A method and device for vital sign monitoring based on longitudinal heart impact signals

Abstract

The application discloses a vital sign monitoring method and device based on longitudinal heart impact signals, and comprises the following steps: deploying the position of the beam direction of the radio frequency sensing device and the longitudinal axis of the target human body; collecting displacement data of the longitudinal heart impact signals through the radio frequency sensing device, and pre-processing the collected displacement data of the longitudinal heart impact signals to realize target human body positioning; performing target human body angle estimation and adaptive beam forming on the pre-processed displacement data to obtain an enhanced signal; performing static elimination and phase extraction on the enhanced signal to obtain a continuous displacement mapping sequence; using variational mode decomposition to adaptively separate the respiratory and heartbeat waveforms from the continuous displacement mapping sequence; and calculating vital sign parameters based on the separated respiratory and heartbeat waveforms. Therefore, the application can realize high-precision, robust and long-term monitoring of human body respiration and heartbeat without the human body carrying any sensor.

Description

Technical Field

[0001] This invention relates to the field of intelligent non-contact sensing technology, specifically to a method and device for monitoring vital signs based on longitudinal cardiac impact signals (Ballistocardiogram, BCG). Background Technology

[0002] Cardiovascular disease is one of the leading causes of death worldwide, and early screening and long-term cardiac health monitoring are crucial for preventing serious complications and improving patient outcomes. In particular, continuous vital sign monitoring during sleep can capture pathological patterns that are difficult to detect in short-term daytime observations, such as bradycardia, sleep-related atrial fibrillation, or abnormal heart rate variability (HRV) patterns. Currently, the gold standard for cardiac monitoring in clinical practice is electrocardiography (ECG), but it requires users to wear multiple electrode pads, resulting in significant discomfort. While wearable devices such as smartwatches and wristbands based on photoplethysmography (PPG) offer more portable options, prolonged wear can still cause skin discomfort and requires frequent charging, adding an extra burden to usage.

[0003] Non-contact radio frequency (RF) sensing technologies, such as millimeter wave (mmWave) and ultra-wideband (UWB), have effectively overcome the physical constraints of traditional contact sensors due to their inherent advantages such as high precision, low cost, and privacy protection. They have become a research hotspot in the field of physiological parameter monitoring and show significant clinical application prospects. Sensing devices can capture sub-millimeter-level skin surface displacements caused by respiration and heartbeat by sending signals and analyzing the echoes reflected from the human body, thereby extracting vital signs and physiological indicators such as respiration and heartbeat.

[0004] However, existing radio frequency sensing heartbeat monitoring technologies all rely on pointing the radio frequency sensing device at the target's chest area, attempting to capture the radial vibrations of the chest wall directly caused by heartbeats. However, this commonly used method faces two long-standing and insurmountable technical challenges in practical applications: 1) Strong respiratory interference: The chest wall displacement caused by respiratory movements (typically 1-12 mm) is much greater than the displacement caused by heartbeats (typically 0.1-0.5 mm), resulting in an extremely low signal-to-interference ratio (SIR) for the heartbeat signal. Due to the extremely low SIR and the nonlinear superposition of respiratory and heartbeat harmonics, it is difficult to separate the weak heartbeat signal from the mixed signal using conventional bandpass filtering or mode decomposition algorithms. When breathing is vigorous, the heartbeat signal is often completely submerged, leading to a significant decrease in monitoring accuracy. 2) High sensitivity to sleep posture: When the user is in a side-lying or prone position, the chest is no longer directly facing the radar device, resulting in a significant attenuation of the captured cardiac signal energy and a substantial decrease in the SIR. This dependence on posture severely hinders the monitoring stability of the system in sleep scenarios.

[0005] In summary, existing non-contact vital sign monitoring technologies still have significant limitations in handling respiratory disturbances and postural changes. There is an urgent need for a new sensing paradigm and method to achieve a robust vital sign monitoring scheme with a high signal-to-interference ratio across various sleep positions. Summary of the Invention

[0006] The present invention aims to at least solve one of the technical problems existing in the prior art. Therefore, in response to the above-mentioned problems, the object of the present invention is to provide a method and apparatus for monitoring vital signs based on longitudinal cardiac impact signals, capable of achieving robust vital sign monitoring with a high signal-to-interference ratio under various postures.

[0007] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows:

[0008] In a first aspect, the vital signs monitoring method based on longitudinal cardiac impact signals provided by the present invention is characterized by comprising: The position of the beam direction of the deployed radio frequency sensing equipment relative to the longitudinal axis of the target human body; Displacement data of longitudinal cardiac impact signals are collected by radio frequency sensing equipment, and the displacement data of longitudinal cardiac impact signals are preprocessed to achieve target human body positioning. Enhanced signals are obtained by estimating the target human body angle and adaptive beamforming using preprocessed displacement data. Static cancellation and phase extraction are performed on the enhanced signal to obtain a continuous displacement mapping sequence; The respiratory and heartbeat waveforms are adaptively separated using variational mode decomposition on a continuous displacement mapping sequence; Vital characteristic parameters are calculated based on separate respiratory and heartbeat waveforms.

