Millimeter wave radar-based vital sign detection method and device, electronic equipment and storage medium

Abstract

The invention provides a vital sign detection method and device based on a millimeter wave radar, electronic equipment and a storage medium. The method comprises the following steps: obtaining a distance frequency domain signal based on a radar echo signal; performing spectral analysis on the distance frequency domain signal, positioning the position of the to-be-detected object, and extracting phase information at the position; filtering through a band-pass filter to obtain a breathing frequency band interval signal and an initial heartbeat frequency band interval signal; performing real-time respiration rate estimation based on the respiration frequency band interval signal to obtain a respiration frequency, and eliminating respiration harmonics in the initial heartbeat frequency band interval signal by using an adaptive wave trap based on the respiration frequency to obtain a heartbeat frequency band interval signal; and with the minimum fuzzy entropy as an optimization target, decomposing the heartbeat frequency band interval signal by using a variational mode decomposition algorithm to obtain a heartbeat signal, and carrying out real-time heart rate estimation based on the heartbeat signal to obtain a real-time heart rate. According to the method, the device, the electronic equipment and the storage medium provided by the invention, the accuracy of heart rate measurement can be improved.

Description

Technical Field

[0001] This invention relates to the field of signal processing technology, and in particular to a method, apparatus, electronic device, and storage medium for detecting vital signs based on millimeter-wave radar. Background Technology

[0002] Accurate and convenient vital sign detection technology can prevent accidents in life, provide effective information for clinicians to develop rehabilitation plans, and can also be integrated into smart home ecosystems to protect the lives of family members.

[0003] Based on the method of collecting vital signs signals, they can be divided into contact and non-contact vital sign detection. Millimeter-wave radar, as a non-contact vital sign detection device, has strong penetrability through clothing and building materials, is unaffected by light, and does not infringe on privacy, making it a promising application in long-term home monitoring.

[0004] Current radar systems have achieved a certain level of accuracy in measuring vital signs, but they still have shortcomings when applied to real-world environments, mainly in the following aspects: (1) The frequency amplitude of the respiratory signal is much greater than that of the heartbeat signal. Therefore, the estimation of heart rate is easily affected by environmental noise and respiratory harmonics, resulting in the final heart rate measurement result being far from the result of the contact device.

[0005] (2) Long-term detection results lack consistency. In experimental environments, most millimeter-wave radar devices only show a heart rate every minute or thirty seconds, which cannot meet the user's requirement to check the actual heart rate at any time.

[0006] (3) Insufficient robustness of long-term detection results. The state of the subjects during routine detection is different from that of the subjects who remain still in the experiment. It is difficult for subjects to remain at rest during routine vital sign detection. In this detection environment, the estimation of vital signs is more easily affected by body movements, resulting in insufficient robustness of the radar detection system.

[0007] In conclusion, there is an urgent need to provide an accurate and stable vital sign detection system suitable for long-term scenarios. Summary of the Invention

[0008] This invention provides a method, device, electronic device, and storage medium for detecting vital signs based on millimeter-wave radar, in order to solve the shortcomings of poor accuracy and stability in long-term vital sign detection scenarios in the prior art.

[0009] This invention provides a vital sign detection method based on millimeter-wave radar, comprising: The radar echo signal is sampled by analog-to-digital conversion to obtain an intermediate frequency signal. The intermediate frequency signal is then subjected to a fast Fourier transform in the range dimension to obtain a range frequency domain signal. Spectral analysis is performed on the distance frequency domain signal. Based on the analysis results, the position of the object to be detected is located, and the phase information at the position is extracted to obtain the phase signal. The phase signal is filtered by a bandpass filter to obtain the respiratory frequency band signal and the initial heartbeat frequency band signal; Real-time respiratory rate estimation is performed based on the respiratory frequency band signal to obtain the respiratory frequency of the subject under test. Based on the respiratory frequency, an adaptive notch filter is used to eliminate respiratory harmonics in the initial heartbeat frequency band signal to obtain the heartbeat frequency band signal. With the minimum fuzzy entropy as the optimization objective, the heartbeat frequency band signal is decomposed using the variational mode decomposition algorithm to obtain the heartbeat signal. Based on the heartbeat signal, real-time heart rate estimation is performed to obtain the real-time heart rate of the object to be detected.

[0010] According to the vital sign detection method based on millimeter-wave radar provided by the present invention, the step of decomposing the heartbeat frequency band signal using a variational mode decomposition algorithm with minimum fuzzy entropy as the optimization objective to obtain the heartbeat signal includes: The heartbeat frequency band signal is initialized and decomposed using a variational mode decomposition algorithm to obtain initial decomposition parameters. The fuzzy entropy is then calculated based on the decomposition parameters, which include the number of modes and a penalty factor. Using the minimum fuzzy entropy as the optimization objective, the initial decomposition parameters are iteratively optimized using the Grey Wolf optimization algorithm to obtain the optimized decomposition parameters; The heartbeat signal is determined based on the optimized decomposition parameters.

[0011] According to the vital sign detection method based on millimeter-wave radar provided by the present invention, determining the heartbeat signal based on the optimized decomposition parameters includes: Based on the optimized decomposition parameters, multiple intrinsic mode functions (IMF) components are obtained; The heartbeat signal is determined from the plurality of IMF components based on the center frequency and kurtosis of each IMF component.

[0012] According to the vital sign detection method based on millimeter-wave radar provided by the present invention, the step of estimating the real-time heart rate based on the heartbeat signal to obtain the real-time heart rate of the subject to be detected includes: Based on the sliding window and linear frequency modulated Z-transform (CZT) algorithm, the heart rate of the heartbeat signal is estimated in real time to obtain the real-time heart rate of the object to be detected.

