A high-precision heart rate measurement method and system based on millimeter wave radar

Abstract

The application relates to the technical field of radar measurement, and discloses a high-precision heart rate measurement method and system based on a millimeter wave radar, which comprises the following steps: searching a chest cavity target through a frequency-modulated continuous wave radar, positioning a heart position, and extracting a complex signal corresponding to heart movement; performing different-dimension calculation on the complex signal, and obtaining a preliminary heart rate value through the calculation results; performing three-level verification mechanisms of motion interference exclusion, clutter interference exclusion and breathing harmonic interference exclusion on the preliminary heart rate value to perform confidence checking and reserve a high-reliability result; performing higher-precision calculation on data in a preset time period based on the high-reliability result of the preliminary heart rate value, and outputting a final heart rate value; the radar measurement heart rate accuracy can be greatly improved, multiple interferences can be effectively inhibited, low-cost hardware can be adapted, the applicability to a wide spectrum of people and complex scenes can be enhanced, only high-confidence heart rate values are output, and dynamic response to heart rate instantaneous changes can be realized.

Description

Technical Field

[0001] This invention relates to the field of radar measurement technology, and specifically to a high-precision heart rate measurement method and system based on millimeter-wave radar. Background Technology

[0002] Radar heart rate detection, as a non-contact technology, has broad application prospects in fields such as medicine and health monitoring. However, its current measurement accuracy of 60%–90% lags behind that of contact sensors (error <1%). The technical challenges and practical difficulties it faces are mainly reflected in the following three aspects: I. Signal strength and interference issues: 1. Weak human body echo signal: The amplitude of chest cavity vibration caused by heart rate is on the millimeter level, and the resulting Doppler frequency shift signal is extremely weak and easily masked by environmental noise. High-sensitivity radar hardware and complex algorithms are required to extract effective information.

[0003] 2. Multipath interference and environmental clutter: Reflections from indoor walls, furniture, and other objects create a multipath effect, causing the radar signal to contain multiple superimposed paths, interfering with the actual human motion signal. Simultaneously, dynamic interference (people walking, objects moving) or static interference (electromagnetic radiation from electrical equipment) in the environment can reduce the signal-to-noise ratio, affecting the accuracy of heart rate signals.

[0004] II. The complexity of physiological movement: 1. Strong interference from respiratory movements: The chest rise and fall caused by breathing reaches the centimeter level, which is much greater than the amplitude of the heartbeat corresponding to the heart rate. Moreover, its frequency (about 0.2-0.5Hz) partially overlaps with the heart rate (about 0.7-3Hz), which can easily cause the respiratory signal to be mixed with the heart rate signal, resulting in spectral leakage or artifacts.

[0005] 2. Interference from other bodily movements: Involuntary bodily movements (body movements, postural changes) or voluntary movements (waving, walking) generate strong Doppler signals, masking subtle heart rate fluctuations. Furthermore, differences in body size and chest wall thickness among individuals can affect the penetration and reflection characteristics of radar signals, leading to unstable signal amplitude and phase variations.

[0006] III. Hardware and Algorithm Challenges: 1. Hardware design limitations: Although high-resolution radar (such as phased array radar) can improve accuracy, it has the problems of high cost and large size, which is not conducive to miniaturized applications such as wearable devices.

[0007] 2. Algorithm Complexity and Robustness: During signal preprocessing, methods such as filtering (bandpass filtering, Kalman filtering) and denoising (principal component analysis) are used to suppress environmental interference, but over-filtering may result in the loss of effective heart rate information. Motion separation algorithms (such as empirical mode decomposition and independent component analysis) need to separate heartbeat and respiratory components from the mixed signal, but they have high computational complexity and poor real-time performance. Moreover, the heart rate signal characteristics of people of different ages, genders, and health statuses vary, and algorithms (especially AI algorithms) that lack adaptive learning capabilities are prone to generalization errors. Summary of the Invention

[0008] In view of this, and considering the shortcomings of existing technologies that affect the accuracy of radar heart rate measurement due to weak signals, susceptibility to interference, and hardware limitations, this invention provides a high-precision heart rate measurement method and system based on millimeter-wave radar. Based on the same subjects and conditions, the accuracy of radar heart rate measurement is significantly improved from 70% to over 99%, greatly expanding the application scope of radar heart rate monitoring technology.

[0009] In a first aspect, the present invention provides a high-precision heart rate measurement method based on millimeter-wave radar, comprising: The chest cavity target is searched by frequency-modulated continuous wave radar to locate the heart and extract the complex signal corresponding to the heart's movement. The complex signal is calculated in different dimensions, and the calculation results are cross-validated to obtain a preliminary heart rate value; A three-level verification mechanism, which excludes motion interference, noise interference, and respiratory harmonic interference, is used to check the confidence level of the preliminary heart rate values ​​and retain highly reliable results. Based on the highly reliable results of the preliminary heart rate values, the data within the most recent preset time period are calculated with higher precision to output the final heart rate value.

