Transmission-type human body heart rate detection device and method based on UHF radio waves

Abstract

The invention discloses a transmission-type human body heart rate detection device and method based on UHF radio waves, and relates to the technical field of wireless equipment. Comprising a transmitter module, a transmitting antenna, a receiver module, a receiving antenna and a signal processing upper computer, a radio frequency interface of the transmitter module is connected with the transmitting antenna, a radio frequency interface of the receiver module is connected with the receiving antenna, and a human body of a tester is located on an LOS link between the transmitting antenna and the receiving antenna. And data interfaces of the transmitter module and the receiver module are respectively connected with a signal processing upper computer. The transmitter module and the receiver module both adopt ADALM-PLUTO SDR equipment of the ADI company, and the transmitting antenna and the receiving antenna both adopt flat directional antennas. According to the invention, the problems of low sensitivity, weak penetrability, high signal processing complexity, weak anti-interference capability and the like of a reflection technology are effectively solved by using the penetration characteristic of low-frequency electromagnetic waves and combining with a processing method for filtering and reconstructing transmitted wave signals.

Description

Technical Field

[0001] The invention relates to the technical field of wireless equipment, and in particular to a transmission-type human heart rate detection device based on UHF radio waves. Background Art

[0002] Human physiological data includes a variety of indicators, such as blood pressure, pulse oxygen saturation, pulse rate, body temperature, respiration and heart rate. Among these indicators, the heartbeat signal is considered to be the most important indicator of physical health. An abnormal heart rate reflects that the human body may be in a sub-healthy or unhealthy state. Heart rate detection equipment is divided into contact and non-contact types. Contact devices usually need to contact the human body and use sensors or electrodes to achieve detection. They have high precision and mature technology and are widely used in the medical field. However, contact devices may not be suitable for some users, such as injured patients, and are complex to deploy and expensive, making them unsuitable for large-scale popularization.

[0003] With the advancement of technology, contactless vital sign monitoring has become important. Radio waves are a type of electromagnetic wave that covers a wide range of frequencies in the environment, ranging from thousands of hertz to hundreds of gigahertz. Radio waves can not only be used to transmit information, but also to sense the human body and the physical world around it. Using radio wave signals to detect human heart rate has become a research focus in recent years. Existing research on the use of radio waves for human heart rate includes three technical solutions: millimeter wave radar, WiFi sensing, and radio frequency identification technology.

[0004] Existing technical solutions are all based on the "electromagnetic reflection" principle, that is, using electromagnetic wave radio frequency signals to measure the reflection fluctuations caused by the tiny movements of the human chest surface during the contraction and relaxation of the heart to detect the heart rate. First, the system transmitter will emit electromagnetic waves to the human body. After the electromagnetic waves come into contact with the human body, part of them is absorbed and part is reflected. The receiver antenna receives the radio wave signals reflected by the human body and environmental objects. Then, by processing the original signal, analyzing the frequency and amplitude changes of the heartbeat signal, filtering out the noise and extracting the heartbeat features. Finally, the heartbeat signal features are sent to the heart rate estimation mathematical model to obtain the heart rate value.

[0005] In the prior art,

[0006] The patent document "Non-invasive respiration and heartbeat detection implementation method based on CSI signal strength in WiFi" (application number 201510593528.9, application publication number CN 105232022 A) applied by Taiyuan University of Technology uses WiFi signals to detect human respiration and heartbeat signals. General WiFi receiving devices can obtain information on one of the subcarriers in the form of channel state information (CSI). The fluctuation of the chest during human respiration and heartbeat will affect CSI. CSI can measure the phase information and amplitude information of each subcarrier. Fast Fourier transform is performed on CSI to extract the frequency domain signals of respiration and heartbeat. The frequency domain signals of respiration and heartbeat are separated by bandpass filters respectively. Then, inverse Fourier transform is performed to extract the time domain signals of respiration and heartbeat, and smoothing is performed to achieve data visualization.

[0007] The patent document applied for by Chang'an University, "A non-contact heart rate detection method and device based on millimeter-wave radar" (application number 202310594406.6, application publication number CN116616736A), uses millimeter-wave radar to realize non-contact heart rate health detection of the human body. First, the radar echo signal reflected by the human target is obtained, the phase of the human target is determined, and the target phase is phase unwrapped and phase differentiated to obtain the original signal, and then filtered to obtain the initial heartbeat signal, and the heartbeat signal is preprocessed to obtain a noisy heartbeat signal, and then processed using the VMD model based on the whale optimization algorithm to obtain a denoised heartbeat signal, and the denoised heartbeat signal is feature extracted using an adaptive threshold function and a sliding window to obtain the final signal peak, and the number of the final signal peaks is determined, and the heart rate is calculated based on the number of the final signal peaks.