[0013] According to the vital sign detection method based on millimeter-wave radar provided by the present invention, before performing spectrum analysis on the range-frequency domain signal, the method further includes: The moving target display algorithm is used to suppress static noise in the range frequency domain signal.

[0014] According to the vital sign detection method based on millimeter-wave radar provided by the present invention, the step of extracting phase information at the location to obtain a phase signal includes: Extract the phase information at the specified location to obtain the initial phase signal; The phase signal is obtained by using phase unwinding to correct the jump signal in the initial phase signal and eliminating the DC component deviation in the initial phase signal.

[0015] According to the vital sign detection method based on millimeter-wave radar provided by the present invention, the step of performing spectral analysis on the range-frequency domain signal and locating the position of the object to be detected based on the analysis results includes: The distance-frequency domain signal is subjected to spectral analysis based on the ordered statistical constant false alarm rate (OS-CFAR) algorithm, and the location of the object to be detected is located based on the analysis results.

[0016] The present invention also provides a vital sign detection device based on millimeter-wave radar, comprising: The signal transformation unit is used to perform analog-to-digital conversion sampling on the radar echo signal to obtain an intermediate frequency signal, and to perform a fast Fourier transform on the intermediate frequency signal in the range dimension to obtain a range frequency domain signal. The signal extraction unit is used to perform spectral analysis on the distance frequency domain signal, locate the position of the object to be detected based on the analysis results, extract the phase information at the position, and obtain the phase signal. The signal filtering unit is used to filter the phase signal through a bandpass filter to obtain the respiratory frequency band signal and the initial heartbeat frequency band signal; The harmonic cancellation unit is used to perform real-time respiratory rate estimation based on the respiratory frequency band interval signal to obtain the respiratory frequency of the object to be detected, and to use an adaptive notch filter to eliminate respiratory harmonics in the initial heartbeat frequency band interval signal based on the respiratory frequency to obtain the heartbeat frequency band interval signal. The signal decomposition unit is used to decompose the heartbeat frequency band signal with the minimum fuzzy entropy as the optimization objective and the variational mode decomposition algorithm to obtain the heartbeat signal. Based on the heartbeat signal, the real-time heart rate is estimated to obtain the real-time heart rate of the object to be detected.

[0017] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the vital signs detection method based on millimeter-wave radar as described above.

[0018] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the vital signs detection method based on millimeter-wave radar as described above.

[0019] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the vital signs detection method based on millimeter-wave radar as described above.

[0020] This invention provides a method, device, electronic equipment, and storage medium for detecting vital signs based on millimeter-wave radar. Considering that the chest cavity vibration displacement caused by the heart and lungs is on the millimeter scale, and that the amplitude of the heartbeat signal is much weaker than that of the respiratory signal, it is often easily submerged in the second harmonic of the respiratory signal. Therefore, an adaptive notch filter is used to eliminate the influence of respiratory harmonics in the initial heartbeat frequency band signal. Furthermore, addressing the issue of easily mis-estimating the heartbeat signal frequency, a variational mode decomposition algorithm is proposed, using minimum fuzzy entropy as the optimization objective, to decompose the heartbeat frequency band signal, thereby improving the accuracy and stability of heart rate measurement. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 This is one of the flowcharts of the vital signs detection method based on millimeter-wave radar provided by the present invention.

[0023] Figure 2 This is a schematic diagram of the structure of the delay canceller provided by the present invention.

[0024] Figure 3 This is a flowchart illustrating the structure of the adaptive notch filter provided by the present invention.

[0025] Figure 4 This is a signal spectrum diagram before and after using an adaptive notch filter, provided by the present invention.

[0026] Figure 5 This is a flowchart illustrating the heartbeat signal determination method provided by the present invention.

[0027] Figure 6 This is the second flowchart of the vital signs detection method based on millimeter-wave radar provided by the present invention.

[0028] Figure 7 This is a schematic diagram of the vital signs detection device based on millimeter-wave radar provided by the present invention.

[0029] Figure 8 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0031] To address the issues of poor accuracy and stability in long-term vital sign detection scenarios in existing technologies, this invention proposes a vital sign detection method based on millimeter-wave radar. In this method, the radar echo signal is sampled using analog-to-digital conversion to obtain an intermediate frequency (IF) signal. The IF signal is then subjected to a fast Fourier transform (FFT) in the range dimension to obtain a range-frequency domain signal. Spectral analysis is performed on the range-frequency domain signal, and the location of the target object is determined based on the analysis results. Phase information at the location is extracted to obtain a phase signal. The phase signal is then filtered using a bandpass filter to obtain a respiratory frequency band signal and an initial heartbeat frequency band signal. Real-time respiratory rate estimation is performed based on the respiratory frequency band signal to obtain the respiratory frequency of the target object. An adaptive notch filter is used to eliminate respiratory harmonics in the initial heartbeat frequency band signal based on the respiratory frequency to obtain the heartbeat frequency band signal. With minimum fuzzy entropy as the optimization objective, variational mode decomposition (VMD) is used to decompose the heartbeat frequency band signal to obtain the heartbeat signal. Real-time heart rate estimation is performed based on the heartbeat signal to obtain the real-time heart rate of the target object.

[0032] In this embodiment of the invention, considering that the chest cavity vibration displacement caused by the heart and lungs is on the millimeter scale, and that the amplitude of the heartbeat signal is much weaker than that of the respiratory signal, it is often easily submerged in the second harmonic of the respiratory signal. Therefore, an adaptive notch filter is used to eliminate the influence of respiratory harmonics in the initial heartbeat frequency band signal. Furthermore, addressing the issue of easily mis-estimating the heartbeat signal frequency, a variational mode decomposition algorithm is proposed to decompose the heartbeat frequency band signal using the minimum fuzzy entropy as the optimization objective, thereby improving the accuracy of heart rate measurement.