[0009] In some possible implementations, the process of positioning the beam direction of the radio frequency sensing device relative to the longitudinal axis of the target human body is as follows: Set the direction of the main beam emitted by the radio frequency sensing device to be parallel to the longitudinal axis of the target human body, or the angle between the two is within a preset threshold range. Setting up typical deployment scenarios ensures that the detection beam and the target human body form a predetermined positional relationship, avoiding the impact of changes in the human body's sleeping posture on the radio frequency sensing device's monitoring of the human body's longitudinal displacement.

[0010] Some possible implementations involve setting up typical deployment scenarios such that the detection beam and the target human body form a predetermined positional relationship, including: Bedside scenario: The radio frequency sensing device is fixed at the head of the bed, and the detection beam shines downward along the longitudinal axis from above the target human head to capture the longitudinal displacement signal of the target human head and shoulders; Bed foot scenario: The radio frequency sensing device is fixed at the foot of the bed, and the detection beam shines upward along the longitudinal axis from the foot to capture the longitudinal displacement signal of the lower body of the target human body, such as the feet, legs or buttocks. Seated scenario: For seated targets, radio frequency sensing devices are deployed under the seat to illuminate upwards or above the seat to illuminate downwards, so that the beam direction coincides with or is parallel to the longitudinal direction of the spine of the seated human body, in order to capture the longitudinal displacement signal of the human hips or head and shoulders.

[0011] In some possible implementations, displacement data of longitudinal cardiac impact signals is acquired using radio frequency sensing devices, and the acquired displacement data of longitudinal cardiac impact signals is preprocessed to achieve target human body localization, including: 1) Preprocessing of data collected by radio frequency sensing devices with distance resolution capability is performed as follows: Obtaining the signal matrix: The radio frequency sensing device samples and processes the received signal to obtain the M matrix. N A signal matrix of dimension P, where M is the size of the fast time dimension, N is the size of the slow time dimension, and P is the number of corresponding receiving antenna channels; Target range bin localization: Calculate the dynamic energy of each range bin within the time window, select the range bin with the highest energy as the range bin containing the target human body, and output the corresponding N. P-signal matrix; 2) Preprocessing of data collected by radio frequency sensing devices lacking distance resolution capability: The collected channel state information is represented as a complex matrix containing multiple subcarriers, with dimension M. N P, where M corresponds to the number of subcarriers, N corresponds to the number of slow time dimension CSIs, and P corresponds to the number of receive antenna channels. Selecting a specific subcarrier, the corresponding N... P-signal matrix output.

[0012] In some possible implementations, the enhanced signal is obtained by estimating the target human body angle and adaptive beamforming using preprocessed displacement data. The process is as follows: For each slow-time dimension sample, the signal is beamformed onto each candidate two-dimensional angle: for a two-dimensional antenna array, a two-dimensional digital beamforming algorithm is used, and for the t-th sampling point, the azimuth angle is calculated. and pitch angle power spectrum ,in, As a guiding vector, For a multi-antenna receiving vector; for a sensing device with only a one-dimensional antenna array, calculate the one-dimensional azimuth angle spectrum. ; Estimate the spatial angle of the target human body for subsequent signal enhancement at that angle: When using a two-dimensional antenna array, the angle of the human body is estimated by frequency accumulation in the time dimension as follows:

[0013] Where T represents the total number of sampling points in the time window, This indicates the accumulation of energy within the frequency range of heartbeat harmonics; Using a defined target angle An optimal beamforming weight vector is constructed for beamforming, and the original signal is weighted and summed to obtain a complex signal with enhanced signal-to-noise ratio. .

[0014] In some possible implementations, static cancellation and phase extraction are performed on the enhanced signal to obtain a continuous displacement mapping sequence, the process of which is as follows: Static cancellation: Static cancellation is performed on complex signals with enhanced signal-to-noise ratio using a fitted circle algorithm or a difference algorithm. ; right Phase calculation and unwinding: Extracting the phase of the complex signal after static elimination. :

[0015] right Phase unwrapping is performed to obtain a continuous displacement mapping sequence. :

[0016] in, The wavelength of the electromagnetic wave in the sensing device.

[0017] In some possible implementations, variational mode decomposition is used to adaptively separate the respiratory and heartbeat waveforms of a continuous displacement mapping sequence, as follows: The displacement sequence is processed using the variational mode decomposition algorithm. Decomposed into K intrinsic mode functions (IMFs); The respiratory component was selected as the IMF with the lowest frequency and the largest energy percentage. ; The autocorrelation function of each IMF is calculated, and its peak intensity within the heart rate delay range is retrieved. If the autocorrelation peak exceeds a threshold, the IMF is selected as the cardiac component. All identified cardiac components are summed to obtain the enhanced reconstructed heartbeat signal. .

[0018] In some possible implementations, the calculation of vital sign parameters includes: Respiratory rate calculation: This involves calculating the extracted respiratory components. Perform zero-crossing detection or spectral peak search to obtain respiratory rate; Heart rate index calculation: The heart rate characteristic peak is located using a bidirectional adaptive search algorithm, the heart rate interval time series is estimated, and various heart rate and heart rate variability indices are calculated.