[0008] In the paper "RF-ECG: Heart Rate Variability Assessment Based on COTS RFIDTag Array", a new heart rate variability (HRV) monitoring method is introduced, namely radio frequency identification (RFID) tag array technology. This technology realizes HRV monitoring through RF-ECG. Compared with the traditional ECG method, RF-ECG has the advantages of small size, light weight, easy portability and simple operation, and can be widely used in HRV monitoring in various environments. Experimental results show that RF-ECG can obtain HRV values ​​comparable to traditional wired technology with high accuracy. In addition, a new heartbeat detection method using RFID tag array is proposed in this paper. This method can be regarded as an extremely lightweight sensor that can effectively distinguish multiple human subjects. The experiment explores the reflection effect of RFID tag array on heart rate signal, and establishes a relationship model between RF signal and heartbeat displacement. The RF-ECG system is designed based on COTS RFID tag array, and heart rate variability assessment is realized through signal preprocessing and heartbeat interval extraction. The experimental results show that under different user conditions, RF-ECG has good average heart rate error and stability, and can accurately detect changes in heart rate variability caused by different reasons.

[0009] The above-mentioned reflection-based technical solution realizes heart rate monitoring through radio waves reflected by the human body. Because the changes in the chest cavity caused by the heartbeat are very weak, sensing this change often requires a complex signal processing process and finding a balance between extracting sensitive features and resisting interference, resulting in low efficiency, high cost or complex implementation of existing technical methods.

[0010] Among them, WiFi signals are affected by factors such as environmental noise, interference and signal attenuation, which may cause signal distortion or instability. In addition, the impact of human breathing and heartbeat signals on WiFi signals is relatively weak, and there are also high requirements for the stability of environmental conditions. In addition, it takes a long time to process CSI, which may cause delays in signal processing and reduce real-time performance.

[0011] Compared with decimeter waves, millimeter waves have a higher frequency and are more sensitive to slight changes. However, their disadvantages are relatively poor penetration, large path loss, and they are more easily absorbed or scattered by clothing and objects. They are more sensitive to interference in the environment, which affects the quality of the detection signal.

[0012] RF-ECG technology has some limitations in practical applications, such as the need to pre-install multiple fixed-position electronic tags on specific parts of the human body, which increases the complexity of system deployment and brings a lot of inconvenience and discomfort to the tester. In addition, the accuracy and robustness of its interbeat interval (IBI) estimation still need to be improved. In some cases, such as movement or emotional changes, signal interference or distortion may occur, resulting in instability in monitoring results.

[0013] There is also Chinese patent CN105105739B, which discloses a short-distance wireless heart rate and heart rate variability detection method. It uses wireless reflected signals to detect heart rate variability, but is susceptible to multipath reflections in the environment and reflection interference from other surrounding objects, resulting in detection blind spots and co-frequency interference. Summary of the invention

[0014] In view of the above problems, the present invention provides a UHF radio wave-based transmission-type human heart rate detection device and method for improving signal stability and suppressing environmental interference.

[0015] The technical solution of the present invention is: a transmission type human heart rate detection device based on UHF radio waves, including a transmitter module, a transmitting antenna, a receiver module, a receiving antenna and a signal processing host computer,

[0016] The radio frequency interface of the transmitter module is connected to the transmitting antenna, and the radio frequency interface of the receiver module is connected to the receiving antenna.

[0017] The tester's body is located on the LOS link between the transmitting antenna and the receiving antenna.

[0018] The data interfaces of the transmitter module and the receiver module are respectively connected to a signal processing host computer.

[0019] The transmitter module and the receiver module both adopt ADALM-PLUTO SDR equipment of ADI Company, and the transmitting antenna and the receiving antenna both adopt flat-plate directional antenna.

[0020] A transmission-type human heart rate detection method based on UHF radio waves, comprising:

[0021] First, the signal processing host computer sends a start control signal to the transmitter module through the data interface. The transmitter module generates a radio frequency signal of a preset frequency and radiates it through the transmitting antenna in the form of electromagnetic waves. The electromagnetic waves penetrate the human body and are modulated by the human body's breathing and heartbeat movements.