[0033] The embodiments of the present invention can be applied to scenarios requiring non-contact vital sign detection. The subject executing this method can be an electronic device such as a terminal device, computer, server, server cluster, or specially designed vital sign detection device, or a vital sign detection device installed in such an electronic device, which can be implemented through software, hardware, or a combination of both.

[0034] In the description of the embodiments of the present invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of the embodiments of the present invention, "multiple" means two or more, unless otherwise explicitly specified.

[0035] Figure 1 This is one of the flowcharts illustrating the vital sign detection method based on millimeter-wave radar provided by the present invention, such as... Figure 1 As shown, the method includes the following steps: Step 110: Perform analog-to-digital conversion sampling on the radar echo signal to obtain the intermediate frequency signal, and perform a fast Fourier transform on the intermediate frequency signal in the range dimension to obtain the range frequency domain signal.

[0036] Specifically, FMCW (Frequency Modulated Continuous Wave) radar scans the surrounding environment by continuously adjusting the frequency of its transmitted signal. The transmitted signal is reflected by the target to form an echo signal. This radar echo signal is received by the receiving antenna to form a received signal. The received signal and the transmitted signal are then mixed by a mixer to form an intermediate frequency (IF) signal. The IF signal is sampled by an analog-to-digital converter to obtain a digital signal that can be processed by a computer. By designing a suitable signal processing flow, vital signs signals can be extracted.

[0037] Table 1 shows the amplitude and frequency range of chest wall vibrations caused by cardiopulmonary activity in healthy adults. As can be seen from Table 1, the amplitude of chest displacement in the human body is far smaller than the minimum range resolution of radar. Therefore, it is impossible to obtain vital signs signals by simply analyzing the frequency difference. This demonstrates that obtaining vital signs signals through changes in the intermediate frequency signal is not feasible. Spectral analysis of the intermediate frequency signal of the subjects is used to determine the target, while the phase information in the echo signal carries the amplitude of chest wall vibrations in the target.

[0038] Table 1 Frequency and amplitude of vital signs

[0039] A frequency-modulated pulse signal transmitted by an FMCW radar is a chirp signal. Taking the sampling process of fast and slow time data, with one chirp signal per frame as an example, in... Found in each frame Each intermediate frequency (IF) signal has a time axis called the slow time axis. Each IF signal is generated after sampling. Each sampling point has a time axis called the fast time axis, which ultimately yields... The matrix is ​​obtained by performing a Fast Fourier Transform (FFT) operation on a frame of data along the fast time axis. The frequency of the spectral peak position represents the distance between the subject and the radar. The phase of the spectral peak position is extracted, and the phase carries the subject's vital signs information. The matrix is ​​then processed along... Perform FFT transformation in the direction, and The direction carries distance information, hence the name range-dimensional FFT, thus obtaining the range-frequency domain signal. Arranging each sample point along the slow time axis yields a duration of... Micro-motion information of the human body.

[0040] Step 120: Perform spectrum analysis on the distance frequency domain signal, locate the position of the object to be detected based on the analysis results, extract the phase information at the position, and obtain the phase signal.

[0041] Specifically, to obtain phase information containing vital signs signals, it is necessary to first select the correct spectral peaks from the spectral peaks as the target for phase extraction. The distance dimension FFT plays a core role in this process. It compresses each frame of discrete signal along the distance dimension and obtains the distance spectrum information within a certain time period by combining the data frame by frame.

[0042] Because the distances between various objects in the test environment and the radar are not exactly the same, and because the human chest cavity also vibrates, the frequency and phase of the radar echo signal will change accordingly. Time-frequency domain analysis of the range-dimensional feature vector can be performed to locate the test target. After determining the location of the test target, its phase information is extracted and arranged frame by frame to obtain the vital signs signals for that time period.

[0043] In some embodiments, before performing spectral analysis on the range frequency domain signal, the method further includes: suppressing static noise in the range frequency domain signal based on a moving target display algorithm.

[0044] Specifically, frequency-modulated continuous wave radar measures vital signs by focusing on the phase changes in the radar echo signal caused by chest vibrations, ultimately achieving high-precision and real-time estimation of vital signs. In practical applications, the presence of multiple stationary objects in the radar's vital sign measurement environment inevitably results in multiple DC components in the radar's range-frequency domain signal. To obtain high-precision vital sign signals, environmental noise needs to be eliminated.

[0045] Extensive experimental results demonstrate that human vital signs differ significantly from other environmental noise signals. The power of environmental noise signals is primarily distributed at zero frequency and integer multiples of the radar repetition frequency, and its power remains constant when no new target object appears in the radar-measurable environment. Based on these characteristics, moving target indication algorithms can effectively suppress this portion of environmental noise.

[0046] In some embodiments, environmental noise in this section can be suppressed by a delay canceller. Figure 2 This is a schematic diagram of the single-stage delay canceller structure provided by the present invention, as shown below. Figure 2 As shown, a single-stage delay-cancellation MTI filter consists of a delay unit and an adder. The delay time is equal to the period of the transmitted pulse. Figure 2 middle, This is represented as the pulse response of the canceller. Represented as input signal, Indicates impulse response and input signal The convolution between the input signals can be observed. Relationship with output information for:

[0047] Then the impulse response of the filter The frequency response and the frequency response are respectively:

[0048]

[0049] In formula (3) , Given the input signal frequency, the power gain of the entire system can be calculated as follows:

[0050] When the frequency of the input signal is a multiple in the spectrum, that is... At that time, the power gain of the frequency response corresponding to the first cancellation delay is

[0051] From equation (5), we can see that when In a single-stage delay canceller, the amplitude of a frequency that repeats at multiples can be reduced to zero, thus suppressing environmental noise. Cascading two single-stage delay cancellers yields a double-stage delay canceller. With increasing cancellation counts, the double-stage delay canceller exhibits better frequency response in both the stopband and passband than the single-stage delay canceller, and the suppression effect of the double-stage delay canceller is even better.