[0019] In some possible implementations, the process of locating heartbeat characteristic peaks based on a bidirectional adaptive search algorithm is as follows: 1) Algorithm initialization and parameter preparation: Input data: Obtain heartbeat waveform ; Coarse heart rate acquisition: through waveform analysis Perform FFT or autocorrelation analysis to estimate the target's rough heart rate. ; Define parameters: based on coarse heart rate Obtain the desired heart rate interval Search tolerance window ; Identify heartbeat waveforms Construct a candidate peak set by finding all local maxima greater than the mean. ; 2) Anchor point selection: From candidate combinations The point with the largest amplitude is selected as the initial anchor point. Store the time index corresponding to that point in the final heartbeat location set. Using this point as a reference, a recursive search will be performed simultaneously in both the forward and reverse time directions. 3) Bidirectional adaptive recursive search: For each search direction Execute the following iteration logic: Target location prediction: Set the current reference peak value as... Current search interval Predicting the possible timing of the next peak : ; Neighborhood peak matching: from candidate set Find the point closest to the predicted location. ; Adaptive Admission and Update: Determining Prediction Bias Is it smaller than the tolerance window? If the constraints are satisfied, then it is considered that... To identify the true heartbeat characteristic peak, add its corresponding time index. Subsequently, the current search interval is adaptively updated. The actual distance between two currently known adjacent peaks, i.e. and order Continue the next round of search; if the constraints are not met, it is considered that the signal quality of the direction is insufficient or the search is completed, and the current loop is exited.

[0020] Secondly, the present invention also provides a vital signs monitoring device based on longitudinal cardiac impact signals, comprising: The data acquisition unit is configured to deploy the beam direction of the radio frequency sensing device and the position of the longitudinal axis of the target human body; it acquires displacement data of the longitudinal cardiac impact signal through the radio frequency sensing device and preprocesses the acquired displacement data of the longitudinal cardiac impact signal to achieve target human body positioning. The data processing unit is configured to perform target human body angle estimation and adaptive beamforming to obtain an enhanced signal using preprocessed displacement data; The displacement sequence acquisition unit is configured to perform static elimination and phase extraction on the enhanced signal to obtain a continuous displacement mapping sequence. The mode decomposition unit is configured to adaptively separate respiratory and heartbeat waveforms from a continuous displacement mapping sequence using variational mode decomposition. The characteristic parameter calculation unit is configured to calculate vital signs parameters based on separate respiratory and heartbeat waveforms.

[0021] Because the present invention adopts the above technical solution, it has the following characteristics: 1. This invention can achieve high-precision, robust and long-term monitoring of human respiration and heartbeat without requiring the human body to carry any sensors.

[0022] 2. This invention proposes to change the position of the radio frequency sensing device, making the main beam direction of the radio frequency sensing device parallel to or at an angle within a preset range to the longitudinal axis of the target human body in the head-to-toe direction, thereby capturing the significant recoil displacement in the longitudinal direction caused by the heart's ejection of blood. The obtained displacement signal has three characteristics: High signal-to-interference ratio (SIIR): Because the direction of chest wall movement caused by respiration is orthogonal or nearly orthogonal to the longitudinal direction of the human body with significant cardiac impact signals, respiratory interference is greatly suppressed at the physical level, thus improving the SIIR of the heartbeat signal. Robust to different sleeping positions: Since changes in sleeping positions such as lying flat, on one's side, or prone do not affect the sensing device's monitoring of the longitudinal displacement of the human head and feet, this deployment method can robustly cope with changes in sleeping positions. It exhibits significant periodic morphological characteristics: within each heartbeat cycle, the waveform signal consists of a main characteristic peak and several smaller peaks before and after it. By searching and locating these main characteristic peaks, accurate estimation of the heartbeat interval and heartbeat variability indicators can be achieved.

[0023] 3. This invention can effectively separate and enhance respiratory and heartbeat signals, thereby improving sensing accuracy.

[0024] In summary, this invention overcomes the two long-standing challenges of respiratory interference and changes in sleep posture, enabling accurate, long-term, and robust monitoring of human vital signs. The provided respiratory and heart rate waveforms can be widely used in sleep monitoring, sleep quality assessment, early disease screening, disease detection, and postoperative recovery tracking. Attached Figure Description

[0025] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. In the drawings: Figure 1 This is a flowchart of a vital signs monitoring method based on longitudinal cardiac impact signals according to an embodiment of the present invention; Figure 2 To illustrate the significance of the recoil signal after a heartbeat in the longitudinal direction, this embodiment of the invention displays a human vascular tree, blood flow direction, and human micro-vibration direction.

[0026] Figure 3 This embodiment of the invention demonstrates a typical bedside deployment scenario in which a radio frequency sensing device is fixed at the head of the bed, and a detection beam shines downward along the longitudinal axis from above the human head to capture the longitudinal micro-displacement of the head and shoulders.

[0027] Figure 4 This is a typical deployment scenario at the foot of the bed, where a radio frequency sensing device is fixed to the foot of the bed and a detection beam shines upwards along the longitudinal axis from the foot of the bed, as shown in this embodiment of the invention.

[0028] Figure 5 This embodiment of the invention demonstrates a seating deployment scenario in which a radio frequency sensing device is deployed under a seat to illuminate the seat, such that the direction of the sensing beam coincides with or is parallel to the longitudinal direction of the spine of a seated person.

