph reflects that the support vector machine model has better accuracy and stability for sample classification.
The above embodiments are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments, and any other changes, modifications, substitutions, combinations, and simplifications which do not depart from the spirit and principle of the present invention should be construed as equivalents thereof, and all such changes, modifications, substitutions, combinations, and simplifications are intended to be included in the scope of the present invention.
Claims (5)
1. A non-contact type brain cognitive load objective detection method based on a distributed radar is characterized by comprising the following steps:
s1, aiming the distributed radar at the thoracic cavity of the human body, and respectively collecting multichannel radar echo signals during sitting and executing cognitive tasks;
s2, obtaining a channel with the highest signal-to-noise ratio of a respiratory signal from the multi-channel radar echo signal, and then filtering the wave to obtain the respiratory signal;
s3, extracting time domain features and frequency domain features from the obtained respiratory signals, and removing individual differences;
and S4, selecting features based on the time domain features and the frequency domain features after the differences are removed, and establishing a support vector machine model to detect the cognitive load level of the brain.
2. The method for objectively detecting cognitive load of a brain based on a distributed radar according to claim 1, wherein the step S1 is performed by aiming the distributed radar at the chest of a human body, and respectively acquiring multi-channel radar echo signals during sitting and performing cognitive tasks as follows:
s101, a distributed radar comprising P radar probes is used, the P radar probes are respectively subjected to frequency division processing, are distributed into sectors around a tester, and respectively acquire radar echo signals of the tester in a sitting state, a low-load cognitive task executing state and a high-load cognitive task executing state, wherein the sitting state is used as a baseline state, a label for executing the low-load cognitive task is set to be-1, and a label for executing the high-load cognitive task is set to be +1, and is used for subsequent classification;
s102, multi-channel synchronous acquisition is carried out by using a multi-bit high-precision ADC of the MCU processor to obtain multi-channel radar echo signals;
s103, wirelessly transmitting the multi-channel radar echo data to a cloud end by using a WiFi technology.
3. The non-contact type brain cognitive load objective detection method based on the distributed radar as claimed in claim 1, wherein the step S2 is to obtain a channel with the highest signal-to-noise ratio of a respiratory signal from the multi-channel radar echo signal, and then obtain the respiratory signal by filtering the wave as follows:
s201, low-pass filtering is carried out on the multi-channel radar echo signals respectively;
s202, I/Q signal calibration is carried out on the radar echo signals of each channel after low-pass filtering, and then arc tangent demodulation is carried out by using an extended DACM algorithm to obtain demodulation signals of each channel;
s203, removing baseline drift of the obtained demodulation signals of each channel, then calculating autocorrelation coefficients, and rejecting the channels with the autocorrelation coefficients lower than a threshold value according to the threshold value;
s204, based on the maximum signal-to-noise ratio principle, selecting a channel signal with the highest signal-to-noise ratio of the respiratory signal from the rest channels as a final output signal, wherein the signal-to-noise ratio calculation formula is as follows:
SNR＝10log(|X is | 2 /E(|X in | 2 ) Equation (1)
Wherein, X is And X in Respectively representing the amplitude values of the breathing signal and the noise of the ith channel in a frequency domain, wherein SNR is the signal-to-noise ratio of the breathing signal in the channel;
and S205, performing band-pass filtering on the final output signal to obtain a respiratory signal.
4. The method for objectively detecting cognitive load of the brain based on the distributed radar as claimed in claim 1, wherein the step S3 is to extract time domain features and frequency domain features from the obtained respiratory signal, and the process of removing the individual difference is as follows:
s301, searching a peak point and a valley point of the respiratory signal, and extracting a respiratory signal time domain feature sequence according to the peak point and the valley point; wherein the time domain feature sequence comprises 8 feature sequences of inspiration time, expiration amplitude, breathing interval, periodic ventilation volume, average respiration rate, average tidal volume and minute inspiration volume which are sequentially and respectively marked as X i I belongs to {1,2,3,4,5,6,7,8}, wherein the average respiration rate, the average tidal volume and the minute inspiration volume are respectively calculated by using a sliding window with the window length of M and the sliding step length of N;
s302, respectively calculating the mean value, the median value, the standard deviation, the root mean square, the maximum value, the minimum value, the variation coefficient and the four-quadrant spacing of the characteristic sequence as time domain characteristics;
wherein the mean value u of the ith characteristic sequence i Standard deviation σ i And root mean square rms i The calculation formulas of (A) are respectively as follows:
wherein, X i [n]Is the N-th element of the i-th signature sequence, N i Represents the sequence length of the ith characteristic sequence;
coefficient of variation CV of ith signature sequence i The calculation formula is as follows:
s303, carrying out fast Fourier transform on the obtained respiratory signal to obtain a frequency spectrum X [ k ], extracting a frequency mean value and a frequency standard deviation as frequency domain characteristics, wherein the calculation formulas are respectively as follows:
wherein f _ mean is a frequency mean value, f _ std is a frequency standard deviation, X [ K ] is a fast Fourier transform spectrum, and K is a signal length;
s304, respectively subtracting the time domain characteristic and the frequency domain characteristic of the respiration signal in the sedentary state from the time domain characteristic and the frequency domain characteristic of the respiration signal in the cognitive task, so as to remove the individual difference and obtain the time domain characteristic and the frequency domain characteristic after the individual difference is removed.
