CN112998701A - Vital sign detection and identity recognition system and method based on millimeter wave radar - Google Patents
Vital sign detection and identity recognition system and method based on millimeter wave radar Download PDFInfo
- Publication number
- CN112998701A CN112998701A CN202110331449.6A CN202110331449A CN112998701A CN 112998701 A CN112998701 A CN 112998701A CN 202110331449 A CN202110331449 A CN 202110331449A CN 112998701 A CN112998701 A CN 112998701A
- Authority
- CN
- China
- Prior art keywords
- vital sign
- subsystem
- signal
- radar
- heartbeat
- 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
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/117—Identification of persons
-
- 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
- A61B5/0205—Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
-
- 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)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- Physiology (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Cardiology (AREA)
- Evolutionary Computation (AREA)
- Signal Processing (AREA)
- Psychiatry (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Fuzzy Systems (AREA)
- Pulmonology (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention relates to a vital sign detection and identity recognition system based on a millimeter wave radar, which comprises a millimeter wave radar subsystem, a radar data preprocessing subsystem, a vital sign characteristic database subsystem, an identity recognition subsystem and an intelligent interactive interface subsystem, wherein the millimeter wave radar subsystem, the radar data preprocessing subsystem, the vital sign characteristic database subsystem, the identity recognition subsystem and the intelligent interactive interface subsystem are in communication connection with each other; the radar data preprocessing subsystem decomposes a respiratory signal and a heartbeat signal in original radar data and extracts vital sign characteristic information; the vital sign characteristic database subsystem is used for storing the identity information and the corresponding vital sign characteristic information; the identity recognition subsystem learns the vital sign characteristic information, classifies and determines corresponding identity information; compared with the prior art, the intelligent interactive interface subsystem has the advantages of high practicability, privacy protection, safety and the like.
Description
Technical Field
The invention relates to the technical field of vital sign detection and identification, in particular to a system and a method for vital sign detection and identity identification based on a millimeter wave radar.
Background
With the development and improvement of the technology level, the technical requirements in various fields such as intelligent life, intelligent home, intelligent medical treatment and the like are increasingly strong. Current vital sign is mainly relied on traditional contact sensor to realize like breathing heartbeat detection technique, like patent CN201811523766.2, but the contact sensor in this patent not only wears the procedure loaded down with trivial details, removes inconveniently, wears for a long time and still can bring uncomfortable sense, can't realize the continuous guardianship of whole day.
The non-contact vital sign detection equipment not only can monitor the health state of people in real time, but also can be applied to the field of security monitoring and used for realizing personnel detection and identity recognition. The relatively common monitoring technology in the field of security monitoring at present is an optical image camera, but the monitoring effect of the camera is greatly influenced by environments such as illumination, smoke dust and shielding, and the security monitoring effect is easily lost under the condition of shielding objects.
Identity recognition based on vital signs such as breath and heartbeat is a non-contact biological feature recognition technology, and the purpose of identity recognition is achieved mainly by extracting respiratory motion features between people. Currently, common identification technologies include fingerprint recognition as described in CN201910431503.7, face recognition as described in CN201811250425.2, and the like. However, face recognition is easily shielded by illumination smoke and dust, and both face recognition and fingerprint recognition have hidden danger of privacy disclosure, which causes information leakage and information safety problems.
Therefore, it is an urgent problem to improve the practicability, universality and privacy protection capability of the identity recognition technology.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a vital sign detection and identity recognition system and method based on a millimeter wave radar, which have high practicability, universality and privacy protection capability.
The purpose of the invention can be realized by the following technical scheme:
a vital sign detection and identity recognition system based on a millimeter wave radar comprises a millimeter wave radar subsystem, a radar data preprocessing subsystem, a vital sign characteristic database subsystem, an identity recognition subsystem and an intelligent interactive interface subsystem which are in communication connection with each other;
the millimeter wave radar subsystem is arranged beside an object to be detected, comprises a transmitting antenna and a receiving antenna, and is used for transmitting a linear frequency modulation continuous wave signal, collecting radar echo reflected on the surface of a static human thorax and obtaining original radar data;
the radar data preprocessing subsystem decomposes a respiratory signal and a heartbeat signal in original radar data and extracts vital sign characteristic information;
the vital sign characteristic database subsystem is used for storing the identity information of a plurality of individuals and the corresponding vital sign characteristic information;
the identity recognition subsystem learns the vital sign characteristic information, classifies and determines corresponding identity information;
the intelligent interactive interface subsystem comprises a display screen and is used for displaying the vital sign information and the identity recognition result of the object to be detected in real time.
Furthermore, a transmitting antenna of the millimeter wave radar subsystem transmits a linear frequency modulation continuous wave signal to an area in a detection range, a receiving antenna receives a radar echo reflected by the surface of the chest of a static human body, the received radar echo and the transmitting signal are subjected to frequency mixing, low-pass filtering is performed, and original radar data are obtained after analog-to-digital conversion, wherein the original radar data comprise vital sign information and various noises.
Further, the radar data preprocessing subsystem decomposes the original radar data signal by adopting a smoothing filtering algorithm, a band-pass filtering algorithm and/or an improved empirical mode decomposition algorithm to obtain a respiration signal waveform and a heartbeat signal waveform.
Furthermore, when the radar data preprocessing subsystem decomposes the respiratory signal and the heartbeat signal in the original radar data, the method specifically comprises the following steps:
1) selecting a signal with a certain frequency modulation period in original radar data, and performing fast Fourier transform to obtain target distance distribution;
2) detecting target points according to the target distance distribution, and extracting target point radar data, wherein the target point radar data are data containing thoracic cavity movement;
3) carrying out deburring processing on target point radar data;
4) and decomposing the original radar data signal to obtain a respiration signal waveform and a heartbeat signal waveform.