[0029] Figure 6 This embodiment of the invention demonstrates a seating deployment scenario in which a radio frequency sensing device is deployed above a seat and illuminates downwards, such that the direction of the sensing beam coincides with or is parallel to the longitudinal direction of the spine of a seated person. Detailed Implementation

[0030] It should be understood that the terminology used herein is for the purpose of describing particular exemplary embodiments only and is not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “described” as used herein may also include the plural forms. The terms “comprising,” “including,” “containing,” and “having” are inclusive and therefore indicate the presence of the stated features, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, elements, components, and / or combinations thereof. The method steps, processes, and operations described herein are not construed as requiring them to be performed in a particular order described or illustrated unless the order of performance is explicitly indicated. It should also be understood that additional or alternative steps may be used.

[0031] Although terms such as first, second, third, etc., may be used in this document to describe multiple elements, components, regions, layers, and / or segments, these elements, components, regions, layers, and / or segments should not be limited by these terms. These terms may be used only to distinguish one element, component, region, layer, or segment from another. Unless the context clearly indicates otherwise, terms such as "first," "second," and other numerical terms used herein do not imply order or sequence. Therefore, the first element, component, region, layer, or segment discussed below may be referred to as the second element, component, region, layer, or segment without departing from the teachings of the exemplary embodiments.

[0032] For ease of description, spatial relative terms may be used in the text to describe the relationship of one element or feature relative to another element or feature as shown in the figure. These relative terms include, for example, "inside," "outside," "middle," "outer," "below," "above," etc. Such spatial relative terms are intended to include different orientations of the device in use or operation, other than those depicted in the figure.

[0033] Existing non-contact vital sign monitoring technologies still have significant limitations in handling respiratory interference and postural changes. This invention provides a vital sign monitoring method and apparatus based on longitudinal cardiac impact signals, comprising: deploying a radio frequency sensing device with its beam direction aligned with the longitudinal axis of the target human body; acquiring displacement data of the longitudinal cardiac impact signal through the radio frequency sensing device, and preprocessing the acquired displacement data to locate the target human body; estimating the target human body angle and adaptively beamforming the preprocessed displacement data to obtain an enhanced signal; performing static elimination and phase extraction on the enhanced signal to obtain a continuous displacement mapping sequence; adaptively separating the respiratory and heartbeat waveforms from the continuous displacement mapping sequence using variational mode decomposition; and calculating vital sign parameters based on the separated respiratory and heartbeat waveforms. Therefore, this invention can effectively suppress respiratory interference and acquire high-quality vital sign signals under various sleeping positions to support applications such as heartbeat detection, respiratory detection, sleep monitoring, early disease screening, and disease detection.

[0034] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the invention and to fully convey the scope of the invention to those skilled in the art.

[0035] Example 1: As Figure 1 As shown, the vital signs monitoring method based on longitudinal cardiac impact signals provided in this embodiment includes: S1. Deployment of sensing devices and construction of spatial geometric constraints: Deploy radio frequency sensing devices so that their beam direction is approximately parallel to the longitudinal axis of the human body (parallel or within a preset angle) to capture longitudinal cardiac impact signal displacement data of the human body.

[0036] In this embodiment, the deployment method of the radio frequency sensing device relative to the human body is changed from the traditional "chest alignment" to "human body longitudinal axis alignment".

[0037] Furthermore, the specific implementation process of deploying radio frequency sensing devices and constructing spatial geometric constraints is as follows: S11, Construction of Physical Criteria: In this embodiment, radio frequency sensing equipment is used to capture the recoil (recoil) displacement data of the human body caused by the momentum change of blood flow during the heart's pumping process, which is then used to extract physiological parameters related to cardiac function. This recoil signal caused by the heart's pumping of blood is called the ballistocardiogram (BCG).

[0038] like Figure 2As shown, based on human anatomy, since the major arteries of the human body are distributed along the head-to-toe longitudinal axis, the mechanical recoil force caused by the heart's ejection of blood also primarily propagates along this direction. This invention defines the direction of the head-to-toe line as the longitudinal axis of the human body. To monitor significant recoil displacement data in this longitudinal direction, the main beam direction (i.e., the radial direction) emitted by the radio frequency sensing device must be parallel to the longitudinal axis of the target human body, or at an angle between them. The size is preferably within a preset threshold range. .

[0039] S12. Set up typical deployment scenarios to ensure that the detection beam and the target human body form a set positional relationship, so as to avoid the impact of changes in the human body's sleeping posture on the radio frequency sensing device's monitoring of the human body's displacement in the longitudinal direction.

[0040] In this embodiment, a typical deployment scenario is set, including: Scene A (Bedside Scene): For example Figure 3 As shown, the radio frequency sensing device is fixed at the head of the bed, and the detection beam shines downward along the longitudinal axis from above the human head to capture the longitudinal displacement signal of the head and shoulders.

[0041] Scene B (Bedside Scene): such as Figure 4 As shown, the radio frequency sensing device is fixed at the foot of the bed, and the detection beam shines upward along the longitudinal axis from the soles of the feet to capture the longitudinal displacement signals of the lower body, including the feet, legs, and buttocks.