[0022] Next, the receiving antenna receives the modulated transmission signal and transmits it to the receiver module through the radio frequency interface for demodulation. The raw data after digital sampling is sent to the signal processing host computer through the data interface and the signal processing host computer performs digital signal processing. The signal processing process includes: downsampling, filtering, feature extraction and heart rate calculation;

[0023] Finally, the target human heart rate information results are displayed on the human-computer interaction interface of the host computer.

[0024] The filtering process includes low-pass filtering, wavelet transform and subsequent filtering.

[0025] The low-pass filter is used to preliminarily filter out high-frequency noise interference while allowing the heart rate signal change frequency to pass;

[0026] The wavelet transform is used to filter out the DC bias component and extract the main component of the heartbeat frequency, further filter out high-frequency clutter, and make the waveform amplitude vary within a range near the zero value, which is more conducive to subsequent signal feature recognition.

[0027] The subsequent filtering is used to filter out abnormal peaks and zero-crossing local extreme points.

[0028] The low-pass filter adopts a Taylor low-pass filter, the order is set to 50, and the cut-off frequency fc is set to 8 Hz.

[0029] The wavelet transform includes wavelet decomposition and wavelet reconstruction.

[0030] Among them, wavelet decomposition includes:

[0031] First, the signal samples need to be decomposed by wavelet transform. The N-layer wavelet transform finally decomposes an estimated coefficient a N and N detail coefficients {d 1 ~d N};

[0032] When the 4th-order sym wavelet basis function type and the number of decomposition layers N=6 are selected, the heartbeat cycle fluctuation characteristics of the signal are analyzed;

[0033] Wavelet reconstruction includes:

[0034] The 6-layer wavelet analysis contains 7 coefficients {d 1 , d 2 , d 3 , d 4 , d 5 , d 6 , a 6}, where a 6 is the estimated coefficient, d 1 , d 2 , d 3 , d4 , d 5 , d 6 is the hierarchical detail coefficient; d 1 , d 2 , d 3 , d 4 represents the high frequency component, d 6 is the low frequency component, d 5 is the intermediate frequency component corresponding to the heart rate, and the extraction coefficient d 5 and coefficient d 6 The wavelet waveform represented by the signal is used to reconstruct the signal; the wavelet reconstruction formula is:

[0035]

[0036] In the formula, w j，k is the kth wavelet coefficient of the jth level, is the wavelet basis function, J is the reconstructed series, and t is time.

[0037] The subsequent filtering includes Hampel filtering and sliding average filtering.

[0038] The Hampel filter is used to prevent, detect and remove abnormal jitter peaks;

[0039] The sliding average filter is used to filter out local extreme values ​​crossing zero points.

[0040] The process of the Hampel filter is as follows:

[0041] Step 1: Set the window size N and threshold T of the moving data. The window contains data {x 1 , x 2 , ..., x N}, where N = 2k + 1, and N is at least 7;

[0042] Step 2: Arrange the data in the window in ascending order and find the median M of the sequence;

[0043] Step 3: Calculate the standard deviation of the median of each sample pair based on the absolute value of the median;

[0044] Find the median of the absolute deviation of each element relative to the window median M; for the new sequence {d i}={|x 1 -M|,|x 2 -M|,...,|x N -M|} sort them in ascending order and find the median D of this new sequence;

[0045] Scale the median absolute deviation; let Q = 1.4826*D, Q replaces D and becomes the new median absolute deviation;

[0046] Step 4: For each element x in the window i , calculate q = |x i -M|; if q<T*Q, then output y i =x i , otherwise replace the original data x with the median i , output y i =M.

[0047] The feature extraction and heart rate calculation include:

[0048] S1: Establish a window with the identified minimum point as the center; within this window, if the point is the minimum point, then retain the feature extraction of the point; if the point is not the minimum point in the window, cancel the feature extraction of the point;

[0049] S2: Calculate the average period of the sampling signal according to the horizontal coordinate position of these minimum points, and finally calculate the frequency and bpm; the specific method is as follows:

[0050] a) Assume that the signal has n minimum points, and the distance between two minimum points is calculated as 1 cycle, then the n-1 cycles are:

[0051] T 1 =x 2 -x 1 , T 2 =x 3 -x 2 , ..., T n-1 =x n -x n-1

[0052] b) Take the average of these n-1 periods and get the average period T during this period:

[0053]

[0054] c) The relationship between frequency and period is:

[0055]

[0056] d) The relationship between heart rate and frequency is:

[0057] bpm=60*f.

