Millimeter wave radar human body heart rate measuring method and system based on heartbeat harmonic analysis

Through a millimeter wave radar method based on heartbeat harmonic analysis, phase information extraction and adaptive notch filter suppression of respiratory harmonics, combined with wavelet transformation to separate heartbeat signals, the traditional heart rate detection equipment is not suitable for infants or skin burn patients and distinguish them. Difficult problem, achieving high accuracy contactless heart rate monitoring.

CN119014840APending Publication Date: 2024-11-26GUANGANMEN HOSPITAL CHINA ACAD OF CHINESE MEDICAL SCI +1
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411120295.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-15
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

Traditional heart rate detection equipment needs to contact the human surface, especially not suitable for infants or skin burn patients. At the same time, it is difficult to effectively distinguish heartbeat harmonics and respiratory harmonics based on electrocardiogram and photoelectric plethysmography, resulting in inaccurate heart rate estimation.

Method used

Using a millimeter wave radar method based on heartbeat harmonic analysis, breath harmonic suppression and multi-stage heart rate estimation steps, the respiratory harmonics are suppressed by arctangent demodulation and adaptive notch filters, and combined with empirical wavelet transformation separation Heartbeat signal, using historical heart rate data to correct heart rate estimates.

Benefits of technology

It improves the accuracy of heart rate detection, reduces the impact of respiratory harmonics on heart rate estimation, and achieves robust contactless heart rate monitoring, suitable for daily health testing and medical monitoring.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119014840A_ABST
    Figure CN119014840A_ABST
Patent Text Reader

Abstract

The invention provides a millimeter wave radar human body heart rate measuring method and system based on heartbeat harmonic analysis. The method comprises the steps of phase information extraction, breathing harmonic suppression and multi-stage heart rate estimation. Phase information extraction can realize automatic discontinuity correction on the basis of obtaining phase information; in the breathing harmonic suppression step, harmonic frequency is obtained according to frequency multiplication of breathing frequency, and suppression is carried out by using an adaptive notch filter so as to reduce the influence of breathing harmonic on heart rate estimation; in the multi-stage heart rate estimation step, firstly, heartbeat and second harmonics thereof are separated out through empirical wavelet transform and serve as reference basis for heart rate estimation; secondly, interference signals are eliminated according to the breathing harmonic information, continuous analysis is carried out according to historical heart rate data so as to correct heart rate estimation under the condition of the weak heartbeat harmonic corresponding relation, and the heartbeat measurement method is simple and convenient in process, high in robustness, high in measurement precision and suitable for being applied to daily life.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of heart rate measurement, and in particular to a method and system for measuring human heart rate using a millimeter wave radar based on heartbeat harmonic analysis. Background Art

[0002] Heart rate is an important physiological indicator for assessing human health. It not only reflects heart function, but also reveals the body's adaptability to various internal and external environmental changes. It fluctuates with the intensity of human activity, emotional state, drug use and even climatic conditions. A stable heart rate and an appropriate range of variation are generally associated with good cardiovascular health and a lower risk of disease. Traditional heart rate detection equipment mainly relies on electrocardiogram or photoplethysmography, which requires contact with the subject's body surface. This wearing method may cause discomfort and is particularly inconvenient for infants or patients with skin burns. With the advancement of radar technology, radar has begun to be used in vital sign detection. Compared with traditional non-contact technology, radar-based methods have stronger penetration and anti-interference capabilities. By emitting low-energy electromagnetic waves to monitor the internal physiological movements of body organs, radar technology can not only achieve non-contact monitoring, but also can be used for sleep quality monitoring and fall prevention monitoring for the elderly. In recent years, the data processing algorithms at each stage of millimeter-wave radar heart rate monitoring have been optimized.

[0003] The signal preprocessing step is usually called phase demodulation. It is essentially the process of combining the received I and Q signals to recover the displacement signal. Among the phase demodulation methods, the most commonly used is the inverse tangent demodulation algorithm. By using an orthogonal receiver system, the limitations of a single receiver channel radar system can be eliminated by considering both channels simultaneously. The inverse tangent algorithm, by combining the two channels, can achieve accurate phase demodulation regardless of the target position or displacement amplitude.