[0042] Scenario C (Seated Scenario): For the seated target, deploy the radio frequency sensing device under the seat (irradiating upwards, such as...) Figure 5 ) or above the seat (shining downwards, such as Figure 6 This allows the direction of the detection beam to coincide with or be parallel to the longitudinal direction of the spine of a seated human body, in order to capture longitudinal displacement signals of the buttocks (the device is located under the seat) or head and shoulders (the device is located above the seat).

[0043] Therefore, the above deployment method ensures that the dominant micro-motion component captured by the radio frequency signal is the longitudinal BCG signal. Along this axis, the direction of chest wall displacement caused by respiration is orthogonal or nearly orthogonal to the radial direction of the beam, thus significantly suppressing respiratory interference at the physical level and improving the signal-to-interference ratio of the heartbeat signal. Since changes in sleeping posture, such as lying flat, on one's side, or prone, do not affect the radio frequency sensing device's monitoring of the longitudinal displacement of the head and toes, this deployment method can robustly handle changes in sleeping posture.

[0044] S2. Data Acquisition and Preprocessing: Acquire displacement data of longitudinal cardiac impact signal and preprocess the acquired displacement data to achieve target human body positioning.

[0045] In this embodiment, for radio frequency sensing devices with different hardware capabilities, this embodiment provides a compatible data acquisition and preprocessing method, including: A. For equipment with range resolution capabilities, such as millimeter-wave radar and UWB radar, the acquisition and preprocessing process is as follows: 1) Obtaining the signal matrix: The radio frequency sensing device samples the received signal to obtain the original ADC analog-to-digital conversion sampling data, forming the M matrix. N A signal matrix with P dimensions, where M is the size of the fast time dimension (corresponding to the number of sampling points in a chirp or pulse), N is the size of the slow time dimension (corresponding to the number of transmitted chirs or pulses), and P corresponds to the (equivalent) number of receiving antenna channels. For FMCW (Frequency Modulated Continuous Wave) signals, an FFT operation is performed on the M sampling points in the fast time dimension to obtain the range domain signal. The range domain signal performs distance-dimensional resolution and separation of the received signal, which can be used to select the signal within the range where the human body is located, while filtering out interference and noise at other distances.

[0046] 2) Target range bin localization: Calculate the dynamic energy of each range bin within the time window, select the range bin with the highest energy as the range bin where the human body is located, and output the corresponding N. P-signal matrix.

[0047] B. For devices lacking distance resolution capabilities, such as Wi-Fi and narrowband radio frequency devices, the collected Channel State Information (CSI) is represented as a complex matrix containing multiple subcarriers, with dimension M. N P and M correspond to the number of subcarriers, N corresponds to the number of slow time dimension CSIs, and P corresponds to the number of receive antenna channels. A specific subcarrier can be selected, and its corresponding N... P-signal matrix output.

[0048] S3, Angle Estimation and Adaptive Beamforming: Preprocessed N The P-signal matrix contains angular information from multiple antenna channels, enabling target human body angle estimation and adaptive beamforming to obtain enhanced signals.

[0049] In this embodiment, to further enhance the reflection intensity of the target location and suppress environmental multipath interference, a multi-antenna array is used to perform spatial filtering on the preprocessed signal to achieve angle estimation and adaptive beamforming. The specific process is as follows: A. Angle spectrum construction: For each slow time dimension sampling, the signal beamforming is applied to each candidate two-dimensional angle.

[0050] For a two-dimensional antenna array, a two-dimensional digital beamforming algorithm is used to calculate the azimuth angle for the t-th sampling point in the slow time dimension. and pitch angle The power spectrum is as follows:

[0051] in, As a guiding vector, This is the multi-antenna receive vector.

[0052] For sensing devices with only a one-dimensional antenna array, only the one-dimensional azimuth angle spectrum is calculated. .

[0053] B. Target Angle Estimation: Estimate the spatial angle of the human target for subsequent signal enhancement at that angle.

[0054] Because human vital signs are periodic, when using a two-dimensional antenna array, this embodiment estimates the angle of the target human body by accumulating frequency over time:

[0055] Where T represents the total number of sampling points in the time window, This indicates the accumulation of energy within the frequency range of heartbeat harmonics (preferably 1.5Hz to 10Hz).

[0056] C. Enhanced beam pointing: Using a defined target angle An optimal beamforming weight vector is constructed for beamforming, and the original signal is weighted and summed to obtain a complex signal with enhanced signal-to-noise ratio. To achieve the angle at which the beam is pointed towards the target. The purpose is to enhance signal reception in that specific direction while suppressing signal energy at other angles.

[0057] S4. Static cancellation and phase extraction: For complex signals with enhanced signal-to-noise ratio. Static elimination and phase extraction are performed to obtain a continuous displacement mapping sequence.

[0058] In this embodiment, static elimination and phase extraction include: A. Static elimination.

[0059] In this embodiment, two methods can be used to eliminate static components: the fitted circle algorithm and the difference algorithm.

[0060] For the circle fitting algorithm, since the subtle movements of the target vital signs appear as arcs in the IQ complex plane, the circle fitting algorithm must first be used to calculate the center of the circle. And perform subtraction: ; In the formula, This corresponds to the human body's micro-motion signals after static elimination.