[0010] This invention locates the heart by searching for targets in the chest cavity and extracting complex signals. It uses at least two independent algorithms to calculate and cross-verify a preliminary heart rate value. A three-level verification mechanism, which excludes motion, clutter, and respiratory harmonic interference, retains a highly reliable result. Then, it outputs a high-precision heart rate value by performing local spectrum analysis on data from the most recent preset time period. This can significantly improve the accuracy of radar heart rate measurement from the traditional 60%-90% to over 99%. At the same time, it effectively suppresses multiple interferences, adapts to low-cost hardware, enhances applicability to a wide range of people and complex scenarios, and outputs only high-confidence heart rate values, dynamically responding to instantaneous changes in heart rate and improving the real-time performance of the algorithm.

[0011] In one optional implementation, the step of searching for chest cavity targets using frequency-modulated continuous wave radar, locating the heart, and extracting the complex signal corresponding to heart motion includes: A linear frequency modulated continuous wave (LFM) continuous wave signal is transmitted using a frequency modulated continuous wave radar. The received echo signal is then mixed with the transmitted signal, and the LFM signal is obtained by low-pass filtering. The LFM signal is then sampled by an IQ dual-channel ADC. The above process is repeated to form a sampling matrix. The range spectrum is obtained by windowing and zero-padding the sampling matrix in a preset fast time dimension and performing a fast Fourier transform. The static reflected clutter is obtained by averaging the accumulated range spectrum in a preset slow time dimension. The range spectrum of moving targets is obtained by subtracting static reflection clutter from the range spectrum, and moving targets with amplitudes greater than a threshold are retained. Targets within the preset range radar range are also selected to determine the heart location. Extract the complex signal from the center of motion within the range of motion corresponding to the heart location.

[0012] This invention extracts complex signals from the motion center, enabling highly sensitive acquisition of cardiac micro-motion signals, effectively suppressing strong motion interference such as respiration, retaining effective heart rate-related components, and providing pure data for subsequent heart rate calculations.

[0013] In one optional implementation, the complex signal is calculated in different dimensions, and the preliminary heart rate value is obtained by cross-validating the calculation results, including: After denoising using principal component analysis, zero-padding and fast Fourier transform are performed, and the first heart rate value is obtained by searching the spectral peak. The phase of the complex signal is extracted, unwound, and differentially processed using the phase unwinding differential method. Then, zero-padding and fast Fourier transform are performed, and the second heart rate value is obtained by searching the spectral peak. After extracting the phase and unwinding the complex signal using variational mode decomposition, variational mode decomposition is performed to separate the signal into respiratory and heartbeat signals. The heartbeat signal obtained by decomposition is zero-padding and subjected to fast Fourier transform, and the third heart rate value is obtained by searching the spectral peak. When the three differences between each pair of the first heart rate value, the second heart rate value, and the third heart rate value are within a preset range, the average of the two original heart rate values ​​corresponding to the minimum difference value is taken as the initial heart rate value.

[0014] The embodiments of the present invention can effectively improve the accuracy and anti-interference ability of heart rate measurement through multi-dimensional signal analysis and cross-validation mechanisms.

[0015] In one optional implementation, the three-level verification mechanism of motion interference exclusion, clutter interference exclusion, and respiratory harmonic interference exclusion, used to check the confidence level of the preliminary heart rate value, includes: Motion interference elimination: The sampling matrix is ​​divided into multiple parts in the preset slow time dimension. The average distance spectrum of the cumulative slow time dimension of each part is calculated. Static reflection clutter is subtracted to retain the moving target. The motion range of the moving target is determined and multiple center distances are obtained. When there are two or more center distance deviations that exceed the preset threshold, the confidence level is low and the calculation ends. Clutter interference elimination: Based on the sampling matrix being divided into multiple parts in the preset slow time dimension, the heart rate value is calculated using the algorithm in the preliminary heart rate value calculation, resulting in multiple heart rate results. These results are then compared, and the confidence level is set to low when two or more heart rate values ​​deviate from the preset threshold, at which point the calculation ends. Respiratory harmonic interference elimination: Search for the highest multiple peaks in the high-frequency part of the energy spectrum and determine whether the peaks are multiples of the initial heart rate value. When the number of peaks below the preset threshold is not a multiple of the initial heart rate value, the confidence level is low, and the calculation ends.

[0016] The embodiments of the present invention employ a three-level verification mechanism that eliminates motion interference, noise interference, and respiratory harmonic interference to filter out interference from body movement, multipath reflections, and respiration, thereby significantly improving the reliability of heart rate values ​​and ensuring the output of clinically reliable data in resting scenarios.

[0017] In one optional implementation, the step of performing a more precise calculation on data within the most recent preset time period based on the highly reliable result of the preliminary heart rate value, and outputting the final heart rate value, includes: Retain data from the most recent preset time period from the denoised complex signal; The retained data is padded with zeros and subjected to a fast Fourier transform. The peak value is searched in the energy spectrum within a preset range centered on the initial heart rate value. If the search is successful, a high-precision heart rate value is obtained; otherwise, the output is discarded.

[0018] Based on the highly reliable results of the preliminary heart rate value, this invention performs higher precision calculations on the data within the most recent preset time period and outputs the final heart rate value. By focusing on high spectral resolution analysis of short-duration data, the accuracy of heart rate measurement can be further improved and the dynamic response optimized.