[0004] In the signal processing stage, the commonly used modal decomposition algorithm is used to separate the heartbeat signal. However, the frequency band of respiratory harmonics is similar to the heartbeat frequency. The modal decomposition algorithm easily causes the spectrum of respiratory harmonics and heartbeat signals to be mixed, and the frequency of the heartbeat signal cannot be distinguished from them. In addition, the heartbeat harmonic signal is not effectively used for heartbeat frequency analysis. A method is needed to comprehensively analyze the heartbeat harmonics and historical heart rate information to obtain a robust heartbeat detection result. Summary of the invention

[0005] The purpose of the present invention is to overcome the existing problems and provide a millimeter wave radar human heart rate measurement method based on heartbeat harmonic analysis for heart rate detection. This method aims to improve the accuracy of heart rate detection through heartbeat harmonics and reduce the influence of respiratory harmonics on heart rate estimation. The method consists of three parts: phase information extraction, respiratory harmonic suppression and multi-stage heart rate estimation. The phase information extraction step aims to achieve automatic discontinuity correction on the basis of obtaining phase information. Considering that respiratory harmonic information is the main influence on heart rate detection, a respiratory harmonic suppression step is introduced. This step obtains the harmonic frequency according to the multiplication of the respiratory frequency, and uses an adaptive notch filter to suppress it to reduce the influence of respiratory harmonics on heart rate estimation. In addition, the present invention designs a multi-stage heart rate estimation step. First, the heartbeat and its second harmonic are separated by empirical wavelet transform as a reference for heart rate estimation; secondly, the interference signal is eliminated according to the respiratory harmonic information, and continuity analysis is performed according to the historical heart rate data to correct the heart rate estimation under the condition of weak heartbeat harmonic correspondence.

[0006] The present invention is implemented by the following technical solutions:

[0007] A method for measuring human heart rate using millimeter wave radar based on heartbeat harmonic analysis comprises the following steps:

[0008] S1: Phase information extraction: used to extract accurate phase signals containing human body information;

[0009] S11: The millimeter wave radar acquires an echo signal, and converts the echo signal into a baseband I / Q signal by down-converting the echo signal using an orthogonal mixer;

[0010] S12: extracting phase information of vital signs from the baseband I / Q signal using arc tangent demodulation;

[0011] S2: Respiratory harmonic suppression:

[0012] Acquire the respiratory frequency based on the phase information, calculate the respiratory harmonic frequency, and suppress the respiratory harmonics using a notch filter based on minimum mean square root; this step can reduce the impact of respiratory harmonics on heart rate estimation;

[0013] S3: Multi-stage heart rate estimation: used to integrate heartbeat harmonic information and robustly measure heartbeat frequency;

[0014] S31: Based on the phase information after respiratory harmonic suppression, separate the signals of the frequency bands where the heartbeat and the heartbeat harmonics are located, and obtain the frequency with higher amplitude in the two frequency bands as the candidate frequency;

[0015] S32: Determine whether there is a harmonic correspondence between the candidate frequencies of the heartbeat frequency band and the heartbeat harmonic frequency band in the current time period; if the harmonic correspondence exists for more than 90% of the time in the current time period, directly derive the estimated heart rate; otherwise, analyze and eliminate the respiratory harmonics, and use the historical heart rate data to correct the current heart rate data.

[0016] In the above technical solution, further, the step S12 is specifically:

[0017] For the acquired I / Q signal, the inverse tangent demodulation is used for the distance unit with the largest input signal amplitude to extract the phase information of the vital signs (inverse tangent demodulation is achieved by calculating the phase angle of the complex number of the received signal): first, constellation correction is performed to eliminate DC signal interference (the received complex signal may contain some undesirable DC components (DC offset), which will affect the accuracy of the phase and amplitude of the signal), and then the phase information is calculated by the inverse tangent function: Where I(t) and Q(t) are the real and imaginary parts of the signal respectively; the phase value obtained after phase demodulation is usually limited to the range of [-π, π], so a continuously changing phase output is finally generated through phase unwrapping.

[0018] Furthermore, the step S2 is specifically as follows: passing the phase signal spectrum information through a bandpass filter to obtain a signal in the respiratory frequency band, with the passband set to [0.1, 0.5] Hz; calculating the peak frequency of the respiratory frequency in this frequency band as the respiratory frequency. The respiratory frequency is doubled to obtain the second harmonic frequency of the respiratory frequency, and the cosine signal of the second harmonic frequency of the respiratory frequency is used as the reference signal of the notch filter, and the phase information is used as the input signal of the notch filter to filter out the respiratory harmonics. This step can eliminate the second harmonic of the respiratory frequency.