[0058] The present invention utilizes the penetrating characteristics of low-frequency electromagnetic waves and combines the processing method of filtering and reconstructing the transmitted wave signal to effectively solve the problems of reflection technology such as low sensitivity, weak penetration ability, high signal processing complexity and weak anti-interference ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the specific implementation of the present invention or the technical solution in the prior art, the following briefly introduces the drawings required for the specific implementation or the prior art description. In the drawings, each part is not necessarily drawn according to the actual scale.

[0060] Figure 1 It is a principle block diagram of the present invention.

[0061] Figure 2 is a signal processing flow chart of the present invention,

[0062] Figure 3 is a diagram of the signal filtering process in the present invention,

[0063] Figure 4 is the amplitude-frequency response diagram of the low-pass filter.

[0064] Figure 5 It is a schematic diagram of the time domain waveform of the test signal before and after low-pass filtering.

[0065] Figure 6 This is a schematic diagram of signal wavelet analysis and order selection.

[0066] Figure 7 It is a schematic diagram of the input and output signal waveform of wavelet transform.

[0067] Figure 8 It is the flow chart of Hampel filtering algorithm.

[0068] Fig. 9 It is the flow chart of sliding average filter algorithm.

[0069] Fig.10 This is a schematic diagram comparing the waveforms of the subsequent filtered input signal and the output signal.

[0070] Fig.11 It is a flow chart for extracting heart rate. DETAILED DESCRIPTION

[0071] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0072] In the description of the present invention, it should be understood that the terms "upper", "lower", "left", "right", "vertical", "horizontal", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific position, be constructed and operated in a specific position, and therefore cannot be understood as limiting the present invention. In the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0073] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0074] The technical implementation scheme of the present invention is further described in detail below with reference to the accompanying drawings.

[0075] (1) Hardware system composition

[0076] The hardware components of the heart rate detection system are as follows: Figure 1 As shown in Figure 1, the system consists of a transmitter module, a transmitting antenna, a receiver module, and a receiving antenna. The transmitter module and the receiver module use the ADALM-PLUTO SDR device of ADI, and the transmitting and receiving antennas use flat directional antennas. The main working parameter settings of the hardware are shown in Table 1.

[0077] Table 1 System hardware parameter settings

[0078] Transmitting frequency 915MHz Transmit antenna bandwidth 880-960MHz Receiving frequency 915MHz Receive antenna bandwidth 880-960MHz Transmitter sampling rate 200Ksps Transmitting antenna gain 6.5dBi Receiver sampling rate 200Ksps Receive antenna gain 6.5dBi Transmitter gain 0dB Antenna polarization Vertical polarization Receiver gain AGC (Automatic Gain Control) Data Frame 1000 sampling points

[0079] Before the test begins, the tester is located on the LOS link between the transmitting antenna and the receiving antenna, the RF interface of the transmitter module is connected to the transmitting antenna, the RF interface of the receiver module is connected to the receiving antenna, and the data interface of the transmitter module and the receiver module is connected to the signal processing host computer.

[0080] At the beginning of the test, the signal processing host computer sends a start control signal to the transmitter module through the data interface. The transmitter module generates a radio frequency signal of a preset frequency and radiates it through the transmitting antenna in the form of electromagnetic waves. The electromagnetic waves penetrate the human body and are modulated by the human body's breathing, heartbeat and other movements. Then, the receiving antenna receives the modulated transmission signal and transmits it to the receiver module through the radio frequency interface for demodulation. The raw data after digital sampling is sent to the signal processing host computer through the data interface and the signal processing host computer performs digital signal processing such as downsampling, filtering, feature extraction, and heart rate calculation. Finally, the target human heart rate information results are displayed on the human-computer interaction interface of the host computer.