[0061] For the difference algorithm, the signal after difference is represented as: .

[0062] B. Phase calculation and unwinding.

[0063] Extracted complex signal phase as follows:

[0064] right Perform phase unwinding to solve Phase ambiguity problem, obtaining a continuous displacement mapping sequence :

[0065] in, The electromagnetic wave wavelength of the radio frequency sensing device.

[0066] S5. Extraction of respiratory and heartbeat waveforms based on variational mode decomposition, i.e., based on continuous displacement mapping sequences. Variational mode decomposition is used to adaptively separate respiratory and heartbeat waveforms.

[0067] In this embodiment, the extraction of respiratory and heartbeat waveforms based on variational mode decomposition includes: A. Adaptive Decomposition: Using the Variational Mode Decomposition (VMD) algorithm to map continuous displacement sequences It is decomposed into K intrinsic mode functions (IMFs). In this embodiment, it is preferred to set... .

[0068] B. Respiratory waveform extraction: Select the IMF (e.g., IMF 7) with the largest energy proportion within the respiratory frequency range as the respiratory component. .

[0069] C. Heartbeat Waveform Extraction: Calculate the autocorrelation result of each IMF and retrieve its peak intensity within the heart rate delay range (e.g., 0.4s to 1.25s, corresponding to 48-150 bpm). If the autocorrelation peak exceeds a threshold (e.g., 0.3), the IMF (e.g., IMF4 to IMF6) is selected as the cardiac component. Sum all the determined heartbeat components to obtain the enhanced reconstructed heartbeat signal. .

[0070] S6. Calculation of vital signs parameters, namely: using a two-way adaptive search algorithm to locate the heartbeat characteristic peak and calculate respiratory rate, heart rate and heart rate variability index, etc.

[0071] In this embodiment, the calculation of vital sign parameters includes: A. Respiratory rate calculation: Calculate the extracted respiratory components. The respiratory rate is obtained by performing zero-pass detection or spectral peak search.

[0072] B. Heartbeat Index Calculation: Heartbeat signal waveforms with high signal-to-interference ratio and significant periodic morphological characteristics were obtained through a specific spatial deployment paradigm. This is characterized by a waveform consisting of a main characteristic peak (heartbeat characteristic peak) and several smaller peaks before and after each heartbeat cycle. Therefore, by searching and locating these heartbeat characteristic peaks, the heartbeat interval time series can be estimated, and various heart rate and heart rate variability (HRV) indicators can be calculated.

[0073] Specifically, this embodiment proposes the following process for locating heartbeat characteristic peaks based on a bidirectional adaptive search algorithm: 1. Algorithm initialization and parameter preparation: Input data: Obtain heartbeat waveform Its sampling rate .

[0074] Coarse heart rate acquisition: through waveform analysis Perform FFT or autocorrelation analysis to estimate the target's rough heart rate. Among them, rough heart rate The existing formula will be used, and no restrictions will be imposed here.

[0075] Parameter definition: Expected heart rate interval : Search tolerance window : , To control the coefficient for the tolerance window size, the present invention preferably sets it to... .

[0076] Identification Construct a candidate peak set by finding all local maxima greater than the mean. .

[0077] 2. Anchor point selection.

[0078] From candidate peak combination Select the point with the largest amplitude as the initial anchor point: :

[0079] Store the time index corresponding to this point in the final heartbeat location set. Using this point as a reference, a recursive search will be performed simultaneously in both forward and reverse time directions.

[0080] 3. Bidirectional adaptive recursive search.

[0081] For each search direction Execute the following iteration logic: Target location prediction: Set the current reference peak value as... (Initially) ), current search interval (Initially) Predicting the likely timing of the next peak. : ; Neighborhood peak matching: from candidate peak set Find the point closest to the predicted location. .

[0082] Adaptive Admission and Update: Determining Prediction Bias Is it smaller than the tolerance window? If the constraints are satisfied, then it is considered... For the true heartbeat characteristic peak, its corresponding time index is... join in Then, the current search interval is adaptively updated. The actual distance between two currently known adjacent peaks, i.e. and order Continue the next round of search. If the constraints are not met, it is assumed that the signal quality of the direction is insufficient or the search is complete, and the current loop is exited.

[0083] Furthermore, IBI (interval between heartbeats) sequence calculation: [The text abruptly ends here, likely due to an incomplete sentence or a formatting error.] Arranged in ascending order, the IBI sequence is calculated as follows:

[0084] in, For set The k-th element after sorting in ascending order. Further, heart rate calculation:

[0085] in, This represents the number of elements in the IBI sequence.

[0086] Furthermore, the HRV index is calculated as follows: Based on the IBI sequence, the RMSSD (root mean square of the difference between adjacent intervals), SDRR (standard deviation), and pNN50 are further calculated. The specific calculation formulas are existing technologies and will not be elaborated here.