[0019] Furthermore, the step S31 is specifically as follows:

[0020] The phase signal after respiratory harmonic suppression is subjected to wavelet transform, and the initial boundary is set according to the heartbeat and heartbeat harmonic frequency range to separate the heartbeat signal and the second harmonic frequency band signal of the heartbeat; the number of peaks in the time domain of the heartbeat signal is counted as the upper limit of the actual frequency, and the spectrum peak is searched within the spectrum range of [0.7Hz, upper cutoff frequency], and the first three peak frequencies are taken in descending order of their amplitudes and included in the candidate frequency array f of the heartbeat. top For the second harmonic frequency band of the heartbeat, search for spectrum peaks in the frequency range of [1.4Hz, upper cutoff frequency], and take the first three peak frequencies in descending order of their amplitudes and classify them into the candidate frequency array f of the heartbeat harmonic frequency band top_2 .

[0021] Furthermore, the step S32 is specifically as follows:

[0022] Calculate f top The double of each frequency and f top_2 The difference Δf i,j , in view of the limitation of spectrum resolution, when the difference is less than 3bpm, the corresponding candidate heartbeat frequency is considered to have a harmonic correspondence with the candidate frequency of the heartbeat harmonic frequency band and is stored in a sliding window (the time length can be 5 seconds). If there is a harmonic correspondence for more than 90% of the time in the sliding window, enter (a), otherwise enter (b):

[0023] (a) If the harmonic correspondence exists for more than 90% of the time in the sliding window, it is considered that there is a strong correlation between the pair of heartbeats and the harmonics, and the heart rate can be directly estimated at f top Index i in * =argmin i (Δf i,j ), stΔf i,j ≤3bpm; index i * The corresponding candidate heart rate is the estimated heart rate;

[0024] (b) Considering the potential impact of respiratory harmonics on heart rate, the heart rate is estimated using the prior information of historical heart rate. The specific method is as follows:

[0025] The second harmonic of breathing is strong and is the main harmonic that affects the heart rate estimation (it has been eliminated in S2). In addition, when the asymmetry of the breathing signal is strong, the third or fourth harmonics are strong.

[0026] First, calculate the asymmetry parameter r of the respiratory signal: Among them, T total Represents the entire breathing cycle time, T inhale represents the inspiratory time; when r>η, η is the set threshold, f is further eliminated top The third and fourth harmonics of the middle breath, the specific method is:

[0027] Calculate the difference between the candidate heart rate and the respiratory rate multiple ∈, the calculation formula is:

[0028] ∈=|f top -μ(r)·rr|

[0029] Wherein, μ(r) represents the respiratory frequency multiple of the calculated error value, μ(r) = 3, 4; rr represents the respiratory rate; if ∈ is less than 3bpm, the frequency is considered to be a respiratory harmonic and is removed from the candidate frequencies;

[0030] Then, calculate the relative distance between the candidate heart rate and the historical heart rate Among them, A topi represent The corresponding amplitude of Represents historical heart rate, Representative candidate heart rate Relative distance from historical heart rate; take the candidate heart rate with the smallest relative distance as the estimated heart rate HR e .

[0031] The present invention also provides a millimeter wave radar human heart rate measurement system based on heartbeat harmonic analysis, which is used to execute the above method, including a phase information extraction module, a respiratory harmonic suppression module and a multi-stage heart rate estimation module;

[0032] The phase information extraction module is used to extract the phase signal containing human body information;

[0033] The respiratory harmonic suppression module is used to effectively suppress respiratory harmonics, obtain the respiratory frequency based on the phase information, calculate the respiratory harmonic frequency, and suppress the respiratory harmonics using a notch filter based on minimum mean square root;

[0034] The multi-stage heart rate estimation module integrates heartbeat harmonic information to robustly measure heartbeat frequency.

[0035] Furthermore, the phase information extraction module is used to extract the phase signal containing human body information, and the specific method is as follows:

[0036] The millimeter wave radar acquires an echo signal and converts the echo signal into a baseband I / Q signal by down-converting the frequency using an orthogonal mixer;

[0037] Phase information of vital signs is extracted from the baseband I / Q signal using arc tangent demodulation.

[0038] The multi-stage heart rate estimation module integrates the heartbeat harmonic information and robustly measures the heartbeat frequency. The specific method is as follows:

[0039] Based on the phase information after respiratory harmonic suppression, the signals of the frequency bands where the heartbeat and heartbeat harmonics are located are separated, and the frequencies with higher amplitudes in the two frequency bands are obtained as candidate frequencies;

[0040] Determine whether there is a harmonic correspondence between the candidate frequencies of the heartbeat frequency band and the heartbeat harmonic frequency band in the current time period; if the harmonic correspondence exists for more than 90% of the time in the current time period, directly obtain the estimated heart rate; otherwise, analyze and eliminate the respiratory harmonics, and use the historical heart rate data to correct the current heart rate data.

[0041] Furthermore, based on the phase information after respiratory harmonic suppression, the signals of the frequency bands where the heartbeat and the heartbeat harmonics are located are separated, and the frequencies with higher amplitudes in the two frequency bands are obtained as candidate frequencies. The specific method is as follows:

[0042] The phase signal after respiratory harmonic suppression is subjected to wavelet transform, and the initial boundary is set according to the heartbeat and heartbeat harmonic frequency range to separate the heartbeat signal and the second harmonic frequency band signal of the heartbeat; the number of peaks in the time domain of the heartbeat signal is counted as the upper limit of the actual frequency, and the spectrum peak is searched within the spectrum range of [0.7Hz, upper cutoff frequency], and the first three peak frequencies are taken in descending order of their amplitudes and included in the candidate frequency array f of the heartbeat. top For the second harmonic frequency band of the heartbeat, in the same way, search for the spectrum peak in the frequency range of [1.4Hz, upper cutoff frequency] and put it into the candidate frequency array f top_2 .

[0043] Furthermore, it is determined whether there is a harmonic correspondence between the candidate frequencies of the heartbeat frequency band and the heartbeat harmonic frequency band in the current time period; if the harmonic correspondence exists for more than 90% of the time in the current time period, the estimated heart rate is directly obtained; otherwise, the respiratory harmonics are analyzed and eliminated, and the historical heart rate data is used to correct the current heart rate data; the specific method is:

[0044] Calculate f top The double of each frequency and f top_2 The difference Δf i,j , when the difference Δf i,j When it is less than 3 bpm, the corresponding candidate heartbeat frequency is considered to have a harmonic correspondence with the candidate frequency of the heartbeat harmonic frequency band and is stored in a sliding window; if the harmonic correspondence exists for more than 90% of the time in the sliding window, enter (a), otherwise enter (b);

[0045] (a) Directly estimate the heart rate at f top Index i in * =argmin i (Δf i,j ), stΔf i,j ≤3bpm; index i * The corresponding candidate heart rate is the estimated heart rate;

[0046] (b) Considering the potential impact of respiratory harmonics on heart rate, the heart rate is estimated using the prior information of historical heart rate. The specific method is as follows:

[0047] First, calculate the asymmetry parameter r of the respiratory signal: Among them, T total Represents the entire breathing cycle time, T inhale represents the inspiratory time; when r>η, η is the set threshold, f is further eliminated top The third and fourth harmonics of the middle breath, the specific method is:

[0048] Calculate the difference between the candidate heart rate and the respiratory rate multiple ∈, the calculation formula is:

[0049] ∈=|f top -μ(r)·rr|

[0050] Wherein, μ(r) represents the respiratory frequency multiple of the calculated error value, μ(r) = 3, 4; rr represents the respiratory rate; if ∈ is less than 3bpm, the frequency is considered to be a respiratory harmonic and is removed from the candidate frequencies;

[0051] Then, calculate the relative distance between the candidate heart rate and the historical heart rate in, represent The corresponding amplitude of Represents historical heart rate, Representative candidate heart rate Relative distance from historical heart rate; take the candidate heart rate with the smallest relative distance as the estimated heart rate HR e .

[0052] The beneficial effects of the present invention are:

[0053] The present invention uses the I / Q signal after time domain filtering and applies the inverse tangent demodulation phase information to provide high demodulation accuracy while having natural resistance to the noise of the input signal. At the same time, the respiratory harmonic frequency is obtained through the spectrum information of the respiratory frequency band, and an adaptive notch filter based on the minimum mean square error LMS is designed to suppress the respiratory harmonics. In the process of heart rate estimation, based on the empirical wavelet transform, the heartbeat and the heartbeat harmonics are separated. If a strong heartbeat and the corresponding relationship between its harmonics are detected, the estimated heart rate is directly obtained; otherwise, the respiratory harmonics are analyzed and eliminated, and the prior knowledge obtained from the historical data is used to correct the current heart rate evaluation in the process. Robust estimation of heart rate under the influence of respiratory harmonics and noise is achieved. It can further serve daily health testing and medical monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 Flow chart of the method of the present invention.

[0055] Figure 2 A comparison chart of the results of using the method of the present invention and the FFT method to estimate the heart rate of a test person. DETAILED DESCRIPTION

[0056] like Figure 1 The present invention also provides a millimeter wave radar human heart rate measurement device based on heartbeat harmonic analysis, which is used to execute the millimeter wave radar human heart rate measurement method based on heartbeat harmonic analysis.

[0057] A millimeter wave radar human heart rate measurement system for heartbeat harmonic analysis includes a phase information extraction module, a respiratory harmonic suppression module and a multi-stage heart rate estimation module.

[0058] 1) Phase information extraction module

[0059] The millimeter-wave radar obtains the echo signal and uses an orthogonal mixer to down-convert it into a baseband I / Q signal.

[0060] Phase information of vital signs is extracted from the baseband I / Q signal using arc tangent demodulation.

[0061] For the acquired I / Q signal, the inverse tangent demodulation is used to extract the phase information of the distance unit with the largest input signal amplitude:

[0062] First, constellation correction is performed to eliminate DC signal interference: the received complex signal may contain some undesirable DC components (DC offset), which will affect the accuracy of the phase and amplitude of the signal. Inverse tangent demodulation is achieved by calculating the phase angle of the complex number of the received signal. The phase information of the signal can be calculated by the inverse tangent function: Where I(t) and Q(t) are the real and imaginary parts of the signal, respectively. The phase value obtained after phase demodulation is usually limited to the range of [-π, π]. A continuously changing phase output is generated through phase unwrapping. The phase correction process is as follows:

[0063] Φ n+1 -Φ n >π:Φ n+1 =Φ n+1 -2π

[0064] Φ n+1 -Φ n <-π:Φ n+1 =Φ n+1 +2π

[0065] The whole process must be performed sequentially, starting from n = 1 and ending at n = M-1. In each iteration, only Φ is updated. n+1 .

[0066] 2) Respiratory harmonic suppression module

[0067] The respiratory frequency is acquired based on the phase information, the respiratory harmonic frequency is calculated, and the respiratory harmonics are suppressed using a notch filter based on the minimum mean square root; this step can reduce the influence of the respiratory harmonics on the heart rate estimation.

[0068] The spectrum information of the phase signal is passed through a bandpass filter to obtain the signal in the respiratory frequency band, and the passband is set to [0.1, 0.5] Hz.

[0069] The peak frequency of the breathing in this frequency band is calculated as the breathing frequency f0.

[0070] The respiratory frequency f0 is doubled to obtain the second harmonic frequency of the respiratory frequency; the cosine signal of the second harmonic frequency of the respiratory frequency is used as the reference signal of the notch filter, and the phase information is used as the input signal of the notch filter to filter out the respiratory harmonic. This step can be eliminated to achieve the effect of filtering out the second harmonic of the respiratory frequency.

[0071] The calculation formula of the cosine signal of the second harmonic frequency of breathing is x(t)=Acos(nw0kt), (n=2), where A is the amplitude and nw0 is the angular frequency of the respiratory harmonic.

[0072] 3) Multi-stage heart rate estimation module

[0073] Firstly, based on the phase information after respiratory harmonic suppression, the signals in the frequency bands of heartbeat and heartbeat harmonics are separated, and the frequencies with higher amplitudes in the two frequency bands are obtained as candidate frequencies.

[0074] The phase signal after respiratory harmonic suppression is subjected to wavelet transformation, and the initial boundary is set according to the heartbeat and heartbeat harmonic frequency range, so as to separate the heartbeat signal and the signal of the second harmonic frequency band of the heartbeat.

[0075] By counting the number of peaks in the heartbeat signal time domain as the upper limit of the actual frequency, the spectrum peaks are searched in the spectrum range of [0.7Hz, upper cutoff frequency], and the first three peak frequencies are taken in descending order of their amplitudes and included in the candidate frequency array f of the heartbeat. top For the second harmonic frequency band of the heartbeat, search for spectrum peaks in the frequency range of [1.4Hz, upper cutoff frequency], and take the first three peak frequencies in descending order of their amplitudes and classify them into the candidate frequency array f of the heartbeat harmonic frequency band. top_2 .

[0076] Then, determine whether there is a harmonic correspondence between the candidate frequencies of the heartbeat frequency band and the heartbeat harmonic frequency band in the current time period; if the harmonic correspondence exists for more than 90% of the time in the current time period, directly obtain the estimated heart rate; otherwise, analyze and eliminate the respiratory harmonics, and use the historical heart rate data to correct the current heart rate data.

[0077] Calculate f top The double of each frequency and f top_2 The difference Δf i,j, in view of the limitation of spectrum resolution, when the difference is less than 3bpm, the corresponding candidate heartbeat frequency is considered to have a harmonic correspondence with the candidate frequency of the heartbeat harmonic frequency band and is stored in a 5-second sliding window. If there is a harmonic correspondence for more than 90% of the time in the sliding window, enter (a), otherwise enter (b):

[0078] (a) Calculate the index i of the candidate heart rate * The formula is as follows:

[0079] i * =argmin i (Δf i,j ), stΔf i,j ≤3bpm

[0080] (b) Considering the potential impact of respiratory harmonics on heart rate, the heart rate is estimated using the prior information of historical heart rate. The specific method is as follows:

[0081] The second harmonic of breathing is strong and is the main harmonic that affects heart rate estimation (it has been eliminated in the respiratory harmonic suppression module). When the asymmetry of the respiratory signal is strong, a more obvious third or fourth harmonic will be generated.

[0082] The asymmetry parameter r of the respiratory signal is calculated as follows: Among them, T total Represents the entire breathing cycle time, T inhale Indicates inspiratory time.

[0083] When r>η (η is the set threshold), the third and fourth harmonics of breathing need to be eliminated, and η is taken as 0.1 here.

[0084] Calculate the difference between the candidate heart rate and the respiratory rate multiple ∈, the calculation formula is:

[0085] ∈=|f top -μ(r)·rr|

[0086] Where ∈ represents the calculated difference, μ(r) represents the respiratory multiple of the calculated error value, μ(r) = 3, 4, and rr represents the respiratory rate. If ∈ is less than 3bpm, this frequency is considered a respiratory harmonic and is removed from the candidate frequencies. After eliminating these possible respiratory harmonic frequencies, the remaining candidate heartbeat frequencies are further evaluated. Based on their relative distance from the historical heart rate data, the frequency with the smallest relative distance is selected as the estimated value of the heart rate. The relative distance between the candidate heart rate and the historical heart rate The calculation formula is as follows:

[0087]

[0088] in, represent The corresponding amplitude of Represents historical heart rate, Representative candidate heart rate Relative distance from historical heart rate; index i * The corresponding candidate heart rate is the estimated heart rate HR e .

[0089] Depend on Figure 2 It can be seen that the heart rate measured by the method of the present invention is relatively close to the reference heart rate. The method has strong robustness and high measurement accuracy, and can further serve daily health detection and medical monitoring.

Claims

1. A method for measuring human heart rate using millimeter wave radar based on heartbeat harmonic analysis, characterized in that: The steps include: S1: Phase information extraction: used to extract the phase signal containing human body information; S11: The millimeter wave radar acquires an echo signal, and converts the echo signal into a baseband I / Q signal by down-converting the echo signal using an orthogonal mixer; S12: extracting phase information of vital signs from the baseband I / Q signal using arc tangent demodulation; S2: Respiratory harmonic suppression: Acquiring the respiratory frequency based on the phase information, calculating the respiratory harmonic frequency, and suppressing the respiratory harmonic using a notch filter based on a minimum mean square root; S3: Multi-stage heart rate estimation: S31: Based on the phase information after respiratory harmonic suppression, separate the signals of the frequency bands where the heartbeat and the heartbeat harmonics are located, and obtain the frequency with higher amplitude in the two frequency bands as the candidate frequency; S32: Determine whether there is a harmonic correspondence between the candidate frequencies of the heartbeat frequency band and the heartbeat harmonic frequency band in the current time period; If the harmonic correspondence exists for more than 90% of the time in the current time period, the estimated heart rate is directly obtained; otherwise, the respiratory harmonics are analyzed and eliminated, and the historical heart rate data is used to correct the current heart rate data.

2. The method for measuring human heart rate using millimeter wave radar based on heartbeat harmonic analysis according to claim 1, characterized in that: The step S12 is specifically as follows: For the acquired I / Q signal, the inverse tangent demodulation is used for the distance unit with the largest input signal amplitude to extract the phase information of the vital signs: firstly, constellation correction is performed to eliminate DC signal interference, and then the phase information is calculated by the inverse tangent function: Where I(t) and Q(t) are the real and imaginary parts of the signal respectively; finally, a continuously changing phase output is generated through phase unwrapping.

3. The method for measuring human heart rate using millimeter wave radar based on heartbeat harmonic analysis according to claim 1, characterized in that: The step S2 is specifically as follows: The phase signal spectrum information is passed through a bandpass filter to obtain the signal of the respiratory frequency band; the peak frequency in the respiratory frequency band is calculated as the respiratory frequency; the respiratory frequency is doubled to obtain the second harmonic frequency of the breathing, and the cosine signal of the second harmonic frequency of the breathing is used as the reference signal of the notch filter, and the phase information is used as the input signal of the notch filter to filter out the respiratory harmonics.

4. The method for measuring human heart rate using millimeter wave radar based on heartbeat harmonic analysis according to claim 1, characterized in that: The step S31 is specifically as follows: Performing wavelet transformation on the phase signal after respiratory harmonic suppression, setting the initial boundary according to the heartbeat and heartbeat harmonic frequency range, and separating the heartbeat signal and the signal of the second harmonic frequency band of the heartbeat; By counting the number of peaks in the heartbeat signal time domain as the upper limit of the actual frequency, the spectrum peaks are searched in the spectrum range of [0.7Hz, upper cutoff frequency], and the first three peak frequencies are taken in descending order of their amplitudes and included in the candidate frequency array f of the heartbeat. top For the second harmonic frequency band of the heartbeat, in the same way, search for the spectrum peak in the frequency range of [1.4Hz, upper cutoff frequency] and put it into the candidate frequency array f top_2 .

5. The method for measuring human heart rate using millimeter wave radar based on heartbeat harmonic analysis according to claim 4, characterized in that: The step S32 is specifically as follows: Calculate f top The double of each frequency and f top_2 The difference Δf i,j , when the difference Δf i,j When it is less than 3 bpm, the corresponding candidate heartbeat frequency is considered to have a harmonic correspondence with the candidate frequency of the heartbeat harmonic frequency band and is stored in a sliding window; If there is a harmonic correspondence for more than 90% of the time in the sliding window, go to (a), otherwise go to (b); (a) Directly estimate the heart rate at f top Index i in * =argmin i (Δf i,j ), stΔf i,j ≤3bpm; index i * The corresponding candidate heart rate is the estimated heart rate; (b) Considering the potential impact of respiratory harmonics on heart rate, the heart rate is estimated using the prior information of historical heart rate. The specific method is as follows: First, calculate the asymmetry parameter r of the respiratory signal: Among them, T total Represents the entire breathing cycle time, T inhale represents the inspiratory time; when r>η, η is the set threshold, f is further eliminated top The third and fourth harmonics of the middle breath, the specific method is: Calculate the difference between the candidate heart rate and the respiratory rate multiple ∈, the calculation formula is: ∈=|f top -μ(r)·rr| Wherein, μ(r) represents the respiratory frequency multiple of the calculated error value, μ(r) = 3, 4; rr represents the respiratory rate; if ∈ is less than 3bpm, the frequency is considered to be a respiratory harmonic and is removed from the candidate frequencies; Then, calculate the relative distance between the candidate heart rate and the historical heart rate in, represent The corresponding amplitude of Represents historical heart rate, Representative candidate heart rate Relative distance from historical heart rate; take the candidate heart rate with the smallest relative distance as the estimated heart rate HR e .

6. A millimeter wave radar human heart rate measurement system based on heartbeat harmonic analysis, used to execute the millimeter wave radar human heart rate measurement method based on heartbeat harmonic analysis as claimed in any one of claims 1 to 5, characterized in that: It includes a phase information extraction module, a respiratory harmonic suppression module and a multi-stage heart rate estimation module; The phase information extraction module is used to extract the phase signal containing human body information; The respiratory harmonic suppression module is used to effectively suppress respiratory harmonics, obtain the respiratory frequency based on the phase information, calculate the respiratory harmonic frequency, and suppress the respiratory harmonics using a notch filter based on minimum mean square root; The multi-stage heart rate estimation module integrates heartbeat harmonic information to robustly measure heartbeat frequency.

7. The millimeter wave radar human heart rate measurement system based on heartbeat harmonic analysis according to claim 6 is characterized in that: The phase information extraction module is used to extract the phase signal containing human body information, and the specific method is as follows: The millimeter wave radar acquires an echo signal and converts the echo signal into a baseband I / Q signal by down-converting the frequency using an orthogonal mixer; Phase information of vital signs is extracted from the baseband I / Q signal using arc tangent demodulation.

8. The millimeter wave radar human heart rate measurement system based on heartbeat harmonic analysis according to claim 6 is characterized in that: The multi-stage heart rate estimation module integrates the heartbeat harmonic information and robustly measures the heartbeat frequency. The specific method is as follows: Based on the phase information after respiratory harmonic suppression, the signals of the frequency bands where the heartbeat and heartbeat harmonics are located are separated, and the frequencies with higher amplitudes in the two frequency bands are obtained as candidate frequencies; Determine whether there is a harmonic correspondence between the candidate frequencies of the heartbeat frequency band and the heartbeat harmonic frequency band in the current time period; If the harmonic correspondence exists for more than 90% of the time in the current time period, the estimated heart rate is directly obtained; otherwise, the respiratory harmonics are analyzed and eliminated, and the historical heart rate data is used to correct the current heart rate data.

9. The millimeter wave radar human heart rate measurement system based on heartbeat harmonic analysis according to claim 6, characterized in that: The method of separating the signals of the frequency bands where the heartbeat and the heartbeat harmonics are located based on the phase information after respiratory harmonic suppression, and obtaining the frequencies with higher amplitudes in the two frequency bands as candidate frequencies is as follows: Performing wavelet transformation on the phase signal after respiratory harmonic suppression, setting the initial boundary according to the heartbeat and heartbeat harmonic frequency range, and separating the heartbeat signal and the signal of the second harmonic frequency band of the heartbeat; By counting the number of peaks in the heartbeat signal time domain as the upper limit of the actual frequency, the spectrum peaks are searched in the spectrum range of [0.7Hz, upper cutoff frequency], and the first three peak frequencies are taken in descending order of their amplitudes and included in the candidate frequency array f of the heartbeat. top For the second harmonic frequency band of the heartbeat, in the same way, search for the spectrum peak in the frequency range of [1.4Hz, upper cutoff frequency] and put it into the candidate frequency array f top_2 .

10. The millimeter wave radar human heart rate measurement system based on heartbeat harmonic analysis according to claim 9, characterized in that: The step of determining whether there is a harmonic correspondence between the candidate frequencies of the heartbeat frequency band and the heartbeat harmonic frequency band in the current time period; If the harmonic correspondence exists for more than 90% of the time in the current time period, the estimated heart rate is directly obtained; otherwise, the respiratory harmonics are analyzed and eliminated, and the historical heart rate data is used to correct the current heart rate data; the specific method is: Calculate f top The double of each frequency and f top_2 The difference Δf i,j , when the difference Δf i,j When it is less than 3 bpm, the corresponding candidate heartbeat frequency is considered to have a harmonic correspondence with the candidate frequency of the heartbeat harmonic frequency band and is stored in a sliding window; If there is a harmonic correspondence for more than 90% of the time in the sliding window, go to (a), otherwise go to (b); (a) Directly estimate the heart rate at f top Index i in * =argmin i (Δf i,j ), stΔf i,j ≤3bpm; index i * The corresponding candidate heart rate is the estimated heart rate; (b) Considering the potential impact of respiratory harmonics on heart rate, the heart rate is estimated using the prior information of historical heart rate. The specific method is as follows: First, calculate the asymmetry parameter r of the respiratory signal: Among them, T total Represents the entire breathing cycle time, T inhale represents the inspiratory time; when r>η, η is the set threshold, f is further eliminated top The third and fourth harmonics of the middle breath, the specific method is: Calculate the difference between the candidate heart rate and the respiratory rate multiple ∈, the calculation formula is: ∈=|f top -μ(r)·rr| Wherein, μ(r) represents the respiratory frequency multiple of the calculated error value, μ(r) = 3, 4; rr represents the respiratory rate; if ∈ is less than 3bpm, the frequency is considered to be a respiratory harmonic and is removed from the candidate frequencies; Then, calculate the relative distance between the candidate heart rate and the historical heart rate in, represent The corresponding amplitude of Represents historical heart rate, Representative candidate heart rate Relative distance from historical heart rate; take the candidate heart rate with the smallest relative distance as the estimated heart rate HR e .

Citation Information

Cited By

  • Millimeter wave radar heart rate signal enhancement method and system based on harmonic reconstruction

    CN120908769A

  • High-precision heart rate estimation method and system based on millimeter wave radar

    CN122074937A