5. The method for objectively detecting a cognitive load of a brain based on a distributed radar according to claim 1, wherein in step S4, feature selection is performed based on time domain features and frequency domain features after individual differences are removed, and a process of establishing a support vector machine model to detect the cognitive load level of the brain is as follows:
s401, respectively carrying out normalization processing on the time domain characteristics and the frequency domain characteristics after the individual differences are removed, and unifying the numerical values of the characteristics to 0-1 to obtain the normalized time domain characteristics and frequency domain characteristics;
s402, performing mutual information screening on the time domain characteristics and the frequency domain characteristics obtained after normalization in the step S401, and selecting the first m characteristics as final brain cognitive load characteristics;
s403, sending the first m features in the step S402 into a support vector machine model for training, and setting a training sample set as { x } m ,y m }，y m E { -1, +1}, m { -1,.. and L, L is the number of samples, in order to solve the nonlinear classification problem, a nonlinear kernel support vector machine is adopted, and a decision function f (x) and an RBF kernel function formula of the nonlinear kernel support vector machine are as follows:
wherein,
is a Lagrange multiplier, b * To classify threshold, x m Is the m-th input variable, x n Is the nth input variable, y m Is the m-th output value, k (x) m ,x n ) Is a kernel function, and sigma is an RBF kernel penalty coefficient;
s404, after feature normalization processing is carried out on a brain cognitive load feature vector sample of an unknown type, the decision function is calculated, brain cognitive load level detection is carried out according to the values of f (x), and if f (x) -1 is a label-1, the brain cognitive load level is represented; if f (x) >1 is the label +1, it represents a high cognitive load level in the brain.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210512666.XA CN114983412B (en) | 2022-05-12 | 2022-05-12 | Non-contact brain cognitive load objective detection method based on distributed radar |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210512666.XA CN114983412B (en) | 2022-05-12 | 2022-05-12 | Non-contact brain cognitive load objective detection method based on distributed radar |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114983412A true CN114983412A (en) | 2022-09-02 |
| CN114983412B CN114983412B (en) | 2024-07-26 |
Family
ID=83027660
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210512666.XA Active CN114983412B (en) | 2022-05-12 | 2022-05-12 | Non-contact brain cognitive load objective detection method based on distributed radar |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114983412B (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109031287A (en) * | 2018-09-21 | 2018-12-18 | 西安交通大学 | ULTRA-WIDEBAND RADAR human body respiration signal detecting method through walls based on Faster-RCNN network |
| CN109875529A (en) * | 2019-01-23 | 2019-06-14 | 北京邮电大学 | A kind of vital sign detection method and system based on ULTRA-WIDEBAND RADAR |
| CN110192850A (en) * | 2019-05-31 | 2019-09-03 | 湖南省顺鸿智能科技有限公司 | Extracting method and system based on heartbeat signal under radar return strong noise background |
| CN110200640A (en) * | 2019-05-14 | 2019-09-06 | 南京理工大学 | Contactless Emotion identification method based on dual-modality sensor |
| CN111657889A (en) * | 2020-06-03 | 2020-09-15 | 大连理工大学 | Non-contact driver fatigue detection method based on millimeter wave radar |
| EP3747359A1 (en) * | 2019-06-06 | 2020-12-09 | Tata Consultancy Services Limited | Method and system for assessment of cognitive workload using breathing pattern of a person |
| CN112137620A (en) * | 2020-08-27 | 2020-12-29 | 广东省地震局 | Ultra-wideband radar-based human body weak respiration signal detection method |
| US20210093203A1 (en) * | 2019-09-30 | 2021-04-01 | DawnLight Technologies | Systems and methods of determining heart-rate and respiratory rate from a radar signal using machine learning methods |
| CN112754444A (en) * | 2021-02-22 | 2021-05-07 | 牧原食品股份有限公司 | Radar-based non-contact pig respiration detection method |
| WO2021109526A1 (en) * | 2019-12-03 | 2021-06-10 | 南京汇君半导体科技有限公司 | Vital sign signal enhancement method and device, and extraction method and device based on millimeter-wave radar |
| CN113261925A (en) * | 2021-03-11 | 2021-08-17 | 山东师范大学 | Human body sensing method and system based on millimeter wave radar |
| CN114052709A (en) * | 2021-12-02 | 2022-02-18 | 中国电子科技南湖研究院 | Stable millimeter wave radar vital sign measuring method |
| CN114305452A (en) * | 2021-12-15 | 2022-04-12 | 南京航空航天大学 | Cross-task cognitive load identification method based on electroencephalogram and field adaptation |
-
2022