Furthermore, the radar data preprocessing subsystem combines a frequency spectrum interpolation method and a signal peak searching algorithm, firstly extracts the respiration rate and the heart rate from the respiration signal waveform and the heartbeat signal waveform, and then extracts a plurality of vital sign characteristic information by combining the respiration signal waveform, the heartbeat signal waveform, the respiration rate and the heart rate.
Further, the vital sign feature information includes:
respiratory amplitude, respiratory frequency, respiratory peak-to-peak time interval, respiratory trough-to-trough time interval, peak-to-trough Euclidean distance, peak-to-trough time, peak-to-trough velocity, trough-to-peak distance, trough-to-peak time, trough-to-peak velocity, tidal volume, heartbeat amplitude, and heartbeat frequency.
Furthermore, the smoothing filter algorithm is used for obtaining a respiratory signal close to the real respiratory motion, the Gaussian weighted average value in each window is calculated through a sliding window, and a smooth value of the respiratory signal changing along with the time is obtained, wherein the range of the sliding window is set as-;
the band-pass filtering algorithm is used for separating heartbeat signals from original radar data signals superposed by breathing and heartbeat, the band-pass filtering algorithm adopts an elliptical band-pass filter, the frequency band of a pass band is set to be Hz-Hz, the frequency of a stop band is set to be Hz-Hz, and the maximum attenuation allowed by the pass band and the stop band is respectively set to be dB and dB;
the improved empirical mode decomposition algorithm is used for decomposing the respiration heartbeat superposition signal to respectively obtain a respiration signal and a heartbeat signal.
Furthermore, the frequency spectrum interpolation method firstly carries out short-time Fourier transform on the respiration signal and the heartbeat signal to obtain discrete frequency spectrum values, then finds out the peak value of the frequency spectrum values, carries out three-point secondary frequency spectrum interpolation on the peak value, finds out the frequency corresponding to the peak value and obtains the respiration rate and the heart rate;
the signal peak searching algorithm sets the minimum interval number between wave peaks according to the range of the heart rate and the respiration rate, and then sets the minimum peak value number according to the extracted respiration and heartbeat amplitude values.
Further, the identity recognition subsystem performs offline or online training on various vital sign characteristic information of different individuals to generate a classification model, and then performs identity recognition by using the classification model;
after collecting the data of a set number of existing identity information individuals or new individuals, the identity recognition subsystem carries out online updating or replacement of the classification model;
the classification model comprises a convolutional neural network, a long-term and short-term memory network, a nearest node algorithm and/or a k-means algorithm.
An identity recognition method of the millimeter wave radar-based vital sign detection and identity recognition system includes the following steps:
s1: the millimeter wave radar transmits linear frequency modulation continuous wave signals to the detection area, radar echoes reflected by the surface of the chest of a static human body are collected, then the received radar echoes and the transmitted waveforms are subjected to low-pass filtering after being subjected to frequency mixing, and original radar data are obtained after analog-to-digital conversion;
s2: carrying out fast Fourier transform on original radar data, filtering static target echoes by utilizing a moving target display algorithm to obtain a sign point phase signal containing human body respiration and heartbeat motion information, then carrying out filtering processing on the phase signal to obtain a respiration signal waveform and a heartbeat signal waveform of a detected object, and then obtaining a respiration rate and a heart rate from the respiration signal waveform and the heartbeat signal waveform;
s3: carrying out time domain and frequency domain data processing on the respiration signal waveform and the heartbeat signal waveform, extracting vital sign characteristic information, and constructing a vital sign characteristic database;
s4: off-line or on-line training is carried out on the vital sign characteristic information of a plurality of individuals to generate a classification model, and classification and identification are carried out on the vital sign characteristic information by utilizing the classification model to obtain identity information;
s5: and updating and displaying the vital sign information of the tested object and the identity information of the tested object on a display screen in real time.
Compared with the prior art, the invention has the following advantages:
1) the invention realizes identity recognition through the respiratory motion characteristics, is not easy to forge, can realize non-contact and remote recognition, does not cause the privacy disclosure problem, realizes comfortable and flexible monitoring of the vital signs all day long, and has high safety;
2) the invention adopts non-contact detection, does not need to wear equipment, can detect all the day, utilizes the millimeter wave radar for detection, has long sensing distance, convenient movement, good flexibility and low environmental influence degree, can work in challenging environments, such as the environment without illumination, with smoke and with non-metallic shielding objects, and has high practicability and universality;
3) the millimeter wave radar of the invention is used for signal acquisition, has high concealment, is not easy to be found and damaged, adopts low-power millimeter waves, is harmless to human bodies, and is safe and reliable.
Drawings
FIG. 1 is a schematic diagram of an application scenario in the structure and embodiment of the present invention;
FIG. 2 is a graph of the detection result of the respiration signal according to the embodiment of the present invention;
FIG. 3 is a diagram illustrating a detection result of a heartbeat signal according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of peaks and troughs of a respiration signal waveform according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a frequency domain interpolation algorithm of the present invention;
FIG. 6 is a flow chart of raw radar data signal preprocessing according to the present invention;
FIG. 7 is a schematic diagram of the operation of the system of the present invention.
The system comprises a millimeter wave radar subsystem 1, a millimeter wave radar subsystem 11, a power supply assembly 13, a transmitting antenna 14, a receiving antenna 2, a radar data preprocessing subsystem 3, a vital sign characteristic database subsystem 4, an identity recognition subsystem 5, an interactive display subsystem 6, an object to be detected 7 and a computer.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments. It is to be understood that the embodiments described are only a few embodiments of the present invention, and not all embodiments. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, shall fall within the scope of protection of the present invention.
Examples
The invention discloses a vital sign detection and identity recognition system based on a millimeter wave radar, which performs identity recognition by utilizing different individual breathing modes, and researches on Girara and Berkenttrit researchers at the France Grainunbell university Hospital center show that the breathing modes of different individuals are very different, the breathing mode is a relatively stable characteristic of adults, and the breathing modes of different people are very different in the aspects of tidal volume, breathing frequency, airflow shape and the like.