[0019] In one optional implementation, the method further includes: Establish a sliding time window to repeat a complete heart rate measurement process at preset time intervals; After each calculation, determine whether a high-precision heart rate value has been obtained; When the high-precision heart rate value is 0 or fails the confidence check, the heart rate result will not be output in this loop; only the high-precision heart rate value that has passed all verifications will be output.

[0020] The embodiments of the present invention establish a sliding time window to repeatedly execute the complete heart rate measurement process at preset intervals, and determine the output based on the validity of high-precision heart rate values, thereby realizing dynamic tracking of heart rate signals and reliable data filtering.

[0021] Secondly, the present invention provides a high-precision heart rate measurement system based on millimeter-wave radar, the system comprising: The heart localization and signal extraction module is used to search for targets in the chest cavity using frequency-modulated continuous wave radar, locate the heart position, and extract the complex signal corresponding to the heart's movement. The preliminary heart rate value calculation module is used to perform calculations on the complex signal in different dimensions, and obtain the preliminary heart rate value by cross-referencing the calculation results. The heart rate confidence check module is used to perform a confidence check on the preliminary heart rate value through a three-level verification mechanism of motion interference elimination, clutter interference elimination, and respiratory harmonic interference elimination, retaining highly reliable results; The heart rate fine calculation module is used to perform more accurate calculations on data within the most recent preset time period based on the highly reliable results of the preliminary heart rate value, and output the final heart rate value.

[0022] Thirdly, the present invention provides a computer device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the high-precision heart rate measurement method based on millimeter-wave radar described in the first aspect or any corresponding embodiment thereof.

[0023] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the high-precision heart rate measurement method based on millimeter-wave radar according to the first aspect or any corresponding embodiment described above.

[0024] Fifthly, the present invention provides a computer program product, including computer instructions for causing a computer to execute the high-precision heart rate measurement method based on millimeter-wave radar according to the first aspect or any corresponding embodiment described above. Attached Figure Description

[0025] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0026] Figure 1This is a flowchart illustrating a high-precision heart rate measurement method based on millimeter-wave radar according to an embodiment of the present invention. Figure 2 This is a structural block diagram of a high-precision heart rate measurement system based on millimeter-wave radar according to an embodiment of the present invention; Figure 3 A schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] To overcome the shortcomings of existing radar heart rate measurement methods, which suffer from three key factors affecting accuracy: weak signals, susceptibility to interference, and hardware limitations, this embodiment provides a high-precision heart rate measurement method based on millimeter-wave radar. Based on low-cost frequency-modulated continuous wave radar, an efficient heart rate detection algorithm is proposed, achieving an overall accuracy of 99.88% in hundreds of hours of heart rate monitoring involving approximately 70 people. Figure 1 This is a flowchart of a high-precision heart rate measurement method based on millimeter-wave radar according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps: Step S1: Search for targets in the chest cavity using frequency-modulated continuous wave radar, locate the heart, and extract the complex signal corresponding to the heart's movement.

[0029] Specifically, radar detects heart rate by detecting the movement of the chest cavity at the heart. Therefore, to achieve high-precision heart rate measurement, the location of the heart must first be found. This embodiment of the invention uses low-cost frequency-modulated continuous wave radar to search for chest cavity targets, specifically including the following steps: Step S11: Transmit linear frequency modulated continuous wave (LFM) chirp signals using frequency modulated continuous wave radar. (If the period is T, then it is sent once every T seconds, and 1 / T times per second).

[0030] Step S12, Receive echo signal After transmitting the signal The signal is mixed and then low-pass filtered to obtain the intermediate frequency signal.

[0031] Step S13: The intermediate frequency signal is sampled by two ADCs, I and Q. Specifically, the signals generated by the I channel (real part) and Q channel (imaginary part) are converted into complex signals by analog-to-digital conversion.

[0032] Step S14: Repeat steps S11-S13 to form a sampling matrix; for example, obtain the sampling matrix SV = [sv(i, j)], where i = 1, 2, 3, ... I, j = 1, 2, 3, ... J, I is the total number of fast-time samples, i is the fast-time sample, J is the total number of slow-time samples within a window, and j is the slow-time sample. For example, T = 10 milliseconds, I = 256, that is, 256 points are sampled per chirp, J = 1200, that is, 1200 chirs are sampled, representing 1200 chirs sampled within 12 seconds, resulting in a 256×1200 sampling matrix.

[0033] Step S15: Window the sampling matrix in the preset fast time dimension, pad with zeros, and perform a Fast Fourier Transform to obtain the range spectrum. Then, average the accumulated range spectra in the preset slow time dimension to obtain the static reflection clutter. Specifically, for example, the 256 points sampled in the fast time dimension (10 milliseconds) are windowed (Hamming window), and zeros are added to the end to bring the total to 1024 points (i.e., N=1024). Then, an N-point Fast Fourier Transform (FFT) is performed to obtain the range spectrum D(i,j). The 1200 range spectra sampled within 12 seconds are accumulated and averaged to obtain the static reflection clutter Rstatic.