[0052] In some embodiments, step 120 involves performing spectral analysis on the distance frequency domain signal and locating the position of the object to be detected based on the analysis results, including: Step 121: Perform spectrum analysis on the distance frequency domain signal based on the ordered statistical constant false alarm rate (OS-CFAR) algorithm, and locate the position of the object to be detected based on the analysis results.

[0053] Specifically, after the MTI moving target display algorithm, static targets in the range frequency domain signal can be effectively removed. Next, the frequency part of the range frequency domain signal is used to estimate the distance information of the subject, that is, to perform spectrum analysis on the range frequency domain signal, and to locate the position of the object to be detected based on the analysis results.

[0054] For the scenario of vital sign detection, this embodiment uses OS-CFAR as the algorithm for target recognition of vital sign signals. The OS-CFAR algorithm improves robustness in multi-target and clutter edge environments by sorting the values ​​of reference cells and selecting the k-th value as the noise estimate. Target detection based on OS-CFAR can determine whether a subject is present at the corresponding location, and it has good shielding properties for multiple peaks in the range-frequency domain signal.

[0055] In some embodiments, step 120 extracts phase information at the location to obtain a phase signal, including: Step 122: Extract the phase information at the specified location to obtain the initial phase signal; Step 123: Correct the jump signal in the initial phase signal by using phase unwinding and eliminate the DC component deviation in the initial phase signal to obtain the phase signal.

[0056] Specifically, the measurement of vital signs signals mainly starts with phase signals, but the value range of phase signals is […]. When the value of the phase signal is greater than At this time, the phase will return to This phenomenon is also known as phase jump. To address this issue, this invention employs a phase unwinding method to process phase jumps in the phase signal. The processing method involves addressing the phenomenon where the difference between two consecutive phase points exceeds a certain threshold. or When this occurs, it is considered that a phase transition has occurred between these two phase points, and its mathematical expression is shown in equation (3.8):

[0057] As can be seen from formula (6), when the phase difference between two adjacent points... Greater than or less than When phase difference is performed, correction is required. After over-unwinding, the phase curve becomes smoother and there is no phase jump phenomenon.

[0058] Even after phase decoupling, a DC component still exists in the phase signal. If this DC component is not removed, it will cause distortion and loss of accuracy in the vital signs signal. When the distance between the subject and other stationary objects is less than the radar's minimum range resolution, the radar equipment cannot distinguish between the two objects, and this will present a spectral peak in the range frequency domain signal. Therefore, the phase signal will also contain a phase sequence obtained without DC offset correction. At this point, the phase signal already exhibits significant distortion because the DC component has not disappeared.

[0059] Therefore, the noise and DC component deviation in the initial phase signal can be eliminated by the DAM algorithm to obtain a noise-removed phase signal.

[0060] Step 130: Filter the phase signal using a bandpass filter to obtain the respiratory frequency band signal and the initial heartbeat frequency band signal.

[0061] Specifically, the phase signal contains both respiratory and heartbeat signals. To separate these two signals, a common method is to design two bandpass filters based on their frequency ranges. The frequency ranges of the respiratory and heartbeat signals are primarily used as key parameters for the bandpass filters. The respiratory signal frequency range is 0.1–0.5 Hz, and the heartbeat signal frequency range is 0.8–2 Hz.

[0062] After the extracted phase signal is processed by a bandpass filter, the respiratory frequency band signal and the initial heartbeat frequency band signal can be obtained respectively.

[0063] Step 140: Real-time respiratory rate estimation is performed based on the respiratory frequency band interval signal to obtain the respiratory frequency of the subject to be detected. Based on the respiratory frequency, an adaptive notch filter is used to eliminate respiratory harmonics in the initial heartbeat frequency band interval signal to obtain the heartbeat frequency band interval signal.

[0064] Specifically, the main targets for extracting vital signs signals are respiratory signals and heartbeat signals. Within the respiratory signal frequency range, there is only one relatively large value. Real-time respiratory rate estimation can be performed based on the respiratory frequency band signal to obtain the respiratory rate of the subject under test.

[0065] In the initial heart rate signal frequency band, there are multiple spectral peaks. Simply relying on spectral peaks to select the heart rate signal is unreliable. If an incorrect spectral peak is selected, it will also lead to incorrect heart rate estimation. These spectral peaks contain not only the second and third harmonics of the respiratory signal but also other noise signals. This can cause respiratory harmonics and other noise signals to be misinterpreted as heart rate signals, thus affecting the accuracy of heart rate prediction.

[0066] In this step, an adaptive notch filter is used to eliminate respiratory harmonics in the initial heartbeat frequency band signal based on the respiratory rate, thus obtaining the heartbeat frequency band signal.

[0067] An adaptive notch filter (ANF) is a filter used to eliminate or reduce specific frequency components in a signal. Its core principle is to dynamically adjust filter parameters through an adaptive algorithm to track and suppress interference frequencies. In vital signs signals, chest vibrations are primarily caused by respiration from the lungs; therefore, respiratory signals are often relatively easy to detect during radar vital sign detection. Taking advantage of this characteristic, an adaptive notch filter can be used to dynamically eliminate the influence of respiratory harmonics.

[0068] Figure 3 This is a flowchart illustrating the structure of the adaptive notch filter provided by the present invention, as follows: Figure 3 As shown in the figure For the input signal, ANF will generate two frequency components according to the specified notch frequency. and , will be mutually orthogonal and By combining accumulators, we can obtain , and The result of subtraction This is the final expected signal.