As shown in fig. 1 and 7, the system includes a millimeter wave radar subsystem 1, a radar data preprocessing subsystem 2, a vital sign feature database subsystem 3, an identity recognition subsystem 4, and an intelligent interactive interface subsystem 5, which are communicatively connected to each other.
The hardware part of the millimeter wave radar subsystem 1 comprises a transmitting antenna 13, a receiving antenna 14, a radar radio frequency front end, a crystal oscillator, a communication component, an analog-to-digital converter, a Micro Control Unit (MCU), a power supply component 11 and a peripheral matching circuit; the hardware part of the radar data preprocessing subsystem 2 comprises a device with certain signal processing capacity, such as an MCU, a Digital Signal Processor (DSP), a field Programmable Gate Array (PGA), a smart phone, a computer and the like; the hardware part of the vital sign characteristic database subsystem 3 has a data storage function, and comprises a mechanical hard disk, a solid state hard disk and the like; the hardware part of the identity recognition subsystem 4 comprises embedded equipment, a smart phone, a computer, a server and other equipment with machine learning and deep learning support; the hardware part of the interactive display subsystem 5 is provided with a display screen, so that the human-computer interaction function can be realized. Hardware facilities such as interfaces are required to be provided between the subsystems, and the functions of communication and data transmission between the subsystems can be realized.
As shown in fig. 1, the millimeter wave radar subsystem 1 is placed beside an object to be measured 6, a transmitting antenna 13 transmits a linear frequency modulation continuous wave signal, and a receiving antenna 14 collects radar echoes reflected on the surface of a static human thorax to obtain original radar data; the radar data preprocessing subsystem 2 decomposes a respiratory signal and a heartbeat signal in original radar data and extracts vital sign characteristic information; the vital sign characteristic database subsystem 3 is used for storing the identity information of a plurality of individuals and the corresponding vital sign characteristic information; the identity recognition subsystem 4 learns the vital sign characteristic information, classifies and determines corresponding identity information; the intelligent interactive interface subsystem 5 comprises a display screen and is used for displaying the vital sign information and the identity recognition result of the object to be detected in real time.
Specifically, the millimeter wave radar subsystem 1 transmits a linear frequency modulation continuous wave signal, collects radar echoes reflected on the surface of a static human thorax, performs low-pass filtering after frequency mixing, and obtains original radar data containing vital sign information and various noises after analog-to-digital conversion.
The millimeter wave radar subsystem 1 has different required radar working parameters according to different application scenes, and can adjust parameters such as the maximum measurement distance, the maximum measurement speed, the distance resolution, the speed resolution, the frame period and the like of the system according to the requirements of the application scenes. The above system parameters are determined by the following radar parameters: the number of antennas of the radar for transmitting linear frequency modulation continuous wave signals, the number of antennas of the radar for receiving echo signals, the number of frequency modulation cycles of each frame, the frequency modulation slope, the frequency modulation cycle, the analog-to-digital conversion sampling rate and the number of sampling points of each frequency modulation cycle.
The main process of the millimeter wave radar subsystem 1 for acquiring the original radar signal is as follows: m transmitting antennas of the millimeter wave radar transmit linear frequency modulation continuous wave signals to an area in a detection range, N receiving antennas are used for receiving radar echoes reflected by the surface of a static human thorax, the received radar echoes and the transmitting signals are subjected to low-pass filtering after frequency mixing, and original radar data containing vital sign information and various noises are obtained after analog-to-digital conversion.
The radar data preprocessing subsystem 2 can perform digital signal processing on original radar data, extracts respiration, heartbeat waveforms and rates containing a test object and vital sign characteristic information, and mainly comprises two steps of performing digital signal processing on the original radar data to extract individual respiration heartbeat information and performing characteristic extraction on the extracted respiration heartbeat information.
Specifically, the process of extracting the individual respiration and heartbeat information in the radar data preprocessing subsystem 2 is as follows:
1. the size of each frame of original radar data collected by the radar is m x n, wherein m represents frequency modulation period index, n represents frequency modulation period sampling point number, and the target distance distribution can be obtained by performing Fast Fourier Transform (FFT) on signals of a certain frequency modulation period.
2. After the target distance distribution is obtained, target detection is needed, threshold detection is carried out on target distance distribution data, target distance points are screened out, and radar data corresponding to the target points are extracted, namely the data containing thoracic cavity movement.
3. After the human body sign data are obtained, interpolation values of each data point and previous and next data points are compared, correction is carried out when the interpolation values are too large, the effect of removing sharp burr clutters contained in the data is achieved, and a median filtering algorithm can be adopted to remove the prominent burrs so as to avoid the influence of wrong data.
The following method can be adopted for data correction:
assuming that the original data is f (x), performing Constant False Alarm Rate (CFAR) detection on f (x), and when f (a) is detected, judging whether f (a-1) and f (a +1) are targets or not, and if not, making h (a) -k × f (a), wherein k is an attenuation factor, and combining the relation between burrs and non-noise signals, and k is 0.4-0.6. And after finishing correcting the single target point, performing second correction, judging the relationship among f (a-1), f (a +1) and f (a) when the first f (a) is detected, and if f (a) is in an ascending trend, making h (a) f (a-1) + m, and if f (a) is in a descending trend, making h (a) f (a-1) + m, combining the relationship between the burr and the non-noise signal, wherein the value range of m is 0.1-0.3.