[0034] Step S16: Subtract the static reflection clutter from the range spectrum to obtain the range spectrum retained for the moving target. Specifically, subtract the static reflection clutter Rstatic from the range spectrum D(i,j) to retain the moving target, resulting in a new range spectrum D'(i,j). Further, find the range gate with the largest amplitude on the new range spectrum D'(i,j). The range gate number multiplied by the range resolution represents the center distance between the radar and the target. For example, if the range resolution is 0.1 meters, and the range gate number with the largest amplitude found on D'(i,j) is 10 (starting from 1, i.e., the 10th range gate), then the distance between the radar and the target is approximately 10 × 0.1 meters = 1 meter. In actual use, users must ensure that the distance between the radar and the chest cavity is appropriate; too close or too far will lead to a decrease in measurement accuracy.

[0035] Step S17: Retain moving targets whose amplitude is greater than a threshold, and filter targets within a preset range of the radar to determine the heart location; specifically, determine the movement range of the moving target through threshold detection, for example, retain a set of values ​​d(i,j) with energy greater than the threshold of 0.2 in D1(i,j), only retain targets within a specified range of the radar, such as 0.2 ~ 2 meters, and discard moving targets outside this range as interference.

[0036] Step S18: Extract the complex signal from the center of motion within the range of motion corresponding to the heart position. In this embodiment of the invention, the center of motion is obtained from the target range of motion, and complex signals C(j) are extracted from D1(i,j), totaling J=1200. This effectively suppresses strong motion interference such as respiration, increases the proportion of effective components in heart rate-related signals, and provides clean data for subsequent heart rate calculations.

[0037] Step S2: Perform calculations on the complex signal in different dimensions, and obtain a preliminary heart rate value by cross-referencing the calculation results.

[0038] Specifically, one of the key technologies for high-accuracy heart rate detection is the orthogonality of algorithm processing. This involves comparing the results of multiple independent algorithms on different dimensions of the same data set, retaining only the consistent results that can be mutually verified, thereby significantly improving accuracy. Optional heart rate detection algorithms include phase unwrapping differential, variational mode decomposition, and principal component analysis. An example is provided below: Algorithm 1: For complex signals, principal component analysis (PCA) is used for denoising, followed by zero-padding and Fast Fourier Transform (FFT). The first heart rate value is then obtained through spectral peak search. Specifically, this includes: 1. Perform principal component analysis on C(j), project the high-dimensional signal onto the low-dimensional principal component space, retain the effective signal component with the largest variance, and suppress the noise component with the smaller variance to obtain CP(j).

[0039] 2. Pad the denoised signal CP(j) to point M, for example, M=2048 and perform an M-point FFT. Search for peaks in the energy spectrum from 0.75Hz to 2.5Hz, corresponding to a heart rate range of 45 BPM (beats / minute) to 150 BPM, to obtain bpm_cplx, for example, 70.5 BPM.

[0040] It should be noted that this solution uses a frequency range of 45-150 BPM to cover the vast majority of the population. 45-150 BPM corresponds to a frequency range of 0.75-2.5 Hz.

[0041] Algorithm 2: For the processing of phase (real signal), the phase unwinding differential method is used to extract the phase of the complex signal, unwind it and perform differential processing, then perform zero padding and fast Fourier transform, and obtain the second heart rate value through spectrum peak search.

[0042] Specifically, the amplitude of chest movement caused by respiration is typically tens of times greater than that caused by heartbeat. This is reflected in the energy spectrum, where the energy of respiration and its harmonics far exceeds that of the heartbeat, thus masking the heartbeat signal. However, the rate and acceleration of respiration and heartbeat are different matters; differential analysis can significantly highlight the heartbeat signal, greatly reducing the difficulty of heart rate detection. The specific steps include: 1. Perform phase extraction on C(j) using the arctangent method to obtain phase(j).

[0043] 2. Unwrap phase(j) to obtain unwrap_phase(j), so that the phase change is between -π and π.

[0044] 3. Perform a difference operation on unwrap_phase(j) to obtain diff_phase(j). The first-order difference reflects the motion rate, and the second-order difference reflects the motion acceleration. Both can extract the heartbeat signal, and one of them can be selected.

[0045] 4. Pad diff_phase(j) to point M and perform an M-point FFT. Similarly, perform a peak search in the energy spectrum from 0.75Hz to 2.5Hz to obtain bpm_diff, for example, 70.8 BPM.

[0046] Algorithm 3: After extracting the phase and unwinding the complex signal using variational mode decomposition, variational mode decomposition is performed to obtain the breathing and heartbeat signals. The heartbeat signal is then zero-padded and subjected to fast Fourier transform. Finally, the third heart rate value is obtained through spectral peak search.

[0047] Specifically, the following steps are included: 1. Perform phase extraction on C(j) using the arctangent method to obtain phase(j); 2. Unwrap phase(j) to obtain unwrap_phase(j), so that the phase change is between -π and π; 3. Perform variational mode decomposition on unwrap_phase(j), where the number of intrinsic mode functions is set to 5 and the penalty factor is set to 500, resulting in 5 sets of intrinsic mode functions IMF(1)~IMF(5). The lowest frequency signal IMF(5) corresponds to the respiratory signal. The superposition signal of IMF(1) + IMF(2), i.e., IMF(1) and IMF(2), contains the heartbeat signal, which may also contain the heartbeat harmonic signal. 4. Zero-padding the superimposed IMF(1) + IMF(2) signal to point M and performing an M-point FFT. Similarly, peak search is performed in the energy spectrum from 0.75Hz to 2.5Hz to obtain bpm_vmd, for example, 71.3 BPM.