[0081] When studying the mechanism of action between electromagnetic waves and the human body, the human body is generally equivalent to a multi-layer dielectric model, and then the electromagnetic field numerical analysis method is used to perform a full-wave solution in 3D space. When the heart is beating, blood circulates between different chambers of the heart, and the heart muscle alternates between two different states of "contraction" and "relaxation". In this process, parameters such as the dielectric thickness and equivalent dielectric constant ε in the multi-layer dielectric equivalent model of the human body change over time, causing the intensity of the transmitted wave to change over time. This is the principle of using transmitted waves to detect heartbeats. The transmitted wave in the present invention is related to the hardware system, deployment method, and the relative position of the tester.

[0082] The present invention has the following characteristics:

[0083] 1) Improve signal stability: The penetration of radio waves is related to the wavelength of radio waves. The shorter the wavelength, the weaker the signal's penetration ability. If higher frequency (centimeter wave, millimeter wave) electromagnetic waves are used for non-contact monitoring of human heart rate, the signal will be easily blocked by the thicker clothes of the person being tested, resulting in greater attenuation, thus affecting the effectiveness of the detection results. The present invention proposes a heart rate detection device and algorithm based on UHF (ultra-high frequency, with a frequency of 300 to 3000 MHz and a wavelength of 1m to 1dm) frequency band radio waves, which utilizes the ability of UHF radio waves to penetrate clothing and the human body to improve the accuracy and stability of heart rate detection results.

[0084] 2) Suppressing environmental interference: Most existing technologies perform feature extraction and filtering based on the disturbance of the phase of the reflected signal of the electromagnetic wave caused by the slight fluctuation of the chest surface. However, the unconscious movements of the subject's limbs and the reflections of multipath objects in the surrounding environment can easily interfere with it. Therefore, the existing technology has high requirements on the subject and the environment.

[0085] The present invention adopts a near-field deployment method with separated transmission and reception, and only receives electromagnetic wave signals that penetrate the human body, effectively suppressing interference from the surrounding environment and multipath reflections of objects, and reducing the environmental requirements for actual deployment.

[0086] (2) Signal processing flow

[0087] The present invention adopts a multi-stage filtering module, including low-pass filtering, wavelet transform, Hampel filtering and sliding filtering. Compared with the single-stage filtering method, the effects of noise filtering, singular point filtering, zero point filtering and DC bias component filtering are enhanced.

[0088] In the feature extraction process, according to the characteristics of the signal waveform, the minimum value feature is more obvious and stable. The method of minimum value position determination and extraction is adopted. Compared with the maximum value determination and extraction method, the obtained feature is more stable.

[0089] Specific as Figure 2 As shown, including:

[0090] Step 1: Downsampling. The purpose of downsampling is to reduce the amount of original data and speed up the efficiency of subsequent signal filtering processing. The sampling rate of the downsampled input original data is 200KHz, and the detection time window is about 20 seconds; the sampling rate of the downsampled output data is 100Hz, and the number of sampling points of the new data sample is about 2000 points.

[0091] Step 2: Signal filtering, such as Figure 3 As shown, including:

[0092] 1) Low-pass filtering

[0093] The role of low-pass filtering is to initially filter out high-frequency noise interference while allowing the heart rate signal to pass through. Usually, a person's heart rate is between 0.67Hz and 3.67Hz, and a normal heart rate is between 1Hz and 1.67Hz. A Taylor low-pass filter is used, the order is set to 50, and the cut-off frequency fc is set to 8Hz. The amplitude-frequency response characteristics of the designed low-pass filter are as follows: Figure 4 As shown in the figure, when the frequency is 3.67Hz, the amplitude of the filter is -1.37db, and the signal attenuation within 3.67Hz is very small. When the amplitude is -3db, the corresponding frequency is 3.8Hz, which means that the filter has a good attenuation effect on high-frequency signals out of the frequency band with a frequency higher than 3.67Hz. Figure 5 The time domain waveform comparison of the signal before and after low-pass filtering is given, which shows that after passing through the low-pass filter, the high-frequency variation components of the signal are filtered out and the low-frequency envelope information of the signal is retained.

[0094] 2) Wavelet transform

[0095] Wavelet Transform is a method used to analyze the frequency of time domain signals at different times, which can balance the time resolution and frequency resolution of signal analysis and filtering. The construction of wavelet transform includes two steps: wavelet decomposition and wavelet reconstruction.

[0096] a) Wavelet decomposition

[0097] First, the signal samples need to be decomposed by wavelet transform. The N-layer wavelet transform finally decomposes an estimated coefficient a N and N detail coefficients {d 1 ~d N After analyzing and optimizing the test signal, the 4th-order sym wavelet basis function type and the decomposition layer number N = 6 can better analyze the heartbeat cycle fluctuation characteristics of the signal (such as Figure 6 shown).