[0069] Assuming the vital signs signal to be processed Composed of two frequency components, it can be expressed as:

[0070] Assumption For the harmonic signals that the adaptive notch filter needs to process, for The initial phase. The frequencies of vital signs signals that need to be preserved. for The initial phase. This can be derived from mathematical formulas:

[0071] in and If the quantities are orthogonal and known, and we can know... Then the signal to be notched can be constructed. Finally, the problem is transformed into solving for the initial phase. Commonly used adaptive algorithms include Least Mean Square (LMS) and Recursive Least Squares (RLS). Since LMS is computationally simple and meets the experimental requirements for eliminating breathing harmonics, it was chosen as the desired signal. The optimization algorithm is as follows. Let the step parameter be... The LMS algorithm updates the filter weight coefficients as follows:

[0072] The system function between the error signal and the input is:

[0073] In formula (10) The amplitude of the reference signal is represented in this system. There is a zero point at this location, which is equivalent to a notch filter, and its bandwidth is:

[0074] Figure 4 The above is a signal spectrum diagram before and after using an adaptive notch filter, as provided by the present invention. It can be seen from the diagram that the breathing harmonic frequency component is suppressed.

[0075] Step 150: Using the minimum fuzzy entropy as the optimization objective, the heartbeat frequency band signal is decomposed using the variational mode decomposition algorithm to obtain the heartbeat signal. Based on the heartbeat signal, real-time heart rate estimation is performed to obtain the real-time heart rate of the object to be detected.

[0076] Specifically, considering that in the actual test environment, respiratory signals have a large amplitude and are relatively easy to extract, while heartbeat signals have a small amplitude and the chest vibrations caused by cardiac pulsation are weaker, and that environmental noise and clutter are easily mixed into the frequency band of the heartbeat signal, which cannot be filtered out by a simple bandpass filter. Therefore, in order to estimate the heart rate more accurately, this step uses a variational mode decomposition algorithm to decompose the heartbeat signal, which is unaffected by respiratory harmonics and other environmental noise.

[0077] In the traditional variational mode decomposition algorithm, the number of IMF components is manually selected. K and penalty factor However, this approach is prone to problems such as under-decomposition of signals and mode aliasing. The decomposition result of VMD depends on the preset number of modes k and the penalty factor α. Presetting the number of modes k and the penalty factor α based on experience is insufficient to handle different scenarios in vital sign detection. If the number of IMFs decomposed is insufficient, mode aliasing will occur; if the number of IMFs decomposed is too high, spurious components will be generated, leading to incorrect estimation of vital sign signals. Another parameter, the penalty factor α, if set too small, will result in excessively large bandwidth of the decomposed intrinsic mode components, leading to insufficient signal decomposition within the heartbeat frequency band when applied to vital sign signal decomposition. Conversely, if the penalty factor α is too large, the bandwidth of the decomposed components will be too small, resulting in the loss of useful information in the signal. Therefore, the selection of the [k, α] combination is also of great importance in the research of variational mode decomposition algorithms for vital sign detection.

[0078] This step introduces the concept of fuzzy entropy, using the minimum fuzzy entropy as the optimization objective. The variational mode decomposition (VMD) algorithm is used to decompose the heartbeat frequency band signal to obtain the heartbeat signal. The VMD algorithm incorporates fuzzy entropy for decomposition; the smaller the fuzzy entropy of each component, the more complete the decomposition of each component, indicating that the VMD algorithm selects […]. k , The more reasonable the combination, the greater the fuzzy entropy of each component, and the greater the complexity of decomposing each component, which indirectly indicates the choice at this time. k , The combination is unreasonable.

[0079] In some embodiments, step 150 specifically includes: Step 151: Initialize the heartbeat frequency band signal by using the variational mode decomposition algorithm to obtain the initial decomposition parameters. Calculate the fuzzy entropy based on the decomposition parameters, which include the number of modes and the penalty factor. Step 152: Using the minimum fuzzy entropy as the optimization objective, the initial decomposition parameters are iteratively optimized using the Grey Wolf optimization algorithm to obtain the optimized decomposition parameters; Step 153: Determine the heartbeat signal based on the optimized decomposition parameters.

[0080] Specifically, Figure 5 This is a flowchart illustrating the heartbeat signal determination method provided by the present invention, as shown below. Figure 5As shown, by combining a bandpass filter and adaptive notch filtering, respiratory harmonic signals that may fall within the heartbeat signal frequency band are removed using easily quantifiable respiratory frequencies. Then, the heartbeat frequency band signal processed by the bandpass filter and adaptive notch filtering is subjected to VMD decomposition. The Grey Wolf optimization algorithm is used to solve for the IMF component signal with the minimum fuzzy entropy as quickly as possible until the preprocessed heartbeat signal is completely decomposed. A suitable IMF component is selected for analysis, followed by minute-by-minute heart rate estimation and real-time heart rate estimation. The respiratory signal can be obtained by simple bandpass filtering followed by FFT transformation to extract the maximum amplitude.

[0081] In some embodiments, step 153 specifically includes: Step 153-1: Based on the optimized decomposition parameters, multiple intrinsic mode functions (IMF) components are obtained; Step 153-2: Determine the heartbeat signal from multiple IMF components based on the center frequency and kurtosis of each IMF component.

[0082] Specifically, numerous IMF components are decomposed using optimization algorithms. Related techniques often rely on experience to select one IMF component from the decomposed signal decomposition algorithm as the heartbeat signal for subsequent calculations, leading to poor stability. This embodiment uses center frequency and kurtosis as two indicators to select the decomposed IMF components. The center frequency is the energy concentration region of each IMF component in the frequency domain, reflecting the dominant frequency part of that component. Its calculation formula is as follows:

[0083] in This is the spectrum of the IMF component after FFT.

[0084] Kurtosis reflects the impulse characteristics of IMF components, and its calculation formula is shown below:

[0085] in The time-domain signal of the IMF component. For the first k The mean of each component, The standard deviation of the k-th component, and the physical meaning of kurtosis is when... When, this component approximates a Gaussian distribution, when When this occurs, it indicates that the component has a sharp peak, and this segment of the signal has been affected by noise. When the signal distribution is flat, it indicates that the signal component is flat.

[0086] Considering that vital sign data were collected at rest in the experimental environment, the kurtosis of the vital sign signal at this time is closer to 3. Taking into account the physical characteristics of the heartbeat signal, the modal component with a center frequency within the heartbeat signal frequency band should be selected first, and the kurtosis of this component should be closer to 3.

[0087] Based on any of the above embodiments, step 150, which involves real-time heart rate estimation based on the heartbeat signal to obtain the real-time heart rate of the object to be detected, includes: Step 154: Based on the sliding window and linear frequency modulated Z-transform (CZT) algorithm, perform real-time heart rate estimation on the heartbeat signal to obtain the real-time heart rate of the object to be detected.

[0088] Specifically, in related technologies, heart rate and respiratory data are often measured as a value per minute. However, this method of generating heart rate and respiratory values ​​does not reflect the condition of human vital signs in a timely manner. To address this issue, this invention employs a sliding window approach, using 10 seconds of vital sign signals to estimate 1 second of vital sign signals.

[0089] In spectrum analysis techniques, the FFT algorithm is undoubtedly a widely used method. However, considering that the FFT algorithm often requires a long observation window, and the sliding window cannot meet the requirement of a sufficiently long observation window for the FFT, this step introduces the linear frequency modulated Z-transform (CZT) algorithm. It can perform refined analysis of the spectrum in a local region, effectively overcoming the inherent limitations of the traditional FFT time window.

[0090] In view of the characteristics of the heartbeat signal and in order to estimate the heart rate more accurately, this step performs CZT spectrum refinement on the heartbeat signal. Comparing the heartbeat frequency band signal processed by FFT and the CZT frequency band signal, it can be seen that the heartbeat signal curve after CZT processing is smoother and the spectral peak value of the heartbeat signal is corrected.

[0091] In addition, the CZT spectrum refinement algorithm has the following advantages: the number of sampling points in the CZT algorithm does not need to be consistent with the length of the input signal, making it effective for short-time dynamic estimation of heart rate using a sliding window. Furthermore, the CZT output does not need to be a set of full-frequency signals like the FFT, due to the initial sampling points... The location can be set autonomously, which allows for high-precision analysis of pre-processed radar echo signals based on the frequency bands of heartbeat and respiratory signals.

[0092] Based on any of the above embodiments Figure 6 This is the second flowchart of the vital sign detection method based on millimeter-wave radar provided by the present invention, as shown below. Figure 6 As shown, the method includes: In the data preprocessing stage, radar echo data is sampled by an ADC to form intermediate frequency (IF) data. A range-dimensional FFT transform is performed on the acquired IF signal, resulting in multiple spikes observed in the spectrum. In the moving target identification stage, a quadratic delay cancellation method is used to eliminate reflected echoes from stationary objects in the radar field of view.

[0093] The CFAR algorithm, by setting a threshold, can accurately identify radar targets. It can be observed that the distance resolution of frequency is much greater than the vibration distance caused by breathing and heartbeats. Therefore, a simple distance-dimensional FFT cannot meet the requirements for measuring vital signs signals; a phase signal needs to be introduced for measurement. However, phase jumps and phase shifts still exist in the phase signal. This invention addresses these phenomena by using phase unwrapping and DC offset correction to obtain the correct phase signal.

[0094] Due to the physical characteristics of vital signs signals, bandpass filters can be used to obtain the respiratory frequency band signal and the initial heartbeat frequency band signal, respectively. However, harmonics of the respiratory signal may fall within the frequency band of the heartbeat signal. To address this issue, this embodiment of the invention utilizes the characteristics of the respiratory signal—large amplitude, strong anti-interference ability, and ease of extraction—to first separate the respiratory signal and calculate the respiratory rate. With the respiratory rate known, an adaptive notch filter is used to eliminate the influence of respiratory harmonics on the heartbeat signal.

[0095] By decomposing vital sign signals using bandpass filtering, it was found that multiple spectral peaks exist within the heartbeat frequency band. Simply relying on spectral peaks to select the heartbeat signal is unreliable; selecting the wrong peak can lead to incorrect heart rate estimation. Therefore, signal decomposition of the heartbeat frequency band signal is necessary. In the signal decomposition stage, minimizing fuzzy entropy is chosen as the optimization objective. The heartbeat frequency band signal is adaptively decomposed using variational mode decomposition (VMD) and the Grey Wolf Optimization Algorithm (GWO-FE-VMD). Among the different IMF components in the decomposition, the center frequency and kurtosis of each IMF component are used to select a suitable IMF component as the heartbeat signal component.

[0096] Finally, in the vital signs signal estimation algorithm, a sliding window is introduced to estimate real-time vital signs signals. Since FFT relies on a long observation window, the sampling points of the sliding window method cannot meet the requirements of FFT. Based on this, the linear frequency modulated Z-transform (CZT) algorithm is introduced to refine the spectrum of the 10-second window data. The transformed CZT spectrum is smoother, and the resulting spectrogram contains more points.

[0097] The vital signs detection device based on millimeter-wave radar provided by the present invention is described below. The vital signs detection device based on millimeter-wave radar described below can be referred to in correspondence with the vital signs detection method based on millimeter-wave radar described above.