[0048] Mutual verification: When the three differences between each pair of the first, second, and third heart rate values ​​fall within a preset range, the average of the two original heart rate values ​​corresponding to the minimum difference is taken as the initial heart rate value. For example, in the example above, bpm_cplx=70.5, bpm_diff=70.8, and bpm_vmd=71.3, the three differences are 0.3, 0.8, and 0.5 respectively, all within the specified range of 1 BPM. In this case, bpm_cplx and bpm_diff corresponding to the minimum difference of 0.3 are selected, and the average of bpm_cplx and bpm_diff is calculated as bpm1=70.65, which is taken as the result of the initial heart rate value calculation.

[0049] This invention effectively improves the accuracy and anti-interference capability of heart rate measurement through multi-dimensional signal analysis and cross-validation mechanisms. The accuracy of bpm1 can be significantly improved to over 90%, and further improvement requires calculating confidence levels and performing screening.

[0050] Step S3 involves a three-level verification mechanism that eliminates motion interference, noise interference, and respiratory harmonic interference to check the confidence level of the preliminary heart rate value and retain highly reliable results.

[0051] Specifically, the second key technology for obtaining highly accurate heart rate in this embodiment of the invention is the calculation and checking of the confidence level of the heart rate value, retaining only the highly reliable results. This includes the following steps: I. Motion Interference Elimination: As a non-contact product, radar heart rate monitoring primarily relies on the subject at rest, such as during sleep, lying down, or sitting. It is not suitable when the subject is in motion. Therefore, motion interference must be eliminated first. Specifically, the sampling matrix is ​​divided into multiple parts along a preset slow-time dimension. The average distance spectrum of each part along the cumulative slow-time dimension is calculated, and static reflection clutter is subtracted to retain the moving target. The movement range of the moving target is determined, resulting in multiple center distances. When two or more center distance deviations exceed a preset threshold, the confidence level is low, and the calculation ends.

[0052] In one embodiment, the following steps are included: 1. Divide matrix D(i,j) into 4 parts in the slow time dimension, namely D1 = [d(i, 1:J / 4)], D2 = [d(i, J / 4+1:J / 2)], .

[0053] 2. Calculate the mean of the distance spectrum of the slow time dimension accumulated in matrices D1 to D4, subtract Rstatic, retain the moving target, determine the motion range of the moving target through threshold detection, and further obtain the four center distances.

[0054] 3. Compare the four center distances. If two or more center distances deviate from the threshold (e.g., 5 cm), the confidence level is low, and the calculation ends.

[0055] II. Clutter Interference Removal: In indoor scenarios, reflections from walls, furniture, and other objects create multipath effects, causing the radar received signal to contain the superposition of multiple paths, thus drowning out the true heart rate signal. Dynamic or static interference such as people moving around and electromagnetic radiation from electrical equipment further reduces the signal-to-noise ratio (SNR). In traditional methods, heart rate errors caused by clutter can reach ±10 BPM or more. Clutter interference removal is a key step in improving accuracy to over 99%. By eliminating outliers caused by non-physiological factors, the radar measurement results approach the clinical accuracy of electrocardiograms (ECG). Specifically, based on the sampling matrix divided into multiple parts in a preset slow time dimension, the algorithm used in the initial heart rate value calculation is used to calculate the heart rate value, obtaining multiple heart rate results. These results are compared, and when two or more heart rate values ​​deviate beyond a preset threshold, the confidence level is low, and the calculation ends.

[0056] In one embodiment, the following steps are included: 1. Based on D1~D4, select one of the algorithms in step S2 (e.g., algorithm processing one) to calculate the heart rate value, and obtain 4 results bpm_d1 ~ bpm_d4.

[0057] 2. Compare the differences between bpm_d1 and bpm_d4. If two or more heart rate values ​​deviate beyond the threshold, the confidence level is low, and the calculation ends. This eliminates occasional noise interference. The threshold is selected based on vital signs, such as 5 BPM. Under normal circumstances, human heart rate fluctuates between 0 and 2 BPM within 3 seconds. Therefore, 80 BPM fluctuates between 40 and 120 BPM within 3 seconds.

[0058] III. Respiratory Interference Elimination: Respiratory harmonics severely interfere with the heartbeat, making the heartbeat signal difficult to detect. However, respiratory harmonic interference is limited, with the main interference coming from the 3rd-4th harmonics, while the 10th and higher harmonics have almost no interference. At this point, the peaks on the energy spectrum mainly consist of the heartbeat or its harmonics. Therefore, we search for the highest number of peaks in the high-frequency part of the energy spectrum and determine whether they are multiples of bpm1. If the number of peaks below the threshold is not a multiple of bpm1, the confidence level is low, and this calculation ends.