[0098] b) Wavelet reconstruction

[0099] exist Figure 6 In the analysis results, the 6-layer wavelet analysis contains 7 coefficients {d 1 , d 2 , d 3 , d 4 , d 5 , d 6 , a 6}, where a 6 is the estimated coefficient, d 1 , d 2 , d 3 , d 4 , d 5 , d 6 is the layered detail coefficient. 1 , d 2 , d 3 , d 4 represents the high frequency component, d 6 is the low frequency component, d 5 is the intermediate frequency component corresponding to the heart rate, so only the coefficient d is extracted 5 and coefficient d 6 The wavelet waveform represented by the signal is used to reconstruct the signal. The wavelet reconstruction formula is:

[0100]

[0101] In the formula, w j，k is the kth wavelet coefficient of the jth level, is the wavelet basis function, J is the reconstructed series, and t is time.

[0102] Figure 7 The time domain waveform comparison of the test signal before and after wavelet reconstruction is given. It can be seen that after the signal is processed by wavelet reconstruction, the DC bias component and useless high-frequency noise superposition are further filtered out, so that the waveform amplitude varies in the range near the zero value, which is more conducive to the subsequent signal feature recognition.

[0103] 3) Subsequent filtering

[0104] a) Hampel filtering

[0105] Abnormal peaks will also have a great impact on the recognition of heartbeat time domain signal features. Adding Hampel filtering can effectively prevent, detect and remove these abnormal jitter peaks. The Hampel filtering algorithm process is as follows: Figure 8 shown.

[0106] Step 1: Set the window size N and threshold T of the moving data. The window contains data {x 1 , x 2 , ..., x N}, where N = 2k + 1, and N is at least 7.

[0107] Step 2: Arrange the data in the window in ascending order (or from large to small order) and find the median M of the sequence.

[0108] Step 3: Calculate the standard deviation of the median of each sample pair based on the absolute value of the median.

[0109] Find the median of the absolute deviations of each element relative to the window median M. For the new sequence {d i}={|x 1 -M|,|x 2 -M|,...,|x N -M|} sort the sequences in ascending order (or descending order) and find the median D of the new sequence.

[0110] Scale the median absolute deviation. Let Q = 1.4826*D. Q replaces D and becomes the new median absolute deviation.

[0111] Step 4: For each element x in the window i , calculate q = |x i -M|. If q<T*Q, then output y i =x i , otherwise replace the original data x with the median i , output y i =M.

[0112] b) Sliding average filter

[0113] The local extreme value of zero crossing will also affect the recognition of the signal periodic characteristics. The sliding average filter can further smooth the signal, especially for eliminating the local mechanism of zero crossing. The principle of the sliding average filter algorithm is to remove the first data of a fixed length sequence (sequence length is N) after a new measurement, and move the remaining N-1 data forward one bit in turn, and insert the newly collected data at the end of the new queue (first-in-first-out principle). Then perform arithmetic operations on this newly formed sequence, and use the result as the result of this measurement. After multiple debugging and optimization, the window size N of the sliding average filter is adjusted to 40.

[0114] The steps of the sliding average filter algorithm are shown in Figure 9.

[0115] For a sequence x containing L elements, 1 , x 2 , ..., x L}

[0116] Step 1: Set a window size N. The larger N is, the smoother the signal will be after the sliding average filter. The value of N cannot be too large or too small. If the value of N is too large, the signal will be too smooth and useful information will be filtered out. If the value of N is too small, the noise added to the useful signal will not be filtered out, which will affect the subsequent signal results.

[0117] Step 2: For the sequence x i ={x 1 , x 2 , ..., x L-N+1}From the first element x 1 First, calculate the average value of all data in window N, and use this value as the new measurement result T i .x L-N+1 The number of subsequent data is smaller than the window size N, and the average value of N data cannot be calculated, so it is directly discarded. The new measurement result T is the signal after the sliding average filter.

[0118] After passing through the Hampel filter and the sliding average filter, the waveform comparison of the test signal is as follows Fig.10 As shown in the figure, it can be seen that the abnormal peaks and local extreme points of the signal crossing zero are well filtered out.

[0119] Step 3: feature extraction and heart rate value calculation, such as Fig.11 shown.