Another method for data correction and deburring is median filtering, assuming that the original data is f (x), and the median filtered value corresponding to the index a is:
4. and separating and extracting respiratory and heartbeat signals by using algorithms such as smooth filtering, band-pass filtering or improved empirical mode decomposition (MEEMD), wherein:
(1) the smoothing filtering algorithm is an effective method for obtaining a respiratory signal closer to a real respiratory motion, and the algorithm content is that a sliding window calculates a Gaussian weighted average value in each window to obtain a smooth value changing along with time. The gaussian weighting uses a gaussian function, i.e. a normal distribution function, whose probability density function is:
where σ is the standard deviation.
When the index is a, μ is a, f (x) is the weight corresponding to the value of the index value x, and when the window is 2n +1, the corresponding value at a after gaussian smoothing can be calculated by the following formula:
since the sum of the weights of 2n +1 points is not 1, normalization is required:
and I (a) is a value after Gaussian smoothing, the window range is generally set to be 80-120 aiming at the respiratory signal, and 0 needs to be supplemented when a-n or a + n does not exist.
(2) The band-pass filtering algorithm is a method for separating accurate heartbeat signals from respiratory heartbeat superposed signals, the algorithm content is to set a pass band and a stop band frequency band of a band-pass filter, so that signals of a specific frequency band in data pass through and signals of other frequency bands are suppressed, the band-pass filter comprises an Infinite Impulse Response (IIR) filter and a Finite Impulse Response (FIR) filter, the fluctuation in an elliptic filter in the Infinite Impulse Response filter is small, a transition band is narrow, the band-pass filter is a band-pass filter with better performance, the heartbeat signals are obtained by adopting the elliptic band-pass filter, the pass band frequency band is generally set to be 0.6 Hz-2 Hz, the stop band frequency is set to be 0.55-3 Hz, and the maximum attenuation allowed by the pass band and the stop band is respectively set to be 3dB and 40 dB.
(3) The respiratory heartbeat superposition signal can be decomposed by improving an empirical mode decomposition algorithm (MEEMD) to obtain respiratory and heartbeat signals, the MEEMD can effectively inhibit modal aliasing, unnecessary false components and false components cannot be generated in a decomposition result, the decomposition result is well matched with an actual signal, and the decomposition effect is good.
For the non-stationary signal s (t), the decomposition process for MEEMD is as follows:
white noise signals n with the average value of 0 are respectively added into the original signals S (t)i(t) and-ni(t), obtaining:
wherein n isi(t) represents an added white noise signal, ciDenotes the magnitude of a white noise signal, i is 1, 2, …, N denotes the logarithm of the added white noise. Are respectively paired
And
performing Empirical Mode Decomposition (EMD) to obtain a first-order Intrinsic Mode Function (IMF) component sequence,
and
the integration averages the above-obtained components:
examination S1(t) is an abnormal signal. If the entropy value is larger than 0.55-0.6, the signal is determined to be an abnormal signal, otherwise, the signal is determined to be a stable signal. If S is1(t) is an abnormal signal, the above steps are continued until the IMF component Sp(t) is a non-abnormal signal. Separating the decomposed first p-1 components from the original signal, namely:
EMD decomposition is performed on the residual signal r (t), and signal components are arranged according to the frequency.
In addition, besides the three algorithms for signal extraction and decomposition with good effect, the algorithm for time frequency analysis such as fourier transform and wavelet transform can also be used for signal decomposition.
5. The method comprises the following steps of extracting a respiratory rate and a heart rate from respiratory and heartbeat signals, wherein the step comprises two implementation methods, namely a frequency spectrum interpolation method and a signal peak searching algorithm, the two methods are supervised with each other and refer to each other, wherein:
(1) spectral interpolation is performed by analyzing and computing the frequency domain characteristics of the signal to derive the respiration rate and the heart rate. First, the signal is short-timedAnd Fourier Transform (STFT) is carried out to obtain discrete spectral values, the peak value of the spectral values is found, and three-point secondary spectrum interpolation is carried out on the peak value. Suppose that the frequency at which the peak is located is f2Peak value of S2Respectively at f2Left and right at the same distance f1And f3Two points, corresponding to a peak value of S1And S3According to the formula:
the frequency f corresponding to the peak can be foundmax。
(2) The signal peak-finding algorithm is to analyze and calculate the time domain characteristics of the signal to obtain the respiration rate and the heart rate. And setting the minimum interval number between wave peaks according to the range of the heart rate and the respiration rate, and setting the minimum peak number according to the extracted respiration and heartbeat amplitude values. The breathing frequency of a normal adult in a resting state is about 12-20 times/minute, a female is 1-2 times/minute faster than a male, and a child is about 20 times/minute; the heartbeat frequency of the adult in the resting state is about 60-100 times/min. Therefore, the respiration rate ranges from 0.15 Hz to 0.35Hz, and the heart rate ranges from 0.85 Hz to 2 Hz.
The vital sign feature information extracted from the radar data preprocessing subsystem 1 comprises a plurality of variables which change along with time, wherein the variables comprise: respiratory amplitude, respiratory frequency, respiratory peak-to-peak time interval, respiratory trough-to-trough time interval, peak-to-trough Euclidean distance, peak-to-trough time, peak-to-trough velocity, trough-to-peak distance, trough-to-peak time, trough-to-peak velocity, tidal volume, heartbeat amplitude, and heartbeat frequency. The variables are defined as follows:
breathing amplitude: the amplitude of the respiration waveform;
breathing frequency: number of breaths per minute;
respiratory peak-to-peak and trough-to-trough time intervals: assuming that the amplitude maximum is A in the nth period of the respiration signal including a peak and a troughnmax, minimum amplitude Anmin, corresponding time tnmax and tnmin, the maximum and minimum amplitude values of the subsequent cycle, i.e. the (n +1) th cycle, being An+1max and An+1min, corresponding time tn+1max and tn+1min, the time interval between respiratory peaks and the time interval between respiratory troughs are respectively tn+1max-tnmax and tn+ 1min-tnmin；
The euclidean distance from the peak to the trough is:
the euclidean distance from trough to crest is:
the peak to trough time is:
|tnmax-tnmin|
the trough to peak times are:
|tn+1max-tnmin|
the peak to trough speed is:
the trough to peak velocities were:
tidal volume: refers to the amount of gas inhaled or exhaled per breath, and can be expressed in an integrated form in the respiratory waveform.