[0087] Example 2: Following the method for monitoring vital signs based on longitudinal cardiac impact signals provided in Example 1, this example provides a device for monitoring vital signs based on longitudinal cardiac impact signals. The device provided in this example can implement the method for monitoring vital signs based on longitudinal cardiac impact signals as described in Example 1. This device can be implemented through software, hardware, or a combination of both. For ease of description, this example is described by dividing the functionality into various units. Of course, in implementation, the functions of each unit can be implemented in one or more software and / or hardware components. For example, the device may include integrated or separate functional modules or units to perform the corresponding steps in the methods of Example 1. Since the device in this example is basically similar to the method example, the description process of this example is relatively simple. Relevant details can be found in the description of Example 1. The embodiment of the vital sign monitoring device based on longitudinal cardiac impact signals provided by this invention is merely illustrative.

[0088] Specifically, this embodiment also provides a vital signs monitoring device based on longitudinal cardiac impact signals, comprising: The data acquisition unit is configured to deploy the beam direction of the radio frequency sensing device and the position of the longitudinal axis of the target human body; it acquires displacement data of the longitudinal cardiac impact signal through the radio frequency sensing device and preprocesses the acquired displacement data of the longitudinal cardiac impact signal to achieve target human body positioning. The data processing unit is configured to perform target human body angle estimation and adaptive beamforming to obtain an enhanced signal using preprocessed displacement data; The displacement sequence acquisition unit is configured to perform static elimination and phase extraction on the enhanced signal to obtain a continuous displacement mapping sequence. The mode decomposition unit is configured to adaptively separate respiratory and heartbeat waveforms from a continuous displacement mapping sequence using variational mode decomposition. The characteristic parameter calculation unit is configured to calculate vital signs parameters based on separate respiratory and heartbeat waveforms.

[0089] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In the description of this specification, the terms "a preferred embodiment," "furthermore," "specifically," "in this embodiment," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the embodiments in this specification. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described can be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0090] 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 monitoring method based on longitudinal heart impact signals, characterized in that, include: The position of the beam direction of the deployed radio frequency sensing equipment relative to the longitudinal axis of the target human body; Displacement data of longitudinal cardiac impact signals are collected by radio frequency sensing equipment, and the displacement data of longitudinal cardiac impact signals are preprocessed to achieve target human body positioning. Enhanced signals are obtained by estimating the target human body angle and adaptive beamforming using preprocessed displacement data. Static cancellation and phase extraction are performed on the enhanced signal to obtain a continuous displacement mapping sequence; The respiratory and heartbeat waveforms are adaptively separated using variational mode decomposition on a continuous displacement mapping sequence; Vital characteristic parameters are calculated based on separate respiratory and heartbeat waveforms.

2. The longitudinal heart-shock signal based vital sign monitoring method of claim 1, wherein, The process of determining the beam direction of the radio frequency sensing device relative to the longitudinal axis of the target human body is as follows: Set the direction of the main beam emitted by the radio frequency sensing device to be parallel to the longitudinal axis of the target human body, or the angle between the two is within a preset threshold range. Setting up typical deployment scenarios ensures that the detection beam and the target human body form a predetermined positional relationship, avoiding the impact of changes in the human body's sleeping posture on the radio frequency sensing device's monitoring of the human body's longitudinal displacement.

3. The longitudinal heart-shock signal based vital sign monitoring method of claim 2, wherein, Typical deployment scenarios are set up to establish a predetermined positional relationship between the detection beam and the target human body, including: Bedside scenario: The radio frequency sensing device is fixed at the head of the bed, and the detection beam shines downward along the longitudinal axis from above the target human head to capture the longitudinal displacement signal of the target human head and shoulders; Bed foot scenario: The radio frequency sensing device is fixed at the foot of the bed, and the detection beam shines upward along the longitudinal axis from the foot to capture the longitudinal displacement signal of the lower body of the target human body, such as the feet, legs or buttocks. Seated scenario: For seated targets, radio frequency sensing devices are deployed under the seat to illuminate upwards or above the seat to illuminate downwards, so that the beam direction coincides with or is parallel to the longitudinal direction of the spine of the seated human body, in order to capture the longitudinal displacement signal of the human hips or head and shoulders.

4. The longitudinal heart-shock signal based vital sign monitoring method of claim 2, wherein, Displacement data of longitudinal cardiac impact signals are acquired using radio frequency sensing equipment, and the acquired displacement data of longitudinal cardiac impact signals is preprocessed to achieve target human body localization, including: 1) Preprocessing of data collected by radio frequency sensing devices with distance resolution capability is performed as follows: Obtaining of signal matrix: the radio frequency sensing device samples and processes the received signal to obtain M N a signal matrix of P dimensions, wherein M is a fast time dimension size, N is a slow time dimension size, and P is a number of corresponding receiving antenna channels. Target range bin positioning: calculate the dynamic energy of each range bin in the time window, select the range bin with the highest energy as the range bin where the target human body is located, and output the corresponding N P signal matrix; 2) Preprocessing of data collected by radio frequency sensing devices lacking distance resolution capability: The collected channel state information is represented as a complex matrix containing multiple subcarriers, with dimension M. N P, where M corresponds to the number of subcarriers, N corresponds to the number of slow time dimension CSIs, and P corresponds to the number of receive antenna channels. Selecting a specific subcarrier, the corresponding N... P-signal matrix output.