[0059] Through the three-level verification mechanism of motion, clutter, and respiratory interference mentioned above, interference from body movement, multipath reflections, and respiration can be filtered out layer by layer, improving the accuracy of heart rate values ​​to over 99% and ensuring the output of clinically reliable data in resting scenarios.

[0060] Step S4: Based on the highly reliable results of the preliminary heart rate value, perform more precise calculations on the data within the most recent preset time period and output the final heart rate value.

[0061] The third key technology for obtaining a highly accurate heart rate in this invention is the fine calculation of the heart rate value, further reducing the error to within 1 BPM or even 0.5 BPM. Specifically, data within the most recent preset time period is retained from the denoised complex signal; the retained data is padded with zeros and subjected to a Fast Fourier Transform; a peak value is searched in the energy spectrum within a preset range centered on the initial heart rate value; if the search is successful, a high-precision heart rate value is obtained; otherwise, the current output is discarded.

[0062] In one embodiment, the following steps are included: 1. Retain the data from the most recent 3 seconds from CP(j), which is CP3s = CP(k...J).

[0063] 2. Pad CP3s to point M and perform an M-point FFT. Search for peak values ​​in the energy spectrum [bpm1-10, bpm1+10]. If the search is successful, obtain bpm2. If no peak value is found, set bpm1 to 0.

[0064] 3. bpm2 is the heart rate calculation result of bpm1 for the most recent 3 seconds. Output bpm2.

[0065] The execution process of steps S1-S4 above is repeated in the radar. A sliding time window is established, and a complete heart rate measurement process is repeated once at a preset time interval. After each calculation, it is determined whether a high-precision heart rate value has been obtained. If the high-precision heart rate value is 0 or fails the confidence check, no heart rate result is output for this loop; only the high-precision heart rate value that has passed all verifications is output. For example, if a calculation is performed once per second, and a satisfactory value bpm2 is not obtained in this calculation, no heart rate result is output for this loop, thus ensuring extremely high accuracy of the output value and enabling dynamic tracking of the heart rate signal and reliable data filtering.

[0066] This embodiment also provides a high-precision heart rate measurement system based on millimeter-wave radar. This system is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the system described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0067] This embodiment provides a high-precision heart rate measurement system based on millimeter-wave radar, such as Figure 2 As shown, it includes: The heart localization and signal extraction module 21 is used to search for chest cavity targets using frequency-modulated continuous wave radar, locate the heart position, and extract the complex signal corresponding to heart movement. The preliminary heart rate value calculation module 22 is used to perform calculations on the complex signal in different dimensions, and obtain the preliminary heart rate value by cross-referencing the calculation results. The heart rate confidence check module 23 is used to check the confidence of the preliminary heart rate value through a three-level verification mechanism of motion interference elimination, clutter interference elimination and respiratory harmonic interference elimination, and retain highly reliable results. The heart rate fine calculation module 24 is used to perform higher precision calculations on data within the most recent preset time period based on the highly reliable results of the preliminary heart rate value, and output the final heart rate value.

[0068] In some alternative implementations, the cardiac localization and signal extraction module 21 includes: The sampling matrix acquisition unit is used to transmit a linear frequency modulated continuous wave signal using a frequency modulated continuous wave radar, receive the echo signal and mix it with the transmitted signal, obtain the intermediate frequency signal through low-pass filtering, and sample the intermediate frequency signal using IQ dual-channel ADC. The above process is repeated to form a sampling matrix. The sampling matrix signal processing unit is used to window, zero-padding, and perform fast Fourier transform on the sampling matrix in a preset fast time dimension to obtain the range spectrum, and to calculate the mean value of the accumulated range spectrum in a preset slow time dimension to obtain static reflected clutter. The heart positioning unit is used to obtain the range spectrum of the moving target by subtracting static reflection clutter from the range spectrum, retain the moving targets whose amplitude is greater than a threshold, and filter targets within the preset range radar range to determine the heart location. The signal extraction unit is used to extract complex signals from the center of motion within the range of motion corresponding to the heart position.

[0069] In some optional implementations, the preliminary heart rate calculation module 22 includes: The first heart rate value calculation unit is used to perform zero padding and fast Fourier transform after denoising using principal component analysis, and obtain the first heart rate value through spectrum peak search. The second heart rate calculation unit is used to extract the phase of the complex signal using the phase unwinding differential method, unwind and differentially process it, perform zero padding and fast Fourier transform, and obtain the second heart rate value through spectrum peak search. The third heart rate calculation unit is used to extract the phase of the complex signal and unwrap it using the variational mode decomposition method, then perform variational mode decomposition to separate the signal to obtain the breathing and heartbeat signals. The heartbeat signal obtained by decomposition is zero-padding and fast Fourier transform, and the third heart rate value is obtained by searching the spectral peak. The preliminary heart rate value calculation unit is used to take the average of the two original heart rate values ​​corresponding to the minimum difference when the three differences between the first heart rate value, the second heart rate value, and the third heart rate value are within a preset range.