[0120] After the second step of signal filtering, the heartbeat time domain signal has obvious periodicity. The next feature extraction is to extract local extreme points. Since the characteristics of the signal's minimum point are more obvious than the maximum point, and the changes near the minimum point are more stable, the moment of the signal's minimum point is extracted as a feature. At the same time, there are still many interfering false minimum points near the signal's maximum point. These interfering points are called false points. The identification of these points will cause large errors in the automatic result statistics, so it is necessary to eliminate the false points. The specific algorithm steps are as follows:

[0121] S1: Establish a window with the identified minimum point as the center. In this window, if the point is the minimum point, then retain the feature extraction of the point; if the point is not the minimum point in the window, cancel the feature extraction of the point.

[0122] S2: Calculate the average period of the sampling signal according to the horizontal coordinate position of these minimum points, and finally calculate the frequency and bpm. The specific method is as follows:

[0123] a) Assume that the signal has n minimum points, and the distance between two minimum points is calculated as 1 cycle, then the n-1 cycle can be obtained as:

[0124] T 1 =x 2 -x 1 , T 2 =x 3 -x 2 , ..., T n-1 =x n -x n-1

[0125] b) Take the average of these n-1 periods and obtain the average period T during this period.

[0126]

[0127] c) The relationship between frequency and period is:

[0128]

[0129] d) The relationship between heart rate (bpm) and frequency is:

[0130] bpm=60*f

[0131] The present invention utilizes the transmission characteristics of UHF radio waves, the interaction mechanism between electromagnetic waves and human tissues, and the deployment method of a device with separate transmission and reception, to design a dedicated signal filtering and heart rate estimation algorithm to achieve accurate heart rate estimation.

[0132] The present invention has the following advantages:

[0133] 1) The radio waves used are UHF band electromagnetic waves, which have a relatively good ability to penetrate the human body.

[0134] 2) The proposed signal processing and filtering algorithm has strong resistance to environmental noise interference, strong environmental adaptability, and is suitable for different usage scenarios and user groups.

[0135] 3) Heart rate detection is contactless, and users do not need to wear any equipment or sensors. Since it does not require contact with the skin, it has deployment flexibility and convenience.

[0136] The present invention can be integrated with the existing UHF band radio frequency identification system, which greatly reduces the deployment difficulty and cost, making it more flexible and diverse in various application scenarios and having a wider application range. It will bring new development opportunities to the field of heart rate monitoring and is expected to provide users with more convenient and accurate heart rate monitoring services.

[0137] Regarding the contents disclosed in this case, there are a few points that need to be explained:

[0138] (1) The drawings of the embodiments disclosed in this case only involve the structures involved in the embodiments disclosed in this case, and other structures can refer to the general design;

[0139] (2) In the absence of conflict, the embodiments and features of the embodiments disclosed in this case can be combined with each other to obtain new embodiments;

[0140] The above are only specific implementation methods disclosed in this case, but the protection scope of the present disclosure is not limited thereto. The protection scope disclosed in this case should be based on the protection scope of the claims.

Claims

1. A transmission type human heart rate detection device based on UHF radio waves, characterized in that: It includes transmitter module, transmitting antenna, receiver module, receiving antenna and signal processing host computer. The radio frequency interface of the transmitter module is connected to the transmitting antenna, and the radio frequency interface of the receiver module is connected to the receiving antenna. The tester's body is located on the LOS link between the transmitting antenna and the receiving antenna. The data interfaces of the transmitter module and the receiver module are respectively connected to a signal processing host computer.

2. A UHF radio wave based transmission type human heart rate detection device according to claim 1, characterized in that: The transmitter module and the receiver module both adopt the ADALM-PLUTO SDR device of ADI Company.

3. The UHF radio wave-based transmission type human heart rate detection device according to claim 1, characterized in that: The transmitting antenna and the receiving antenna are both flat-plate directional antennas.

4. A transmission-type human heart rate detection method based on UHF radio waves, characterized in that: A transmission-type human heart rate detection device based on UHF radio waves as described in any one of claims 1 to 3, comprising: First, the signal processing host computer sends a start control signal to the transmitter module through the data interface. The transmitter module generates a radio frequency signal of a preset frequency and radiates it through the transmitting antenna in the form of electromagnetic waves. The electromagnetic waves penetrate the human body and are modulated by the human body's breathing and heartbeat movements. Next, the receiving antenna receives the modulated transmission signal and transmits it to the receiver module through the radio frequency interface for demodulation. The raw data after digital sampling is sent to the signal processing host computer through the data interface and the signal processing host computer performs digital signal processing. The signal processing process includes: downsampling, filtering, feature extraction and heart rate calculation; Finally, the target human heart rate information results are displayed on the human-computer interaction interface of the host computer.