5. The vital signs monitoring method based on longitudinal cardiac impact signals according to claim 2, characterized in that, The enhanced signal is obtained by estimating the target human body angle and adaptive beamforming using preprocessed displacement data. The process is as follows: For each slow-time dimension sample, the signal is beamformed onto each candidate two-dimensional angle: for a two-dimensional antenna array, a two-dimensional digital beamforming algorithm is used, and for the t-th sampling point, the azimuth angle is calculated. and pitch angle power spectrum ,in, As a guiding vector, For a multi-antenna receiving vector; for a sensing device with only a one-dimensional antenna array, calculate the one-dimensional azimuth angle spectrum. ; Estimate the spatial angle of the target human body for subsequent signal enhancement at that angle: When using a two-dimensional antenna array, the angle of the human body is estimated by frequency accumulation in the time dimension as follows: Where T represents the total number of sampling points in the time window, This indicates the accumulation of energy within the frequency range of heartbeat harmonics; Using a defined target angle An optimal beamforming weight vector is constructed for beamforming, and the original signal is weighted and summed to obtain a complex signal with enhanced signal-to-noise ratio. .

6. The vital signs monitoring method based on longitudinal cardiac impact signals according to claim 2, characterized in that, Static cancellation and phase extraction are performed on the enhanced signal to obtain a continuous displacement mapping sequence. The process is as follows: Static cancellation: Static cancellation is performed on complex signals with enhanced signal-to-noise ratio using a fitted circle algorithm or a difference algorithm. ; right Phase calculation and unwinding: Extracting the phase of the complex signal after static elimination. : right Phase unwrapping is performed to obtain a continuous displacement mapping sequence. : in, The wavelength of the electromagnetic wave in the sensing device.

7. The vital signs monitoring method based on longitudinal cardiac impact signals according to claim 6, characterized in that, The respiratory and heartbeat waveforms are adaptively separated using variational mode decomposition on a continuous displacement mapping sequence. The process is as follows: The displacement sequence is processed using the variational mode decomposition algorithm. Decomposed into K intrinsic mode functions (IMFs); The respiratory component was selected as the IMF with the lowest frequency and the largest energy percentage. ; The autocorrelation function of each IMF is calculated, and its peak intensity within the heart rate delay range is retrieved. If the autocorrelation peak exceeds a threshold, the IMF is selected as the cardiac component. All identified cardiac components are summed to obtain the enhanced reconstructed heartbeat signal. .

8. The vital signs monitoring method based on longitudinal cardiac impact signals according to claim 7, characterized in that, Calculation of vital signs parameters, including: Respiratory rate calculation: This involves calculating the extracted respiratory components. Perform zero-crossing detection or spectral peak search to obtain respiratory rate; Heart rate index calculation: The heart rate characteristic peak is located using a bidirectional adaptive search algorithm, the heart rate interval time series is estimated, and various heart rate and heart rate variability indices are calculated.

9. The vital signs monitoring method based on longitudinal cardiac impact signals according to claim 8, characterized in that, The process of locating heartbeat characteristic peaks based on the bidirectional adaptive search algorithm is as follows: 1) Algorithm initialization and parameter preparation: Input data: Obtain heartbeat waveform ; Coarse heart rate acquisition: through waveform analysis Perform FFT or autocorrelation analysis to estimate the target's rough heart rate. ; Define parameters: based on coarse heart rate Obtain the desired heart rate interval Search tolerance window ; Identify heartbeat waveforms Construct a candidate peak set by finding all local maxima greater than the mean. ; 2) Anchor point selection: From candidate combinations The point with the largest amplitude is selected as the initial anchor point. Store the time index corresponding to that point in the final heartbeat location set. Using this point as a reference, a recursive search will be performed simultaneously in both the forward and reverse time directions. 3) Bidirectional adaptive recursive search: For each search direction Execute the following iteration logic: Target location prediction: Set the current reference peak value as... Current search interval Predicting the possible timing of the next peak : ; Neighborhood peak matching: from candidate set Find the point closest to the predicted location. ; Adaptive Admission and Update: Determining Prediction Bias Is it smaller than the tolerance window? If the constraints are satisfied, then it is considered that... To identify the true heartbeat characteristic peak, add its corresponding time index. Subsequently, the current search interval is adaptively updated. The actual distance between two currently known adjacent peaks, i.e. and order Continue the next round of search; if the constraints are not met, it is considered that the signal quality of the direction is insufficient or the search is completed, and the current loop is exited.

10. A vital signs monitoring device based on longitudinal cardiac impact signals, characterized in that, include: The data acquisition unit is configured to be positioned relative to the longitudinal axis of the target human body, along with the beam direction of the deployed radio frequency sensing device. Displacement data of longitudinal cardiac impact signals are collected by radio frequency sensing equipment, and the displacement data of longitudinal cardiac impact signals are preprocessed to achieve target human body positioning. The data processing unit is configured to perform target human body angle estimation and adaptive beamforming to obtain an enhanced signal using preprocessed displacement data; The displacement sequence acquisition unit is configured to perform static elimination and phase extraction on the enhanced signal to obtain a continuous displacement mapping sequence. The mode decomposition unit is configured to adaptively separate respiratory and heartbeat waveforms from a continuous displacement mapping sequence using variational mode decomposition. The characteristic parameter calculation unit is configured to calculate vital signs parameters based on separate respiratory and heartbeat waveforms.

Images