[0070] In some alternative implementations, the heart rate confidence check module 23 includes: The motion interference elimination unit is used to divide the sampling matrix into multiple parts in a preset slow time dimension, calculate the average distance spectrum of each part in the cumulative slow time dimension, subtract static reflection clutter to retain the moving target, determine the motion range of the moving target, and obtain multiple center distances. When there are two or more center distance deviations that exceed a preset threshold, the confidence level is low, and the calculation ends. The clutter interference elimination unit is used to divide the sampling matrix into multiple parts in a preset slow time dimension, calculate the heart rate value using the algorithm in the preliminary heart rate value calculation, obtain multiple heart rate results, and compare the multiple heart rate results. When there are two or more heart rate values ​​with a deviation exceeding a preset threshold, the confidence level is low, and the calculation ends. The respiratory harmonic interference elimination unit is used to search for the highest multiple peaks in the high-frequency part of the energy spectrum and determine whether the peaks are multiples of the initial heart rate value. When the number of peaks below a preset threshold is not a multiple of the initial heart rate value, the confidence level is low, and the calculation ends.

[0071] In some alternative implementations, the heart rate fine calculation module 24 includes: The data filtering unit is used to retain data within the most recent preset time period from the denoised complex signal; The data processing unit is used to pad the retained data with zeros and perform a fast Fourier transform. It searches for a peak value in the energy spectrum within a preset range centered on the initial heart rate value. If the search is successful, a high-precision heart rate value is obtained; otherwise, the current output is discarded.

[0072] In an optional implementation, the system further includes: a loop control module, used to establish a sliding time window and repeatedly execute a complete heart rate measurement process at a preset time interval; after each calculation, it is determined whether a high-precision heart rate value has been obtained; when the high-precision heart rate value is 0 or fails the confidence check, the heart rate result is not output in this loop, and only the high-precision heart rate value that has passed all verifications is output.

[0073] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0074] The high-precision heart rate measurement system based on millimeter-wave radar in this embodiment is presented in the form of functional units. Here, a unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0075] This invention also provides a computer device having the above-described features. Figure 2 The system shown is a high-precision heart rate measurement system based on millimeter-wave radar.

[0076] Please see Figure 3 , Figure 3 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 3 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 3 Take a processor 10 as an example.

[0077] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0078] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0079] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0080] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0081] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0082] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0083] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0084] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A high-precision heart rate measurement method based on millimeter wave radar, characterized in that, include: The chest cavity target is searched by frequency-modulated continuous wave radar to locate the heart and extract the complex signal corresponding to the heart's movement. The complex signal is calculated in different dimensions, and the calculation results are cross-verified to obtain a preliminary heart rate value. This process includes: independently calculating the complex signal using three different heart rate detection algorithms to obtain three heart rate values; and when the differences between any two of the three heart rate values ​​fall within a preset range, the average of the two heart rate values ​​corresponding to the minimum difference is used as the preliminary heart rate value. A three-level verification mechanism, consisting of motion interference elimination, clutter interference elimination, and respiratory harmonic interference elimination, is used to perform a confidence check on the preliminary heart rate values, retaining highly reliable results, including the preliminary heart rate values ​​that pass the confidence check. The motion interference elimination verification mechanism includes: dividing the sampling matrix into multiple parts in a preset slow time dimension, averaging the distance spectrum of each part in the cumulative slow time dimension, subtracting static reflection clutter to retain the moving target, determining the range of motion of the moving target, obtaining multiple center distances, and determining that the confidence level of the preliminary heart rate value is low when there are two or more center distance deviations exceeding a preset threshold, and the calculation ends. The verification mechanism for eliminating clutter interference includes: dividing the sampling matrix into multiple parts in a preset slow time dimension, using the algorithm in the preliminary heart rate value calculation to calculate the heart rate value of each part, obtaining multiple heart rate results, and comparing the multiple heart rate results. When there are two or more heart rate values ​​with deviations exceeding a preset threshold, the confidence level of the preliminary heart rate value is determined to be low, and the calculation ends. The verification mechanism for eliminating respiratory harmonic interference includes: searching for the highest number of peaks in the high-frequency part of the energy spectrum, determining whether the peaks are multiples of the initial heart rate value, and determining that the confidence level of the initial heart rate value is low when the number of peaks below a preset threshold is not a multiple of the initial heart rate value, and the calculation ends. Based on the initial heart rate value that has passed the confidence check, the data within the most recent preset time period is calculated with higher precision, and the final heart rate value is output.

2. The high-precision heart rate measurement method based on millimeter-wave radar according to claim 1, characterized in that, The process of searching for targets in the chest cavity using frequency-modulated continuous wave radar, locating the heart, and extracting the complex signal corresponding to heart motion includes: A linear frequency modulated continuous wave (LFM) continuous wave signal is transmitted using a frequency modulated continuous wave radar. The received echo signal is then mixed with the transmitted signal, and the LFM signal is obtained by low-pass filtering. The LFM signal is then sampled by an IQ dual-channel ADC. The above process is repeated to form a sampling matrix. The range spectrum is obtained by windowing and zero-padding the sampling matrix in a preset fast time dimension and performing a fast Fourier transform. The static reflected clutter is obtained by averaging the accumulated range spectrum in a preset slow time dimension. The range spectrum of moving targets is obtained by subtracting static reflection clutter from the range spectrum, and moving targets with amplitudes greater than a threshold are retained. Targets within the preset range radar range are also selected to determine the heart location. Extract the complex signal from the center of motion within the range of motion corresponding to the heart location.