[0098] Based on the above embodiments, Figure 7 This is a schematic diagram of the vital signs detection device based on millimeter-wave radar provided by the present invention, as shown below. Figure 7 As shown, the device includes: The signal conversion unit 710 is used to perform analog-to-digital conversion sampling on the radar echo signal to obtain an intermediate frequency signal, and to perform a fast Fourier transform on the intermediate frequency signal in the range dimension to obtain a range frequency domain signal. The signal extraction unit 720 is used to perform spectrum analysis on the distance frequency domain signal, locate the position of the object to be detected based on the analysis results, extract the phase information at the position, and obtain the phase signal. The signal filtering unit 730 is used to filter the phase signal through a bandpass filter to obtain the respiratory frequency band interval signal and the initial heartbeat frequency band interval signal; The harmonic cancellation unit 740 is used to perform real-time respiratory rate estimation based on the respiratory frequency band interval signal to obtain the respiratory frequency of the subject to be detected, and to use an adaptive notch filter to eliminate respiratory harmonics in the initial heartbeat frequency band interval signal based on the respiratory frequency to obtain the heartbeat frequency band interval signal. The signal decomposition unit 750 is used to decompose the heartbeat frequency band signal with the minimum fuzzy entropy as the optimization objective and the variational mode decomposition algorithm to obtain the heartbeat signal. Based on the heartbeat signal, the real-time heart rate is estimated to obtain the real-time heart rate of the object to be detected.

[0099] Based on any of the above embodiments, the signal decomposition unit is specifically used for: The heartbeat frequency band signal is initialized and decomposed using a variational mode decomposition algorithm to obtain initial decomposition parameters. The fuzzy entropy is then calculated based on the decomposition parameters, which include the number of modes and a penalty factor. Using the minimum fuzzy entropy as the optimization objective, the initial decomposition parameters are iteratively optimized using the Grey Wolf optimization algorithm to obtain the optimized decomposition parameters; The heartbeat signal is determined based on the optimized decomposition parameters.

[0100] Based on any of the above embodiments, the signal decomposition unit is specifically used for: Based on the optimized decomposition parameters, multiple intrinsic mode functions (IMF) components are obtained; The heartbeat signal is determined from the plurality of IMF components based on the center frequency and kurtosis of each IMF component.

[0101] Based on any of the above embodiments, the signal decomposition unit is specifically used for: Based on the sliding window and linear frequency modulated Z-transform (CZT) algorithm, the heart rate of the heartbeat signal is estimated in real time to obtain the real-time heart rate of the object to be detected.

[0102] Based on any of the above embodiments, the device further includes a noise suppression unit, used for: The moving target display algorithm is used to suppress static noise in the range frequency domain signal.

[0103] Based on any of the above embodiments, the signal extraction unit is specifically used for: Extract the phase information at the specified location to obtain the initial phase signal; The phase signal is obtained by using phase unwinding to correct the jump signal in the initial phase signal and eliminating the DC component deviation in the initial phase signal.

[0104] Based on any of the above embodiments, the signal extraction unit is specifically used for: The distance-frequency domain signal is subjected to spectral analysis based on the ordered statistical constant false alarm rate (OS-CFAR) algorithm, and the location of the object to be detected is located based on the analysis results.

[0105] Figure 8 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 8 As shown, the electronic device may include: a processor 810, a communications interface 820, a memory 830, and a communications bus 840, wherein the processor 810, the communications interface 820, and the memory 830 communicate with each other through the communications bus 840. The processor 810 can call logic instructions in the memory 830 to execute a vital sign detection method based on millimeter-wave radar. This method includes: performing analog-to-digital conversion sampling on the radar echo signal to obtain an intermediate frequency (IF) signal; performing a fast Fourier transform (FFT) on the IF signal in the range dimension to obtain a range-frequency domain signal; performing spectral analysis on the range-frequency domain signal; locating the position of the object to be detected based on the analysis results; extracting phase information at the position to obtain a phase signal; filtering the phase signal using a bandpass filter to obtain a respiratory frequency band signal and an initial heartbeat frequency band signal; estimating the respiratory rate in real time based on the respiratory frequency band signal to obtain the respiratory frequency of the object to be detected; eliminating respiratory harmonics in the initial heartbeat frequency band signal using an adaptive notch filter based on the respiratory frequency to obtain a heartbeat frequency band signal; decomposing the heartbeat frequency band signal using a variational mode decomposition algorithm with minimum fuzzy entropy as the optimization objective to obtain a heartbeat signal; and estimating the real-time heart rate of the object to be detected based on the heartbeat signal.

[0106] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0107] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the vital sign detection method based on millimeter-wave radar provided by the above methods. The method includes: performing analog-to-digital conversion sampling on the radar echo signal to obtain an intermediate frequency signal; performing a fast Fourier transform on the intermediate frequency signal in the range dimension to obtain a range-frequency domain signal; performing spectrum analysis on the range-frequency domain signal; locating the position of the object to be detected based on the analysis results; and extracting the phase signal at the location. The process involves: obtaining a phase signal; filtering the phase signal using a bandpass filter to obtain a respiratory frequency band signal and an initial heartbeat frequency band signal; estimating the respiratory rate in real time based on the respiratory frequency band signal to obtain the respiratory frequency of the subject; eliminating respiratory harmonics in the initial heartbeat frequency band signal using an adaptive notch filter based on the respiratory frequency to obtain the heartbeat frequency band signal; decomposing the heartbeat frequency band signal using a variational mode decomposition algorithm with the minimum fuzzy entropy as the optimization objective to obtain the heartbeat signal; and estimating the real-time heart rate of the subject based on the heartbeat signal to obtain the real-time heart rate of the subject.

[0108] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the vital sign detection method based on millimeter-wave radar provided by the above methods. The method includes: performing analog-to-digital conversion sampling on the radar echo signal to obtain an intermediate frequency signal; performing a fast Fourier transform on the intermediate frequency signal in the range dimension to obtain a range frequency domain signal; performing spectral analysis on the range frequency domain signal; locating the position of the object to be detected based on the analysis results; extracting phase information at the position to obtain a phase signal; filtering the phase signal through a bandpass filter to obtain a respiratory frequency band signal and an initial heartbeat frequency band signal; performing real-time respiratory rate estimation based on the respiratory frequency band signal to obtain the respiratory frequency of the object to be detected; using an adaptive notch filter to eliminate respiratory harmonics in the initial heartbeat frequency band signal based on the respiratory frequency to obtain a heartbeat frequency band signal; decomposing the heartbeat frequency band signal using a variational mode decomposition algorithm with minimum fuzzy entropy as the optimization objective to obtain a heartbeat signal; and performing real-time heart rate estimation based on the heartbeat signal to obtain the real-time heart rate of the object to be detected.

[0109] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0110] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0111] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A vital sign detection method based on millimeter-wave radar, characterized in that, include: The radar echo signal is sampled by analog-to-digital conversion to obtain an intermediate frequency signal. The intermediate frequency signal is then subjected to a fast Fourier transform in the range dimension to obtain a range frequency domain signal. Spectral analysis is performed on the distance frequency domain signal. Based on the analysis results, the position of the object to be detected is located, and the phase information at the position is extracted to obtain the phase signal. The phase signal is filtered by a bandpass filter to obtain the respiratory frequency band signal and the initial heartbeat frequency band signal; Real-time respiratory rate estimation is performed based on the respiratory frequency band signal to obtain the respiratory frequency of the subject under test. Based on the respiratory frequency, an adaptive notch filter is used to eliminate respiratory harmonics in the initial heartbeat frequency band signal to obtain the heartbeat frequency band signal. With the minimum fuzzy entropy as the optimization objective, the heartbeat frequency band signal is decomposed using the variational mode decomposition algorithm to obtain the heartbeat signal. Based on the heartbeat signal, real-time heart rate estimation is performed to obtain the real-time heart rate of the object to be detected.

2. The vital sign detection method based on millimeter-wave radar according to claim 1, characterized in that, The heartbeat frequency band signal is decomposed using a variational mode decomposition algorithm with the minimum fuzzy entropy as the optimization objective, resulting in a heartbeat signal including: The heartbeat frequency band signal is initialized and decomposed using a variational mode decomposition algorithm to obtain initial decomposition parameters. The fuzzy entropy is then calculated based on the decomposition parameters, which include the number of modes and a penalty factor. Using the minimum fuzzy entropy as the optimization objective, the initial decomposition parameters are iteratively optimized using the Grey Wolf optimization algorithm to obtain the optimized decomposition parameters; The heartbeat signal is determined based on the optimized decomposition parameters.

3. The vital sign detection method based on millimeter-wave radar according to claim 2, characterized in that, Determining the heartbeat signal based on the optimized decomposition parameters includes: Based on the optimized decomposition parameters, multiple intrinsic mode functions (IMF) components are obtained; The heartbeat signal is determined from the plurality of IMF components based on the center frequency and kurtosis of each IMF component.

4. The vital sign detection method based on millimeter-wave radar according to claim 1, characterized in that, The step of estimating the real-time heart rate based on the heartbeat signal to obtain the real-time heart rate of the object to be detected includes: Based on the sliding window and linear frequency modulated Z-transform (CZT) algorithm, the heart rate of the heartbeat signal is estimated in real time to obtain the real-time heart rate of the object to be detected.

5. The vital sign detection method based on millimeter-wave radar according to claim 1, characterized in that, Before performing spectral analysis on the distance-frequency domain signal, the method further includes: The moving target display algorithm is used to suppress static noise in the range frequency domain signal.

6. The vital sign detection method based on millimeter-wave radar according to claim 1, characterized in that, The step of extracting the phase information at the location to obtain the phase signal includes: Extract the phase information at the specified location to obtain the initial phase signal; The phase signal is obtained by using phase unwinding to correct the jump signal in the initial phase signal and eliminating the DC component deviation in the initial phase signal.

7. The vital sign detection method based on millimeter-wave radar according to claim 1, characterized in that, The step of performing spectral analysis on the distance-frequency domain signal and locating the position of the object to be detected based on the analysis results includes: The distance-frequency domain signal is subjected to spectral analysis based on the ordered statistical constant false alarm rate (OS-CFAR) algorithm, and the location of the object to be detected is located based on the analysis results.

8. A vital sign detection device based on millimeter-wave radar, characterized in that, include: The signal transformation unit is used to perform analog-to-digital conversion sampling on the radar echo signal to obtain an intermediate frequency signal, and to perform a fast Fourier transform on the intermediate frequency signal in the range dimension to obtain a range frequency domain signal. The signal extraction unit is used to perform spectral analysis on the distance frequency domain signal, locate the position of the object to be detected based on the analysis results, extract the phase information at the position, and obtain the phase signal. The signal filtering unit is used to filter the phase signal through a bandpass filter to obtain the respiratory frequency band signal and the initial heartbeat frequency band signal; The harmonic cancellation unit is used to perform real-time respiratory rate estimation based on the respiratory frequency band interval signal to obtain the respiratory frequency of the subject to be detected, and to use an adaptive notch filter to eliminate respiratory harmonics in the initial heartbeat frequency band interval signal based on the respiratory frequency to obtain the heartbeat frequency band interval signal. The signal decomposition unit is used to decompose the heartbeat frequency band signal with the minimum fuzzy entropy as the optimization objective and the variational mode decomposition algorithm to obtain the heartbeat signal. Based on the heartbeat signal, the real-time heart rate is estimated to obtain the real-time heart rate of the object to be detected.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the vital signs detection method based on millimeter-wave radar as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the vital signs detection method based on millimeter-wave radar as described in any one of claims 1 to 7.

Images