Heartbeat amplitude: amplitude of the heartbeat waveform.
Heartbeat frequency: number of heartbeats per minute.
The individual's link differences between the peaks and troughs of the respiration can be further estimated on the basis of these features, including but not limited to the included information of time, space, power spectral density, etc.
The vital sign feature database subsystem 3 stores the extracted vital sign feature data set, specifically, stores a plurality of individual identities and a plurality of vital sign features corresponding thereto.
The identity recognition subsystem 4 is used for learning the vital sign characteristic information, classifying and judging, determining the identity information, performing offline or online training on various vital sign characteristic data of different people to generate a classification model, then performing identity recognition by using the model, and performing online updating or replacement of the classification model after collecting a certain amount of data of an existing identity information individual or a new individual, wherein the classification model includes but is not limited to: convolutional Neural Networks (CNN), long short term memory networks (LSTM), nearest neighbor node algorithms (KNN), and k-means algorithms (KMN).
The interactive display subsystem 5 is configured to update and display the respiration heartbeat waveform and the rate information of the test object and the identification result of the test object on the display screen in real time, and specifically update and display the respiration heartbeat waveform and the rate information of the test object and the identification result of the test object on the display screen in real time.
As shown in fig. 7, the present invention further provides an identity recognition method of the vital sign detection and identity recognition system based on the millimeter wave radar, comprising the following steps:
s1: the millimeter wave radar transmits linear frequency modulation continuous wave signals to the detection area, radar echoes reflected by the surface of the chest of a static human body are collected, then the received radar echoes and the transmitted waveforms are subjected to low-pass filtering after being subjected to frequency mixing, and original radar data are obtained after analog-to-digital conversion;
s2: carrying out fast Fourier transform on original radar data, filtering static target echoes by utilizing a moving target display algorithm to obtain sign point phase signals containing human respiration and heartbeat motion information, carrying out filtering processing on the phase signals to obtain respiration signal waveforms and heartbeat signal waveforms of a detected object, and carrying out peak searching or frequency domain processing on the respiration signal waveforms and the heartbeat signal waveforms to obtain respiration rate and heart rate;
s3: carrying out time domain and frequency domain data processing on the respiration signal waveform and the heartbeat signal waveform, extracting vital sign characteristic information, and constructing a vital sign characteristic database;
s4: off-line or on-line training is carried out on the vital sign characteristic information of a plurality of individuals to generate a classification model, and classification and identification are carried out on the vital sign characteristic information by utilizing the classification model to obtain identity information;
s5: and updating and displaying the vital sign information of the tested object and the identity information of the tested object in real time on a display screen, wherein the vital sign information specifically comprises the respiration heartbeat waveform and the rate information of the tested object of 6 and the identity information of the tested object.
A specific application scenario of this embodiment is given below, and as shown in fig. 1, a specific implementation process is as follows:
(1) arranging a millimeter wave radar and signal processing integrated module in a scene, connecting a computer 7 through a USB data line, establishing communication and starting the radar;
(2) standing or sitting an object 6 to be detected in a detection area of the millimeter wave radar, keeping the object static, detecting the thoracic cavity movement of a human body by the millimeter wave radar, and acquiring original digital intermediate frequency data containing the thoracic cavity movement caused by the respiration and heartbeat of the human body;
(3) processing the original data on a computer 7, respectively extracting a respiration signal waveform and a heartbeat signal waveform by utilizing smooth filtering and band-pass filtering, and processing the respiration and heartbeat waveforms by adopting a frequency domain interpolation algorithm and a time domain peak searching algorithm to obtain a respiration rate and a heart rate, wherein the extracted respiration signal waveform and the extracted heartbeat signal waveform are shown in figures 2 and 3, and the frequency domain interpolation algorithm is shown in figure 5;
(4) repeating the step (2) and the step (3), obtaining respiratory waveforms of 5 persons, carrying out short-time Fourier transform to obtain one-dimensional respiratory waveforms and two-dimensional respiratory time-frequency characteristic graphs, constructing corresponding one-dimensional and two-dimensional vital sign characteristic data sets, and storing 100 samples of each person in a computer 7;
(5) respectively sending the one-dimensional and two-dimensional vital sign feature data sets to a designed long-short term memory network (LSTM) and a designed Convolutional Neural Network (CNN) on a computer 7, and training to obtain a multi-classification model;
(6) and calling a multi-classification model on the computer to classify the vital sign features transmitted in real time, so as to realize the identification based on the vital sign features, and displaying the identified identities.
The radar parameter setting of the embodiment includes two transmitting antennas 13 and four receiving antennas 14, the frequency modulation starting frequency is 77.666GHz, the frequency modulation slope is 96.80009, the frequency modulation period is 300us, the number of frequency modulation periods per frame is 64, the frame period is 31ms, the ADC sampling rate is 3.515MHz, the number of ADC sampling points per frequency modulation period is 128, the maximum measurement distance of the radar determined by these parameters is 5.44m, the maximum measurement speed is 3.14m/s, the distance resolution is 4.25cm, and the speed resolution is 9.83 cm/s.
The structure diagram of the frequency domain interpolation algorithm of this embodiment is shown in fig. 5, and first, a Short Time Fourier Transform (STFT) is performed on a signal, a black dot represents an obtained discrete spectral value, a peak value of the spectral value is found, and three-point secondary spectrum interpolation is performed on the peak value. Suppose that the frequency at which the peak is located is f2Peak value of S2Respectively at f2Left and right at the same distance f1And f3Two points, corresponding to a peak value of S1And S3According to the formula, the frequency f corresponding to the peak value can be foundmax。
The original data signal processing flow of this embodiment is as shown in fig. 6, first performing channel data screening on the acquired 8 channels and 64-chirp original radar data, selecting 1 channel with the least influence of noise, and then performing coherent enhancement processing on the 64-chirp data of this channel, which can effectively improve the signal-to-noise ratio, thereby improving the accuracy of target point detection, then performing target point detection, extracting the phase and phase difference signal of the target point, where the phase difference signal is obtained by subtracting the previous frame from the next frame of the phase signal, then performing deburring and denoising processing on the signal, then obtaining a respiration signal and a heartbeat signal by smooth filtering and band-pass filtering, and finally calculating the respiration rate and the heart rate by using a frequency domain interpolation algorithm and a time domain peak-finding algorithm.
The system work flow chart of the embodiment is shown in fig. 7, and the work flow is as follows:
firstly, starting a radar in a millimeter wave radar subsystem, transmitting a linear frequency modulation continuous wave signal, collecting radar echo reflected on the surface of a static human thorax, performing low-pass filtering after frequency mixing, and obtaining original radar data containing vital sign information and various noises, namely an intermediate frequency signal, after analog-to-digital conversion;
then, the radar data preprocessing subsystem can perform digital signal processing on original radar data, screen out channel signals which are slightly affected by noise, perform distance fast Fourier transform after multi-phase interference enhancement, and extract respiratory waveforms and rate, heartbeat waveforms and rate and vital sign characteristic information of a tested object;
finally, the vital sign feature database subsystem stores the extracted vital sign feature data set; the identity recognition subsystem is used for learning vital sign characteristic information, classifying and judging to determine identity information; the interactive display subsystem realizes real-time updating and displaying of the respiration heartbeat waveform and the speed information of the test object and the identity recognition result of the test object on the display screen.
While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and those skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A vital sign detection and identity recognition system based on a millimeter wave radar is characterized by comprising a millimeter wave radar subsystem (1), a radar data preprocessing subsystem (2), a vital sign feature database subsystem (3), an identity recognition subsystem (4) and an intelligent interactive interface subsystem (5), which are in communication connection with each other;
the millimeter wave radar subsystem (1) is placed beside an object to be detected (6), comprises a transmitting antenna (13) and a receiving antenna (14), and is used for transmitting a linear frequency modulation continuous wave signal, collecting a radar echo acting on the static human thorax surface to be reflected, and obtaining original radar data;
the radar data preprocessing subsystem (2) decomposes a respiratory signal and a heartbeat signal in original radar data and extracts vital sign characteristic information;
the vital sign characteristic database subsystem (3) is used for storing the identity information of a plurality of individuals and the corresponding vital sign characteristic information;
the identity recognition subsystem (4) learns the vital sign characteristic information, classifies and determines corresponding identity information;
the intelligent interactive interface subsystem (5) comprises a display screen and is used for displaying the vital sign information and the identity recognition result of the object to be detected (6) in real time.
2. The system for detecting vital signs and identifying identity based on millimeter wave radar according to claim 1, wherein the transmitting antenna of the millimeter wave radar subsystem (1) transmits a chirp continuous wave signal to a region within the detection range, the receiving antenna receives a radar echo reflected by the surface of the chest of a stationary human body, the received radar echo and the transmitting signal are mixed, low pass filtering is performed, and the raw radar data is obtained after analog-to-digital conversion, wherein the raw radar data contains vital sign information and a plurality of noises.
3. The millimeter wave radar-based vital sign detection and identity recognition system according to claim 1, wherein the radar data preprocessing subsystem (2) decomposes the original radar data signal using a smoothing filter algorithm, a band-pass filter algorithm and/or an improved empirical mode decomposition algorithm to obtain a respiration signal waveform and a heartbeat signal waveform.
4. The millimeter wave radar-based vital sign detection and identity recognition system according to claim 3, wherein the radar data preprocessing subsystem (2) specifically comprises the following steps when decomposing respiratory signals and heartbeat signals in the original radar data:
1) selecting a signal with a certain frequency modulation period in original radar data, and performing fast Fourier transform to obtain target distance distribution;
2) detecting target points according to the target distance distribution, and extracting target point radar data, wherein the target point radar data are data containing thoracic cavity movement;
3) carrying out deburring processing on target point radar data;
4) and decomposing the original radar data signal to obtain a respiration signal waveform and a heartbeat signal waveform.
5. The vital sign detection and identity recognition system based on the millimeter wave radar as claimed in claim 3, wherein the radar data preprocessing subsystem (2) combines a spectrum interpolation method and a signal peak finding algorithm to extract a respiration rate and a heart rate from a respiration signal waveform and a heartbeat signal waveform, and then combines the respiration signal waveform, the heartbeat signal waveform, the respiration rate and the heart rate to extract a plurality of vital sign feature information.
6. The millimeter wave radar-based vital sign detection and identity recognition system of claim 5, wherein the vital sign feature information comprises:
respiratory amplitude, respiratory frequency, respiratory peak-to-peak time interval, respiratory trough-to-trough time interval, peak-to-trough Euclidean distance, peak-to-trough time, peak-to-trough velocity, trough-to-peak distance, trough-to-peak time, trough-to-peak velocity, tidal volume, heartbeat amplitude, and heartbeat frequency.
7. The millimeter wave radar-based vital sign detection and identity recognition system according to claim 3, wherein the smoothing filter algorithm is used for obtaining a respiratory signal close to a real respiratory motion, a Gaussian weighted average value in each window is calculated through a sliding window, a smooth value of the respiratory signal changing along with time is obtained, and the range of the sliding window is set to be 80-120;
the band-pass filtering algorithm is used for separating heartbeat signals from original radar data signals superposed by breathing and heartbeat, an elliptical band-pass filter is adopted in the band-pass filtering algorithm, the frequency band of a pass band is set to be 0.6 Hz-2 Hz, the frequency of a stop band is set to be 0.55-3 Hz, and the maximum allowed attenuation of the pass band and the stop band is respectively set to be 3dB and 40 dB;
the improved empirical mode decomposition algorithm is used for decomposing the respiration heartbeat superposition signal to respectively obtain a respiration signal and a heartbeat signal.
8. The system for detecting and identifying vital signs based on the millimeter wave radar as claimed in claim 5, wherein the spectrum interpolation method comprises the steps of firstly performing short-time Fourier transform on a respiration signal and a heartbeat signal to obtain discrete spectral values, then finding a peak value of the spectral values, performing three-point secondary spectrum interpolation on the peak value to find a frequency corresponding to the peak value to obtain a respiration rate and a heart rate;
the signal peak searching algorithm sets the minimum interval number between wave peaks according to the range of the heart rate and the respiration rate, and then sets the minimum peak value number according to the extracted respiration and heartbeat amplitude values.
9. The millimeter wave radar-based vital sign detection and identity recognition system according to claim 1, wherein the identity recognition subsystem (4) performs offline or online training on a plurality of vital sign feature information of different individuals to generate a classification model, and then performs identity recognition by using the classification model;
after collecting the data of a set number of existing identity information individuals or new individuals, the identity recognition subsystem (4) carries out online updating or replacement of the classification model;
the classification model comprises a convolutional neural network, a long-term and short-term memory network, a nearest node algorithm and/or a k-means algorithm.
10. An identity recognition method of the millimeter wave radar-based vital sign detection and identity recognition system according to any one of claims 1 to 9, comprising the steps of:
s1: the millimeter wave radar transmits linear frequency modulation continuous wave signals to the detection area, radar echoes reflected by the surface of the chest of a static human body are collected, then the received radar echoes and the transmitted waveforms are subjected to low-pass filtering after being subjected to frequency mixing, and original radar data are obtained after analog-to-digital conversion;
s2: carrying out fast Fourier transform on original radar data, filtering static target echoes by utilizing a moving target display algorithm to obtain a sign point phase signal containing human body respiration and heartbeat motion information, then carrying out filtering processing on the phase signal to obtain a respiration signal waveform and a heartbeat signal waveform of a detected object, and then obtaining a respiration rate and a heart rate from the respiration signal waveform and the heartbeat signal waveform;
s3: carrying out time domain and frequency domain data processing on the respiration signal waveform and the heartbeat signal waveform, extracting vital sign characteristic information, and constructing a vital sign characteristic database;
s4: off-line or on-line training is carried out on the vital sign characteristic information of a plurality of individuals to generate a classification model, and classification and identification are carried out on the vital sign characteristic information by utilizing the classification model to obtain identity information;
s5: and updating and displaying the vital sign information of the tested object and the identity information of the tested object on a display screen in real time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110331449.6A CN112998701A (en) | 2021-03-27 | 2021-03-27 | Vital sign detection and identity recognition system and method based on millimeter wave radar |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110331449.6A CN112998701A (en) | 2021-03-27 | 2021-03-27 | Vital sign detection and identity recognition system and method based on millimeter wave radar |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112998701A true CN112998701A (en) | 2021-06-22 |
Family
ID=76408347
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110331449.6A Pending CN112998701A (en) | 2021-03-27 | 2021-03-27 | Vital sign detection and identity recognition system and method based on millimeter wave radar |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112998701A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113545764A (en) * | 2021-07-16 | 2021-10-26 | 中国人民解放军国防科技大学 | FMCW radar personnel identity verification data processing method and system |
| CN113640792A (en) * | 2021-07-06 | 2021-11-12 | 南京理工大学 | Machine learning-based millimeter wave radar detection method for life body in vehicle |
| CN113892931A (en) * | 2021-10-14 | 2022-01-07 | 重庆大学 | Method for extracting and analyzing intra-abdominal pressure by FMCW radar based on deep learning |
| CN114098679A (en) * | 2021-12-30 | 2022-03-01 | 中新国际联合研究院 | Vital sign monitoring waveform recovery method based on deep learning and radio frequency perception |
| CN114533004A (en) * | 2022-01-13 | 2022-05-27 | 山东师范大学 | Vital sign detection method and system based on millimeter wave radar |
| CN114931368A (en) * | 2022-04-08 | 2022-08-23 | 大连海事大学 | Dynamic detection method for vital signs of millimeter wave radar |
| CN115089143A (en) * | 2022-08-24 | 2022-09-23 | 江苏亿连通信技术有限公司 | Millimeter wave radar vital sign signal extraction and measurement method |
| CN115470864A (en) * | 2022-09-28 | 2022-12-13 | 中国人民解放军总医院 | Identity recognition method based on pulse ultra-wideband radar |
| CN115778352A (en) * | 2023-02-02 | 2023-03-14 | 北京华声医学科技开发公司 | Sleep quality evaluation method, device, system and medium based on millimeter wave radar |
| WO2023080018A1 (en) * | 2021-11-04 | 2023-05-11 | オムロン株式会社 | Biological information processing device, biological information processing method, and program |
| CN116131963A (en) * | 2023-02-02 | 2023-05-16 | 广东工业大学 | Fiber link multipath interference noise equalization method based on LSTM neural network |
| CN116236173A (en) * | 2023-03-02 | 2023-06-09 | 康力元(天津)医疗科技有限公司 | Intelligent care monitoring management system based on millimeter wave radar |