- 2022-05-12 CN CN202210512666.XA patent/CN114983412B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109031287A (en) * | 2018-09-21 | 2018-12-18 | 西安交通大学 | ULTRA-WIDEBAND RADAR human body respiration signal detecting method through walls based on Faster-RCNN network |
| CN109875529A (en) * | 2019-01-23 | 2019-06-14 | 北京邮电大学 | A kind of vital sign detection method and system based on ULTRA-WIDEBAND RADAR |
| CN110200640A (en) * | 2019-05-14 | 2019-09-06 | 南京理工大学 | Contactless Emotion identification method based on dual-modality sensor |
| CN110192850A (en) * | 2019-05-31 | 2019-09-03 | 湖南省顺鸿智能科技有限公司 | Extracting method and system based on heartbeat signal under radar return strong noise background |
| EP3747359A1 (en) * | 2019-06-06 | 2020-12-09 | Tata Consultancy Services Limited | Method and system for assessment of cognitive workload using breathing pattern of a person |
| US20210093203A1 (en) * | 2019-09-30 | 2021-04-01 | DawnLight Technologies | Systems and methods of determining heart-rate and respiratory rate from a radar signal using machine learning methods |
| WO2021109526A1 (en) * | 2019-12-03 | 2021-06-10 | 南京汇君半导体科技有限公司 | Vital sign signal enhancement method and device, and extraction method and device based on millimeter-wave radar |
| CN111657889A (en) * | 2020-06-03 | 2020-09-15 | 大连理工大学 | Non-contact driver fatigue detection method based on millimeter wave radar |
| CN112137620A (en) * | 2020-08-27 | 2020-12-29 | 广东省地震局 | Ultra-wideband radar-based human body weak respiration signal detection method |
| CN112754444A (en) * | 2021-02-22 | 2021-05-07 | 牧原食品股份有限公司 | Radar-based non-contact pig respiration detection method |
| CN113261925A (en) * | 2021-03-11 | 2021-08-17 | 山东师范大学 | Human body sensing method and system based on millimeter wave radar |
| CN114052709A (en) * | 2021-12-02 | 2022-02-18 | 中国电子科技南湖研究院 | Stable millimeter wave radar vital sign measuring method |
| CN114305452A (en) * | 2021-12-15 | 2022-04-12 | 南京航空航天大学 | Cross-task cognitive load identification method based on electroencephalogram and field adaptation |
Non-Patent Citations (1)
| Title |
|---|
| 文乐;方晖;张兴敢;: "基于MUSIC算法的生命探测雷达信号处理方法", 南京大学学报(自然科学), no. 05, 30 September 2016 (2016-09-30), pages 946 - 952 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114983412B (en) | 2024-07-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109907752A (en) | A kind of cardiac diagnosis and monitoring method and system of the interference of removal motion artifacts and ecg characteristics detection | |
| Samieinasab et al. | Fetal phonocardiogram extraction using single channel blind source separation | |
| TW201902416A (en) | Cardiovascular testing device and method | |
| John et al. | A multimodal data fusion technique for heartbeat detection in wearable IoT sensors | |
| US20220110590A1 (en) | System and Method for Assessing Physiological Signal Quality | |
| WO2019100563A1 (en) | Method for assessing electrocardiogram signal quality | |
| CN109833034A (en) | The method and device of blood pressure data is extracted in a kind of pulse wave signal | |
| CN104305992B (en) | A kind of interactive fast automatic extracting method of Fetal ECG | |
| CN114469124B (en) | Method for identifying abnormal electrocardiosignals in movement process | |
| TW201711628A (en) | Heart rate detection method and heart rate detection device | |
| KR20210001575A (en) | Real-time cardiac rate detection apparatus in noisy environment and method thereof | |
| CN113558584A (en) | Pulse wave preprocessing method based on signal quality evaluation | |
| CN112998690A (en) | Pulse wave multi-feature fusion-based respiration rate extraction method | |
| Zhang et al. | Using rear smartphone cameras as sensors for measuring heart rate variability | |
| Zheng et al. | Identification of chronic heart failure using linear and nonlinear analysis of heart sound | |
| CN110070013A (en) | A kind of ECG signal noise pollution degree and classification intelligent evaluation method | |
| CN110403595B (en) | Signal quality evaluation method of wearable electrocardiogram monitoring equipment | |
| CN114983412B (en) | Non-contact brain cognitive load objective detection method based on distributed radar | |
| CN115363598B (en) | Electrocardiosignal processing method and system | |
| CN105050493B (en) | For determining the apparatus and method of the appearance of the QRS complex in ECG data | |
| Quanyu | Pulse signal analysis based on deep learning network | |
| Yu et al. | Fusion of detected multi-channel maternal electrocardiogram (ECG) R-wave peak locations | |
| JP5933138B2 (en) | Apparatus and method for determining occurrence of QRS complex in ECG data | |
| CN114343664A (en) | Heart and lung endurance evaluation method and device based on heart rate stimulation-recovery trend | |
| CN111345815B (en) | Method, device, equipment and storage medium for detecting QRS wave in electrocardiosignal |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |