WO2022104868A1 - Millimeter wave radar-based non-contact real-time vital sign monitoring system and method - Google Patents

Abstract

A millimeter wave radar-based non-contact real-time vital sign monitoring system and method. The monitoring system comprises: a millimeter wave transceiving module (1), a real-time signal acquisition module (2), a real-time signal processing module (3), and a visualization module (4). The millimeter wave transceiving module (1) is used for transmitting millimeter waves and receiving echo signals thereof; the real-time signal processing module (3) is used for extracting respiratory signals and heart rate signals by means of clutter suppression; and the visualization module (4) is used for performing waveform fitting on the respiratory signals and heart rate signals, and displaying, by means of a visual interface, the position, respiration, and heart rate of a subject in real time. The clutter suppression effectively overcomes the effect of environmental noise, such that the accuracy of the monitoring system is improved. In addition, the respiratory signals and heart rate signals are fitted by means of an iterative algorithm to visually display relevant waveforms and monitoring results.

Description

A non-contact real-time vital sign monitoring system and method based on millimeter wave radar technical field

The invention belongs to the technical field of radar, and more particularly, relates to a non-contact real-time vital sign monitoring system and method based on millimeter wave radar.

Background technique

Non-contact vital signs monitoring mostly uses technologies such as WiFi, ultra-wideband radar, and millimeter-wave radar. By monitoring the periodic changes of human body reflected signals, human vital signs (such as respiration and heart rate, etc.) are monitored. Among them, non-contact vital sign monitoring based on millimeter-wave radar has the advantages of low transmission power, low cost, and high precision, and has gradually become a hot research field in academia and industry in recent years.

The non-contact vital signs monitoring system based on millimeter-wave radar, by transmitting and receiving millimeter waves, detects the periodic small displacement of the thoracic cavity at mm level caused by the human body due to breathing (inhalation/exhalation) and heart beating. Heart rate is monitored. Specifically, after the monitoring system emits low-power millimeter waves, it measures the time it takes for the signal to bounce back from the body. Taking breathing as an example, when the human body inhales, its chest expands and moves forward, reducing the reflex time; conversely, it increases the reflex time. Through the measurement and analysis of the periodic changes of the thoracic cavity, the waveform and frequency of respiration and heart rate can be extracted to realize the monitoring of human vital signs.

But the existing technology still has many deficiencies. Mainly reflected in: first, the existing system is easily disturbed by various noises from the environment, and the stability is poor; secondly, most of the existing systems have certain defects, or cannot export and visualize the breathing and heart rate data in real time, Or the monitoring results of relevant vital signs cannot be displayed in real time and accurately.

Therefore, building a non-contact vital signs monitoring system with good real-time performance, strong robustness and high accuracy has become an urgent problem to be solved in the art.

SUMMARY OF THE INVENTION

In view of the defects of the prior art, the purpose of the present invention is to provide a non-contact real-time vital sign monitoring system and method based on millimeter wave radar, aiming to solve the problem that the prior art is susceptible to environmental noise due to lack of clutter suppression. Interference, resulting in inaccurate vital sign monitoring and poor visualization.

The invention provides a non-contact real-time vital sign monitoring system based on millimeter wave radar, which includes: a millimeter wave transceiver module, a real-time signal acquisition module, a real-time signal processing module and a visualization module; the input end of the real-time signal acquisition module is connected to the millimeter wave The output end of the wave transceiver module, the first output end of the real-time signal acquisition module is connected to the input end of the millimeter wave transceiver module, the second output end of the real-time signal acquisition module is connected to the input end of the real-time signal processing module, the input end of the visualization module Connect to the output end of the real-time signal processing module; the millimeter wave transceiver module is used to transmit millimeter waves and receive their echo signals; the real-time signal acquisition module is used to collect millimeter wave signals in real time and package and output the echo signals; the real-time signal processing module is used for It is used to extract the breathing signal and the heart rate signal through clutter suppression; the visualization module is used to perform waveform fitting on the breathing signal and the heart rate signal, and display the subject's position, breathing and heart rate in real time through a visual interface.

Further, the real-time signal acquisition module includes: a parameter adjustment unit, used to determine the detection range of the millimeter-wave radar; a data capture unit, used to monitor the UDP port and capture UDP data packets in real time; a data transmission unit, used to periodically Added data splicing, packaging and transmission; sliding window setting unit for setting sliding windows.

距离分辨率为

其中，c为光速3×10 ⁸m/s，F _s为线性调频信号上样本点的采样频率，K _slope为线性调频信号的斜率，B为线性调频信号的扫频带宽。 Further, the maximum monitoring distance in the detection range is

The distance resolution is

where c is the speed of light 3×10 ⁸ m/s, F _s is the sampling frequency of the sample point on the chirp signal, K _slope is the slope of the chirp signal, and B is the swept bandwidth of the chirp signal.

Furthermore, the real-time signal processing module includes: a clutter suppression unit, used for processing data in the window, to suppress clutter interference and to perform echo selection; a bandpass filter unit, used to realize the respiration signal and the echo selection through bandpass filtering. Extraction of heart rate signal; vital sign calculation unit, used for extracting breathing frequency and heartbeat frequency through frequency domain analysis method.

Further, the clutter suppression unit includes: a first unit for suppressing stationary noise within the millimeter wave measurement range; and a second unit for suppressing non-stationary noise within the millimeter wave measurement range.

Further, the vital sign calculation unit includes: a breathing frequency calculation unit, used for searching for the peak of the breathing signal through a time-domain peak-finding algorithm, and obtaining the breathing frequency by calculating the frequency of the peak; a heartbeat frequency calculation unit, used for adopting a modified The periodogram power spectral density estimation method calculates the power spectral density of the sequence to estimate the heartbeat frequency, and uses fitting to eliminate the frequency domain offset to further optimize and calibrate the heartbeat frequency.

The present invention also provides a non-contact real-time vital sign monitoring method based on millimeter wave radar, comprising the following steps:

transmitting millimeter waves and receiving echo signals reflected by the millimeter waves;

Perform clutter suppression and echo selection on the echo signal to extract the breathing signal and the heart rate signal;

By fitting the breathing signal and the heart rate signal, the real-time visual display of the signal and the detection result is realized.

Further, the clutter suppression of the echo signal is specifically: suppressing stationary noise in the millimeter wave measurement range through adaptive background subtraction by weighting coefficient fitting; realizing non-stationary noise in the millimeter wave measurement range through singular value decomposition. Noise suppression.

Further, the echo selection is specifically: selecting signals related to vital signs from the echo signals.

Further, the breathing signal and the heart rate signal are extracted by means of band-pass filtering, the peak of the breathing signal is searched by a time-domain peak-finding algorithm, and the breathing frequency is obtained by calculating the frequency of the peak; a modified periodogram is used. The power spectral density estimation method calculates the power spectral density of the sequence to estimate the heartbeat frequency, and uses fitting to eliminate the frequency domain offset to further optimize and calibrate the heartbeat frequency.

Compared with the prior art, the present invention has the following obvious outstanding features and remarkable technological progress:

(1) The present invention effectively overcomes the influence of environmental noise through clutter suppression, and improves the precision of the monitoring system.

(2) The present invention realizes real-time collection, real-time analysis and visualization of data; specifically, real-time capture of UDP data through the Socket module and back to the host computer to realize real-time data collection; through time-frequency domain analysis, the realization of human life Real-time detection of physical signs; fit the breathing signal and heart rate signal through an iterative algorithm, and visualize the relevant waveforms and monitoring results.

Description of drawings

FIG. 1 is a schematic block diagram of a non-contact real-time vital sign monitoring system based on a millimeter wave radar provided by an embodiment of the present invention.

FIG. 2 is an implementation flowchart of a non-contact real-time vital sign monitoring method based on a millimeter wave radar according to an embodiment of the present invention.

FIG. 3 is a schematic structural diagram of a millimeter wave transceiver module in a non-contact real-time vital sign monitoring system based on a millimeter wave radar according to an embodiment of the present invention.

FIG. 4 is a schematic diagram of a transmit signal and an echo signal of an FMCW radar provided by an embodiment of the present invention.

FIG. 5 is a block diagram of a sliding window provided by an embodiment of the present invention.

6 is the experimental result of the system and method provided by the embodiment of the present invention, wherein (a) is a schematic diagram of the respiration, heart rate and ECG control waveforms of the first subject, (b) is the respiration, Schematic diagram of heart rate and ECG control waveforms, (c) is a schematic diagram of the third subject's respiration, heart rate and ECG control waveforms.

7 is a visualization interface after fitting provided by an embodiment of the present invention, wherein (a) is a schematic diagram of the real-time detection results of the original waveform and the fitted waveform of the subject's respiration and the respiratory frequency and the distance of the subject, (b) ) are the original waveform and the fitted waveform of the subject's heart rate, as well as the schematic diagram of the real-time detection result of the heartbeat frequency.

Detailed ways

In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention.

The invention effectively eliminates stationary noise and non-stationary noise from the environment through clutter suppression on the basis of real-time signal acquisition; dynamically tracks the location of the subject through echo selection, and improves the accuracy of vital sign detection; The optimization effectively eliminates the frequency domain offset of the heart rate signal, and realizes the accurate estimation of the heart rate; it performs high-quality fitting and reconstruction of the heart rate and respiration signals in real time through an iterative algorithm.

FIG. 1 shows a principle block diagram of a non-contact real-time vital sign monitoring system based on a millimeter-wave radar provided by an embodiment of the present invention. For the convenience of description, only the part related to the embodiment of the present invention is shown, and the details are as follows:

The real-time, stable and high-precision non-contact vital signs monitoring system provided by the present invention includes: a millimeter wave transceiver module 1, a real-time signal acquisition module 2, a real-time signal processing module 3 and a visualization module 4; wherein, the real-time signal acquisition module The input end of 2 is connected to the output end of the millimeter wave transceiver module 1, the first output end of the real-time signal acquisition module 2 is connected to the input end of the millimeter wave transceiver module 1, and the second output end of the real-time signal acquisition module 2 is connected to the real-time signal The input end of the processing module 3, the input end of the visualization module 4 are connected to the output end of the real-time signal processing module 3, the millimeter wave transceiver module 1 transmits millimeter waves and receives its echo signal, and its echo signal is packaged by the real-time signal acquisition module 2 It is transmitted to the host computer in real time, then processed and analyzed by the real-time signal processing module 3, and finally the waveform fitting and visual display are performed by the visualization module 4.

The millimeter wave transceiver module 1 is mainly used to transmit millimeter waves and receive millimeter wave echo signals. It mainly includes three parts: millimeter-wave radar transceiver, high-precision AD conversion, and digital signal processing. Among them, the transmitting and receiving part of the millimeter wave radar adopts MIMO antenna technology, including two transmitting antennas Tx1, Tx2 and four receiving antennas Rx1, Rx2, Rx3, and Rx4, which are respectively composed of parallel microstrip antennas. Each transmit antenna has independent phase and amplitude control, and can transmit chirps from 77GHz to 81GHz; each receive antenna can work individually or simultaneously. The high-precision AD conversion part performs 16-bit high-precision analog-to-digital conversion on the signal received by the receiving antenna. The digital signal processing part uses FPGA or DSP to preprocess the echo signal.

The real-time signal acquisition module 2 is used to set millimeter wave parameters, and perform real-time data capture, storage and splicing operations. First adjust the parameters to determine the detection range of the millimeter-wave radar; then monitor the UDP port through the Socket module, capture UDP data packets in real time and save the original data on the host computer; It is transmitted to the real-time signal processing module; finally, the sliding window is set to prepare for subsequent signal processing.

The real-time signal processing module 3 is used to extract the breathing signal and the heart rate signal; firstly, the data in the sliding window is preprocessed to suppress clutter interference and echo selection; then the band-pass filtering operation is performed to extract the breathing signal and the heart rate signal; The improved frequency domain analysis method, on the basis of power spectral density estimation, eliminates the offset of the signal frequency domain by fitting, and extracts the heartbeat frequency.

The visualization module 4 is used to realize the visualization of real-time data and its monitoring results. First, the iterative algorithm is used to fit the breathing signal and the heart rate signal, and then save it in real time; then the (i) position, (ii) the breathing and heart rate waveform, (iii) the breathing frequency and the heartbeat frequency of the subject are displayed in real time through the visual interface.

The invention can solve the problems of clutter suppression and signal fitting. First, the echo signals of millimeter waves may include various clutter interference, such as stationary noise from static objects (reflected signals) such as desks and walls, non-stationary noises from moving objects (reflection signals), etc., which may affect vital signs. Monitoring, especially heart rate monitoring, causes great trouble. In traditional clutter suppression, the environmental noise in a specific environment is generally collected first, and then the environmental noise is subtracted from the millimeter wave echo signal to eliminate static clutter interference. However, this method is not universal, cannot adapt to changes in the environment, and has poor clutter suppression effect. In order to adaptively and effectively suppress clutter, the present invention designs adaptive filtering algorithms for stationary noise and non-stationary noise based on the analysis of environmental noise to suppress various noises from the environment. Secondly, due to vital sign monitoring, especially heart rate monitoring, it is necessary to extract weak mm-level useful signals from strong noise and strong interference, which requires extremely high back-end signal processing algorithms. Traditional signal reconstruction/fitting either uses wavelet transform and other methods for simple band-pass filtering, or uses EEMD (Ensemble Empirical Mode Decomposition) and other methods for reconstruction. But in general, the filtering effect of the former is poor, and the selection of the latter modal components is difficult to achieve universality and adaptability, that is, for subjects with low or high heart rate, it is difficult for EEMD to adapt. A suitable modal component or combination of modal components is selected. Based on the analysis and comparison of many traditional algorithms, the invention proposes a waveform fitting method based on an iterative algorithm, and performs high-quality fitting and reconstruction on the heart rate and breathing signals in real time.

Fig. 2 shows the implementation flow chart of the non-contact real-time vital sign monitoring method based on the millimeter wave radar provided by the embodiment of the present invention, which is described in detail as follows:

The non-contact real-time vital sign monitoring method based on the millimeter wave radar provided by the embodiment of the present invention includes the following steps:

(1) The millimeter wave transceiver module realizes millimeter wave transceiver, high-precision AD conversion and digital signal processing operations.

The specific processing process is as follows: the present invention adopts FMCW (Frequency Modulated Continuous Wave, frequency modulated continuous pulse) millimeter-wave radar, and its transceiver module is shown in Figure 3, and the transceiver module includes 2 transmit antennas Tx and 4 receive antennas Rx. The millimeter-wave radar transmit and echo signals are shown in Figure 4.

式中A _Tx是传输信号的幅值，K _slope为锯齿波斜率，f _c为雷达发射信号中心频率，

为发射信号的初始相位。 (1-1) Transmitter. As shown in Figure 3, the oscillator provides a reference signal to generate a fundamental frequency signal after a phase-locked loop, and a linear frequency modulation signal (Chirp) is obtained after a frequency multiplier, which is sent by a transmitting antenna after a power amplifier. The carrier of the transmitted signal is a sawtooth wave, as shown in FIG. 4 , its period is T _f , the frequency modulation bandwidth is B; the frame period (ie, the repetition period of the sawtooth wave) is T _i . The expression of the transmitted signal is: S _Tx (t) =

where A _Tx is the amplitude of the transmitted signal, K _slope is the slope of the sawtooth wave, f _c is the center frequency of the radar transmit signal,

is the initial phase of the transmitted signal.

(1-2) Receiver. As shown in FIG. 3 , the receiving antenna R _X receives the millimeter wave echo signal (ie, the signal reflected by the millimeter wave radar after irradiating an object such as a human body) and processes it. The specific process is as follows: after the echo signal passes through the low-noise amplifier, it is orthogonally mixed with the original transmit signal, the high-frequency signal is filtered out by the low-pass filter to obtain the intermediate frequency signal, and the digital signal is obtained by AD conversion after the intermediate frequency amplifier, and then the The microcontroller (main control unit) sends the relevant digital signals to a high-precision FPGA or DSP module for preprocessing. This process can be described by the following formula,

(1-2-1) The expression of the echo signal is: S _Rx (t)=A _Rx S _Tx (t-τ)...(2-1);

式中c为光速，d ₀为毫米波雷达到目标胸腔运动的中心距离，A _rsin(2πf _rt)和A _hsin(2πf _ht)分别为呼吸和心跳导致的胸腔位移，A _r和A _h分别为人体呼吸和心跳造成的胸腔最大位移，f _r和f _h为呼吸频率和心跳频率。 where A _Rx is the echo signal amplitude, A _Rx is inversely proportional to the distance from the millimeter-wave radar to the target; τ is the delay of the radar echo signal,

where c is the speed of light, d ₀ is the distance from the millimeter-wave radar to the center of the thoracic motion of the target, _{Ar sin(2πf r t} ) and A _h sin(2πf _h t) are the thoracic displacement caused by breathing and heartbeat, respectively, _Ar and A _h is the maximum displacement of the thoracic cavity caused by human respiration and heartbeat, respectively, and fr and _fh are the respiratory _rate and the heartbeat frequency.

经过低通滤波后即可得到该式中的中频信号，其频率f _IF＝2πK _slopeτ，相位

(1-2-2) The echo signal is mixed with the original transmitted signal,

After low-pass filtering, the intermediate frequency signal in this formula can be obtained, and its frequency f _IF =2πK _slope τ, phase

(2) Set radar parameters through the real-time signal acquisition module, perform real-time capture, storage and splicing operations of millimeter wave echo signals, and set up sliding windows to prepare for back-end signal processing. The specific processing process is as follows:

其中，c为光速3×10 ⁸m/s，F _s为Chirp上样本点的采样频率，K _slope为Chirp的斜率，B为Chirp的扫频带宽，见图4。例如：若F _s＝2×10 ⁶sps，K _slope＝70MHz/us，B＝4GHz(其中sps为每秒采样次数)，则d _max＝4.29m，d _res＝3.76cm。 (2-1) Radar parameter setting: Millimeter waves are non-contact and non-interfering, and can pass through materials such as plastic, drywall, and clothing, and can measure a wide range. Its maximum detection distance d _max and distance resolution d _res can be flexibly adjusted by modifying the relevant parameters of the millimeter wave radar,

Among them, c is the speed of light 3×10 ⁸ m/s, F _s is the sampling frequency of the sample points on Chirp, K _slope is the slope of Chirp, and B is the frequency sweep bandwidth of Chirp, as shown in Figure 4. For example: if F _s =2×10 ⁶ sps, K _slope =70MHz/us, and B=4GHz (where sps is the number of samples per second), then d _max =4.29m, d _res =3.76cm.

(2-2) Data capture, storage and splicing. Monitor the UDP port through the Socket module, capture UDP data packets in real time and save the original data in the host computer. Next, a timer is set, and when Δt=t _step (for example: Δt=1s), the newly added data is spliced and packaged and transmitted to the real-time signal processing module.

(2-3) Sliding window. Set the sliding window, the window length and step size are t _window and t _step , respectively. For example: when t _window =30s and t _step =1s, the block diagram is shown in Figure 5 .

(3) Through the real-time signal processing module, perform clutter suppression, echo selection, and band-pass filtering operations on the data in the sliding window, extract respiratory signals and heart rate signals, and calculate their frequencies.

The specific processing process is as follows:

Q表示原始回波信号、滤除平稳杂波后的回波信号和滤除非平稳杂波后的回波信号，三者均为N×M的矩阵，则滤波过程可表示如下： (3-1) Clutter suppression. The echo signals of millimeter waves may include various clutter interferences, such as stationary noise from static objects (reflected signals) such as desks and walls, and non-stationary noises from moving objects (reflection signals). The center frequency of the clutter power spectrum is close to zero frequency, which is close to the thoracic vibration frequency caused by human respiration and heartbeat (breathing 0.1Hz~0.6Hz, heart rate 0.8Hz~2Hz), and the two are easily aliased, which will cause damage to the monitoring of human vital signs. Great interference, so the suppression of clutter signals is essential. If used separately

Q represents the original echo signal, the echo signal after filtering out stationary clutter, and the echo signal after filtering out non-stationary clutter, all of which are N×M matrices, the filtering process can be expressed as follows:

其中

和B _n(m)分别表征t _n时刻的原始回波信号、滤除平稳杂波后的回波信号和背景噪声估计，三者均为M×1的一维向量。其中，

式中，λ为加权系数，λ∈[0，1]，例如λ＝0.95。 (3-1-1) Suppression of stationary clutter. The invention adopts adaptive background subtraction based on weighting coefficient fitting to suppress stationary noise within the millimeter wave measurement range. As shown in Figure 4, at a slow time t _n , the filtering of stationary clutter can be expressed as:

in

and B _n (m) represent the original echo signal at time t _n , the echo signal after filtering out stationary clutter, and the estimated background noise, and all three are M×1 one-dimensional vectors. in,

In the formula, λ is a weighting coefficient, λ∈[0, 1], for example, λ=0.95.

分解为正交矩阵，

其中U和V分别是N×N和M×M阶酉矩阵，H表示共轭转置；Σ是N×M阶对角阵，包含了矩阵

的奇异值。接着，将对角阵Σ中除最大奇异值λ _max之外的所有奇异值置0，得到对角阵

最后，信号重构，

式中Q即为滤除非平稳杂波后的信号。 (3-1-2) Suppression of non-stationary clutter. The present invention adopts singular value decomposition (SVD) to suppress non-stationary noise in the millimeter wave measurement range. First, the signal matrix

Decomposed into orthogonal matrices,

where U and V are N×N and M×M order unitary matrices, respectively, H represents the conjugate transpose; Σ is an N×M order diagonal matrix, including the matrix

singular value of . Next, set all singular values except the largest singular value λ _max in the diagonal matrix Σ to 0 to obtain a diagonal matrix

Finally, the signal reconstruction,

where Q is the signal after filtering out non-stationary clutter.

(3-2) Echo selection. The distance of the subject is precisely positioned, and the echo signal of the distance unit is selected, and the signal contains the periodic change information of the chest cavity caused by the subject's breathing and heartbeat.

其中，m∈[1，M]。第三，找出最大能量和max(E(m))所在的列，将其列索引记为m _max，该列所表征的距离单元即为受试者的测试距离。第四，从矩阵Q中提取其第m _max列信号，计算相位并执行相位解缠操作，将其结果记为序列x(n),n∈[1，N]。 Echo selection refers to the extraction of raw signals related to the subject's vital signs from ambient noise while accurately locating the subject's distance unit. The specific operations are as follows: First, one-dimensional FFT is performed on each row of the matrix Q to obtain an N×M distance matrix R. Each column of matrix R represents a distance unit, as shown in Figure 4. For example: the distance unit represented by the mth column is m×d _res , and d _res is the distance resolution (for example: d _res =3.76cm), see formula (4). Second, calculate the sum of energies over each distance cell,

where m∈[1,M]. Third, find the column where the maximum energy sum max(E(m)) is located, and record its column index as m _max , and the distance unit represented by this column is the test distance of the subject. Fourth, extract its m _max column signal from the matrix Q, calculate the phase and perform the phase unwrapping operation, and record the result as the sequence x(n),n∈[1,N].

其中呼吸和心率的通带[f _L，f _H]分别为：[0.1-0.6]Hz和[0.8-2.5]Hz。最后通过逆小波变换分别提取心率信号和呼吸信号。该过程可采用如下面公式进行描述：

(3-3) Bandpass filtering. Design a bandpass filter to extract the respiration signal and the heart rate signal respectively. Band-pass filtering is performed using wavelet transform, which specifically includes: firstly calculating the wavelet transform of x(n), DWT{x(n)}, and then performing band-pass filtering in the wavelet domain,

where the passbands [f _L , f _H ] for respiration and heart rate are: [0.1-0.6]Hz and [0.8-2.5]Hz, respectively. Finally, the heart rate signal and respiration signal are extracted by inverse wavelet transform. This process can be described by the following formula:

(3-4) Calculation of vital signs. Process the sequence x(n) in the sliding window, (including the filtered respiration signal sequence x _br (n) and the heart rate signal sequence x _hr (n)), and calculate the respiratory rate and heartbeat frequency respectively; as shown in Figure 2 :

(3-4-1) Calculation of respiratory rate:

The time-domain peak-finding algorithm findpeaks is used to find the peaks of the respiratory signal x _br (n) and calculate its frequency.

(3-4-2) Heartbeat frequency calculation:

Welch (modified periodogram power spectral density estimation method) was used to calculate the power spectral density of the sequence x _hr (n), and the frequency of the heartbeat was initially estimated. The improved frequency domain analysis algorithm was used to calibrate and optimize the heartbeat frequency.

其中，FFT{x _hr(n)}为序列x _hr(n)的Fourier变换，n∈[1，N]。为减小误差，本发明采用分段平均法对功率谱密度P _hr进行平滑，即：将序列x _hr(n)分成L个小段，每小段含W个采样点，对每小段信号分别进行功率谱估计后，取其平均值作为整个序列x _hr(n)的功率谱估计。其中，N≤L×W，若L个小段互不重叠，则N＝L×W。功率谱密度最大值P _max所对应的频率，即为初步估计的心跳频率f _hr0。 Specifically include: (3-4-2-1) Spectrum estimation: The power spectral density of the sequence x _hr (n) can be calculated by the following formula:

Among them, FFT{x _hr (n)} is the Fourier transform of the sequence x _hr (n), n∈[1,N]. In order to reduce the error, the present invention adopts the piecewise averaging method to smooth the power spectral density P _hr , that is: divide the sequence x _hr (n) into L small sections, each small section contains W sampling points, and the power spectral density of each small section signal is carried out respectively. After spectrum estimation, take its average as the power spectrum estimate for the entire sequence x _hr (n). Among them, N≤L×W, if the L small segments do not overlap each other, then N=L×W. The frequency corresponding to the maximum value P _max of the power spectral density is the preliminarily estimated heartbeat frequency f _hr0 .

该方法可精确地消除信号频域的偏移，对心跳频率进行校准优化。 (3-4-2-1) Calibration optimization: first perform FFT processing on x _hr (n), and use findpeaks to find its frequency domain peak; select the peak closest to f _hr0 value as the frequency domain main peak; select the frequency domain peak The main peak of the domain and its two peaks on the left and right sides are used as useful data; other data are set to 0; IFFT transformation is performed to obtain the reconstructed time domain signal and its slope kx _slope is calculated. Then the heart rate per minute can be obtained by the following formula:

This method can accurately eliminate the offset of the signal frequency domain, and calibrate and optimize the heartbeat frequency.

(4) By fitting the breathing signal and the heart rate signal, the real-time and visual display of the signal and the detection result is realized.

The specific processing process is as follows:

(4-1) Iterative fitting:

其中，b ₁～b ₄为参数，在迭代算法中进行拟合。b ₁，b ₂，b ₃，b ₄分别为类正弦函数的振幅因子、平移因子、缩放因子、偏移因子。 Perform sine-like function fitting on the respiratory signal and heart rate signal sequence (iex _br (n) and x _hr (n)), and save it to the host computer in real time,

Among them, b ₁ to b ₄ are parameters, which are fitted in an iterative algorithm. b ₁ , b ₂ , b ₃ , and b ₄ are the amplitude factor, translation factor, scaling factor, and offset factor of the sine-like function, respectively.

(4-2) Visualization. Through the visual interface, the position of the subject, the waveform of the subject's respiration and heart rate, the subject's respiration rate and heartbeat rate are displayed in real time. In order to verify the reliability of the vital sign monitoring system and method proposed in this embodiment, a plurality of subjects are recruited in this embodiment to participate in a test with a duration of 100 seconds. Three representative subjects were selected, and the heart rates of the three subjects were located in three different frequency ranges: high, medium, and low, respectively. Their heart rate and respiratory waveforms are shown in Figure 6, (a), (b) ) and (c) show the visualization interface of the waveforms during the testing of the three subjects, respectively. The test results are shown in Table 1. It can be seen from Table 1 that the test results of this embodiment are highly consistent with the test results of the wearable ECG device (collected synchronously during the experiment) no matter in which interval. In this embodiment, the subject's distance, respiration and heart rate waveforms are updated and displayed every second, thereby realizing real-time detection and accurate measurement of vital signs. The visualization results are shown in Figure 7. (a) and (b) show the fitted heart rate and respiration signals updated every minute, as well as real-time detection results, including: the subject's position, the number of heartbeats per minute and breathing rate. It can be seen that this embodiment realizes real-time monitoring and accurate measurement of vital signs.

Table 1. Results of vital signs monitoring

Those skilled in the art can easily understand that the above are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention, etc., All should be included within the protection scope of the present invention.

Claims (10)

A non-contact real-time vital sign monitoring system based on millimeter-wave radar, characterized in that it comprises: a millimeter-wave transceiver module (1), a real-time signal acquisition module (2), a real-time signal processing module (3) and a visualization module (4) ); The input end of the real-time signal acquisition module (2) is connected to the output end of the millimeter wave transceiver module (1), and the first output end of the real-time signal acquisition module (2) is connected to the millimeter wave transceiver module ( 1), the second output of the real-time signal acquisition module (2) is connected to the input of the real-time signal processing module (3), and the input of the visualization module (4) is connected to the real-time signal processing module (3). an output end of the signal processing module (3); The millimeter wave transceiver module (1) is used for transmitting millimeter waves and receiving echo signals thereof; The real-time signal acquisition module (2) is used to collect millimeter wave signals in real time and package and output the echo signals; The real-time signal processing module (3) is used to extract the breathing signal and the heart rate signal through clutter suppression; The visualization module (4) is used to perform waveform fitting on the respiration signal and the heart rate signal, and display the subject's position, respiration and heart rate in real time through a visualization interface. The non-contact real-time vital sign monitoring system according to claim 1, wherein the real-time signal acquisition module (2) comprises: The parameter adjustment unit is used to determine the detection range of the millimeter wave radar; Data capture unit, used to monitor UDP ports and capture UDP packets in real time; A data transmission unit, which is used to splicing, packaging and transmitting the newly added data periodically; The sliding window setting unit is used to set the sliding window. The non-contact real-time vital sign monitoring system according to claim 2, wherein the maximum monitoring distance in the detection range is

The distance resolution is

where c is the speed of light 3×10 ⁸ m/s, F _s is the sampling frequency of the sample point on the chirp signal, K _slope is the slope of the chirp signal, and B is the swept bandwidth of the chirp signal. The non-contact real-time vital sign monitoring system according to any one of claims 1-3, wherein the real-time signal processing module (3) comprises: The clutter suppression unit is used to process the data in the window to suppress clutter interference and perform echo selection; Band-pass filtering unit, used for extracting breathing signal and heart rate signal through band-pass filtering; The vital sign calculation unit is used to realize the extraction of the breathing frequency and the heartbeat frequency through the time-frequency domain analysis method. The non-contact real-time vital sign monitoring system according to claim 4, wherein the clutter suppression unit comprises: The first unit is used to suppress stationary noise in the millimeter wave measurement range; The second unit is used to suppress non-stationary noise in the millimeter wave measurement range. The non-contact real-time vital sign monitoring system according to claim 4 or 5, wherein the vital sign calculating unit comprises: a respiratory frequency calculation unit, used for finding the peak of the respiratory signal through a time-domain peak-finding algorithm, and obtaining the respiratory frequency by calculating the frequency of the peak; The heartbeat frequency calculation unit is used for calculating the power spectral density of the sequence by using the modified periodogram power spectral density estimation method to estimate the heartbeat frequency, and using fitting to eliminate the frequency domain offset to further optimize and calibrate the heartbeat frequency. A non-contact real-time vital sign monitoring method based on millimeter-wave radar, characterized in that it comprises the following steps: transmitting millimeter waves and receiving echo signals reflected by the millimeter waves; Perform clutter suppression and echo selection on the echo signal to extract the breathing signal and the heart rate signal; By fitting the breathing signal and the heart rate signal, the real-time visual display of the signal and the detection result is realized. The non-contact real-time vital sign monitoring method according to claim 7, wherein the clutter suppression on the echo signal is specifically: The suppression of stationary noise in the millimeter-wave measurement range is achieved by adaptive background subtraction fitted with weighting coefficients; The non-stationary noise in the millimeter wave measurement range is suppressed by singular value decomposition. The non-contact real-time vital sign monitoring method according to claim 7 or 8, wherein the echo selection is specifically: selecting a signal related to the vital sign from the echo signals. The non-contact real-time vital sign monitoring method according to any one of claims 7 to 9, wherein the breathing signal and the heart rate signal are extracted by means of bandpass filtering, and a time-domain peak finding algorithm is used to find the The peak of the respiratory signal is obtained by calculating the frequency of the peak to obtain the respiratory frequency; the modified periodogram power spectral density estimation method is used to calculate the power spectral density of the sequence to estimate the heartbeat frequency, and the frequency domain offset is eliminated by fitting. Further Picture Control.

Images