3. The high-precision heart rate measurement method based on millimeter-wave radar according to claim 2, characterized in that, The complex signal is calculated in different dimensions, and the preliminary heart rate value is obtained by cross-referencing the calculation results, including: After denoising using principal component analysis, zero-padding and fast Fourier transform are performed, and the first heart rate value is obtained by searching the spectral peak. The phase of the complex signal is extracted, unwound, and differentially processed using the phase unwinding differential method. Then, zero-padding and fast Fourier transform are performed, and the second heart rate value is obtained by searching the spectral peak. After extracting the phase and unwinding the complex signal using variational mode decomposition, variational mode decomposition is performed to separate the signal into respiratory and heartbeat signals. The heartbeat signal obtained by decomposition is zero-padding and subjected to fast Fourier transform, and the third heart rate value is obtained by searching the spectral peak. When the three differences between each pair of the first heart rate value, the second heart rate value, and the third heart rate value are within a preset range, the average of the two original heart rate values ​​corresponding to the minimum difference value is taken as the initial heart rate value.

4. The high-precision heart rate measurement method based on millimeter-wave radar according to claim 1, characterized in that, Based on the highly reliable results of the preliminary heart rate value, a higher precision calculation is performed on the data within the most recent preset time period to output the final heart rate value, including: Retain data from the most recent preset time period from the denoised complex signal; The retained data is padded with zeros and subjected to a fast Fourier transform. The peak value is searched in the energy spectrum within a preset range centered on the initial heart rate value. If the search is successful, a high-precision heart rate value is obtained; otherwise, the output is discarded.

5. The high-precision heart rate measurement method based on millimeter-wave radar according to any one of claims 1-4, characterized in that, Also includes: Establish a sliding time window to repeat a complete heart rate measurement process at preset time intervals; After each calculation, determine whether a high-precision heart rate value has been obtained; When the high-precision heart rate value is 0 or fails the confidence check, the heart rate result will not be output in this loop; only the high-precision heart rate value that has passed all verifications will be output.

6. A high-precision heart rate measurement system based on millimeter-wave radar, characterized in that, include: The heart localization and signal extraction module is used to search for targets in the chest cavity using frequency-modulated continuous wave radar, locate the heart position, and extract the complex signal corresponding to the heart's movement. The preliminary heart rate value calculation module is used to perform calculations on the complex signal in different dimensions and obtain a preliminary heart rate value by cross-validating the calculation results. The calculations on the complex signal in different dimensions and the cross-validation of the calculation results to obtain the preliminary heart rate value include: independently calculating the complex signal using three heart rate detection algorithms to obtain three heart rate values; when the differences between any two of the three heart rate values ​​are within a preset range, the average of the two heart rate values ​​corresponding to the minimum difference is taken as the preliminary heart rate value. The heart rate confidence check module is used to perform a confidence check on the preliminary heart rate value through a three-level verification mechanism of motion interference elimination, clutter interference elimination, and respiratory harmonic interference elimination, retaining highly reliable results. The highly reliable results include the preliminary heart rate value that has passed the confidence check. The heart rate confidence check module includes: a motion interference elimination unit, a clutter interference elimination unit, and a respiratory harmonic interference elimination unit. The motion interference elimination unit is used to divide the sampling matrix into multiple parts in a preset slow time dimension, calculate the average distance spectrum of each part in the cumulative slow time dimension, subtract static reflection clutter to retain the moving target, determine the range of motion of the moving target, and obtain multiple center distances. When there are two or more center distance deviations exceeding a preset threshold, the confidence level of the preliminary heart rate value is determined to be low, and the calculation ends. The clutter interference elimination unit is used to divide the sampling matrix into multiple parts in a preset slow time dimension, calculate the heart rate value of each part using the algorithm in the preliminary heart rate value calculation, obtain multiple heart rate results, and compare the multiple heart rate results. When there are two or more heart rate values ​​with a deviation exceeding a preset threshold, the confidence level of the preliminary heart rate value is determined to be low, and the calculation ends. The respiratory harmonic interference elimination unit is used to search for the highest number of peaks in the high-frequency part of the energy spectrum, determine whether the peaks are multiples of the initial heart rate value, and determine that the confidence level of the initial heart rate value is low when the number of peaks below the preset threshold is not a multiple of the initial heart rate value, and the calculation ends. The heart rate fine calculation module is used to perform higher precision calculations on data within the most recent preset time period based on the preliminary heart rate value that has passed the confidence check, and output the final heart rate value.

7. A computer device, characterized in that, include: A memory and a processor are interconnected, the memory storing computer instructions, and the processor executing the computer instructions to perform the high-precision heart rate measurement method based on millimeter-wave radar as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the high-precision heart rate measurement method based on millimeter-wave radar as described in any one of claims 1 to 5.

9. A computer program product, characterized in that, Includes computer instructions for causing a computer to execute the high-precision heart rate measurement method based on millimeter-wave radar as described in any one of claims 1 to 5.

Images