| CN116702015A (en) * | 2023-08-09 | 2023-09-05 | 中国科学技术大学 | Signal processing method, device, unmanned aerial vehicle and computer readable storage medium |
-
2021
- 2021-03-27 CN CN202110331449.6A patent/CN112998701A/en active Pending
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113640792A (en) * | 2021-07-06 | 2021-11-12 | 南京理工大学 | Machine learning-based millimeter wave radar detection method for life body in vehicle |
| CN113640792B (en) * | 2021-07-06 | 2024-02-13 | 南京理工大学 | Machine learning-based millimeter wave radar detection method for in-vehicle living body |
| CN113545764B (en) * | 2021-07-16 | 2022-03-25 | 中国人民解放军国防科技大学 | FMCW radar personnel identity verification data processing method and system |
| CN113545764A (en) * | 2021-07-16 | 2021-10-26 | 中国人民解放军国防科技大学 | FMCW radar personnel identity verification data processing method and system |
| CN113892931A (en) * | 2021-10-14 | 2022-01-07 | 重庆大学 | Method for extracting and analyzing intra-abdominal pressure by FMCW radar based on deep learning |
| CN113892931B (en) * | 2021-10-14 | 2023-08-22 | 重庆大学 | Method for extracting and analyzing intra-abdominal pressure by FMCW radar based on deep learning |
| WO2023080018A1 (en) * | 2021-11-04 | 2023-05-11 | オムロン株式会社 | Biological information processing device, biological information processing method, and program |
| CN114098679A (en) * | 2021-12-30 | 2022-03-01 | 中新国际联合研究院 | Vital sign monitoring waveform recovery method based on deep learning and radio frequency perception |
| CN114098679B (en) * | 2021-12-30 | 2024-03-29 | 中新国际联合研究院 | Vital sign monitoring waveform recovery method based on deep learning and radio frequency sensing |
| CN114533004A (en) * | 2022-01-13 | 2022-05-27 | 山东师范大学 | Vital sign detection method and system based on millimeter wave radar |
| CN114931368A (en) * | 2022-04-08 | 2022-08-23 | 大连海事大学 | Dynamic detection method for vital signs of millimeter wave radar |
| CN115089143A (en) * | 2022-08-24 | 2022-09-23 | 江苏亿连通信技术有限公司 | Millimeter wave radar vital sign signal extraction and measurement method |
| CN115470864B (en) * | 2022-09-28 | 2023-05-23 | 中国人民解放军总医院 | Identity recognition method based on pulse ultra-wideband radar |
| CN115470864A (en) * | 2022-09-28 | 2022-12-13 | 中国人民解放军总医院 | Identity recognition method based on pulse ultra-wideband radar |
| CN115778352A (en) * | 2023-02-02 | 2023-03-14 | 北京华声医学科技开发公司 | Sleep quality evaluation method, device, system and medium based on millimeter wave radar |
| CN116131963A (en) * | 2023-02-02 | 2023-05-16 | 广东工业大学 | Fiber link multipath interference noise equalization method based on LSTM neural network |
| CN116236173A (en) * | 2023-03-02 | 2023-06-09 | 康力元(天津)医疗科技有限公司 | Intelligent care monitoring management system based on millimeter wave radar |
| CN116236173B (en) * | 2023-03-02 | 2024-03-26 | 康力元(天津)医疗科技有限公司 | Intelligent care monitoring management system based on millimeter wave radar |
| CN116702015A (en) * | 2023-08-09 | 2023-09-05 | 中国科学技术大学 | Signal processing method, device, unmanned aerial vehicle and computer readable storage medium |
| CN116702015B (en) * | 2023-08-09 | 2023-11-28 | 中国科学技术大学 | Signal processing method, device, unmanned aerial vehicle and computer readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112998701A (en) | Vital sign detection and identity recognition system and method based on millimeter wave radar | |
| CN108416367B (en) | Sleep staging method based on multi-sensor data decision-level fusion | |
| Nenadic et al. | Spike detection using the continuous wavelet transform | |
| CN102499664B (en) | Video-image-based method and system for detecting non-contact vital sign | |
| US20220280087A1 (en) | Visual Perception-Based Emotion Recognition Method | |
| CN108388912A (en) | Sleep stage method based on multisensor feature optimization algorithm | |
| Lee et al. | Smartwatch-based driver vigilance indicator with kernel-fuzzy-C-means-wavelet method | |
| CN113925475B (en) | Non-contact human health monitoring system and monitoring method | |
| CN112754431A (en) | Respiration and heartbeat monitoring system based on millimeter wave radar and lightweight neural network | |
| CN112998668B (en) | Millimeter wave-based non-contact far-field multi-human-body respiration heart rate monitoring method | |
| CN112401856A (en) | Nursing home monitoring method and system based on millimeter wave radar | |
| CN112716474A (en) | Non-contact sleep state monitoring method and system based on biological microwave radar | |
| CN116138745B (en) | Sleep respiration monitoring method and device integrating millimeter wave radar and blood oxygen data | |
| CN115736854A (en) | Respiration and heartbeat monitoring system based on millimeter wave radar | |
| CN107495939B (en) | Living body biological characteristic monitoring method, device and system | |
| CN112998689B (en) | Non-contact multi-person real-time respiration condition evaluation system, device and storage medium | |
| CN114305355B (en) | Breathing heartbeat detection method, system and device based on millimeter wave radar | |
| CN113273978B (en) | Ultra-wideband radar-based human body respiration and heartbeat frequency detection method | |
| Imbeault-Nepton et al. | Filtering data bins of UWB radars for activity recognition with random forest | |
| Wang et al. | Ppg signal reconstruction using deep convolutional generative adversarial network | |
| Wang et al. | HeRe: Heartbeat signal reconstruction for low-power millimeter-wave radar based on deep learning | |
| CN111685760B (en) | Human body respiratory frequency calculation method based on radar measurement | |
| CN106419884B (en) | A kind of rate calculation method and system based on wavelet analysis | |
| CN215820931U (en) | Vital sign detects and identity recognition system based on millimeter wave radar | |
| CN116602640A (en) | Millimeter wave radar heart rate detection method based on adaptive clutter filtering |
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 |