5. The method for detecting human heart rate based on UHF radio waves according to claim 1, characterized in that: The filtering process includes low-pass filtering, wavelet transform and subsequent filtering. The low-pass filter is used to preliminarily filter out high-frequency noise interference while allowing the heart rate signal change frequency to pass; The wavelet transform is used to filter out DC bias components and high-frequency clutter, so that the waveform amplitude varies within a range near zero; The subsequent filtering is used to filter out abnormal peaks and zero-crossing local extreme points.

6. The method for detecting human heart rate based on UHF radio waves according to claim 1, characterized in that: The low-pass filter adopts a Taylor low-pass filter, the order is set to 50, and the cut-off frequency fc is set to 8 Hz.

7. The method for detecting human heart rate based on UHF radio waves according to claim 1, characterized in that: The wavelet transform includes wavelet decomposition and wavelet reconstruction. Among them, wavelet decomposition includes: First, the signal samples need to be decomposed by wavelet transform. The N-layer wavelet transform finally decomposes an estimated coefficient a N and N detail coefficients {d1~d N }; When the 4th-order sym wavelet basis function type and the number of decomposition layers N=6 are selected, the heartbeat cycle fluctuation characteristics of the signal are analyzed; Wavelet reconstruction includes: The 6-layer wavelet analysis contains 7 coefficients {d1, d2, d3, d4, d5, d6, a6}, where a6 is the estimated coefficient, d1, d2, d3, d4, d5, d6 are the hierarchical detail coefficients; d1, d2, d3d4 represent high-frequency components, d6 is the low-frequency component, and d5 is the intermediate frequency component corresponding to the heartbeat frequency. The wavelet waveform represented by coefficients d5 and d6 is extracted for signal reconstruction; the wavelet reconstruction formula is: In the formula, w j，k is the kth wavelet coefficient of the jth level, is the wavelet basis function, J is the reconstructed series, and t is time.

8. The UHF radio wave-based transmission-type human heart rate detection method according to claim 1, characterized in that: The subsequent filtering includes Hampel filtering and sliding average filtering. The Hampel filter is used to prevent, detect and remove abnormal jitter peaks; The sliding average filter is used to filter out local extreme values ​​crossing zero points.

9. The method for detecting human heart rate based on UHF radio waves according to claim 1, characterized in that: The process of the Hampel filter is as follows: Step 1: Set the window size N and threshold T of the moving data. The window contains the data {x1, x2, ..., x N }, where N = 2k + 1, and N is at least 7; Step 2: Arrange the data in the window in ascending order and find the median M of the sequence; Step 3: Calculate the standard deviation of the median of each sample pair based on the absolute value of the median; Find the median of the absolute deviation of each element relative to the window median M; for the new sequence {d i }={|x1-M|, |x2-M|,...,|x N -M|} sort them in ascending order and find the median D of this new sequence; Scale the median absolute deviation; let Q = 1.4826*D, Q replaces D and becomes the new median absolute deviation; Step 4: For each element x in the window i , calculate q = |x i -M|; If q<T*Q, then output y i =x i Otherwise, replace the original data x with the median i , output y i =M.

10. The UHF radio wave-based transmission-type human heart rate detection method according to claim 1, characterized in that: The feature extraction and heart rate calculation include: S1: Establish a window with the identified minimum point as the center; within this window, if the point is the minimum point, then retain the feature extraction of the point; if the point is not the minimum point in the window, cancel the feature extraction of the point; S2: Calculate the average period of the sampling signal according to the horizontal coordinate position of these minimum points, and finally calculate the frequency and bpm; the specific method is as follows: a) Assume that the signal has n minimum points, and the distance between two minimum points is calculated as 1 cycle, then the n-1 cycles are: T1=x2-x1,T2=x3-x2,...,T n-1 =x n -x n-1 b) Take the average of these n-1 periods and get the average period T during this period: c) The relationship between frequency and period is: d) The relationship between heart rate and frequency is: bpm=60*f.

Images