Double-channel fused high-precision blood pressure measurement device and blood pressure value estimation method

Abstract

The invention discloses a double-channel fused high-precision blood pressure measurement device and a blood pressure value estimation method. The device comprises an air bag cuff, an air bag pressure sensor, an array type skin pressure sensor and a micro-processing unit, the air bag pressure sensor obtains air pressure data in the air bag, the array type skin pressure sensor obtains pulse pressure data, and the microprocessing unit is used for receiving the data and conducting fusion processing to obtain a final blood pressure estimation value. The skin pulse signal pressure sensing array is introduced on the basis of cuff pressure detection, local pulse waveforms can be directly collected, external noise interference is reduced, and signal stability is improved; besides, by means of double-channel fusion, air bag pressure signals and skin pulse array signals are combined, blood pressure estimation refers to cuff pressure changes and is combined with continuous pulse waveforms, adaptability to different individuals and high-precision blood pressure measurement are achieved, meanwhile, the recognition capacity of weak pulse signals is improved, and the measurement failure or error probability is reduced.

Description

Technical Field

[0001] This invention belongs to the field of biomedical engineering and wearable medical device technology, specifically relating to a dual-channel fusion high-precision blood pressure measurement device and a blood pressure value estimation method. Background Technology

[0002] Blood pressure is an important physiological indicator reflecting the cardiovascular health of the human body. Sustained abnormal blood pressure levels are a significant risk factor for many diseases, including hypertension, heart disease, stroke, and kidney disease. Therefore, accurate, continuous, and non-invasive blood pressure measurement is of great importance for disease prevention, clinical diagnosis, and daily health management.

[0003] Traditional blood pressure measurement methods mainly include auscultation (mercury / mercury sphygmomanometers), oscillometric methods (electronic sphygmomanometers), and catheter aspiration (CAD). While CAD provides highly accurate blood pressure data, it is invasive and typically only used in intensive care or surgical settings, limiting its clinical application. Non-invasive blood pressure measurement methods, such as cuff-based auscultation and oscillometric methods, are widely used in home and hospital settings due to their simplicity, portability, and non-invasiveness, but they still have certain limitations in practical application.

[0004] The auscultatory method relies on the operator to determine systolic and diastolic blood pressure by interpreting the start and end points of Korotkoff sounds with a stethoscope. Therefore, the measurement results are easily affected by operator experience and environmental noise. Furthermore, traditional mercury sphygmomanometers are bulky, posing a risk of mercury leakage, and are not suitable for portable or home use. The oscillometric method calculates blood pressure values ​​based on cuff oscillation signals. While convenient, its measurement accuracy is affected by individual vascular compliance and hemodynamic differences, and is easily interfered with by arm movement or muscle tension. Moreover, blood pressure calculation is highly dependent on internal algorithms and calibration data, lacking individualized adjustment methods. Therefore, although the auscultatory and oscillometric methods are widely used in non-invasive blood pressure measurement, their operational dependence and limited accuracy still restrict their application.

[0005] Auscultatory sphygmomanometers mainly consist of a cuff, a mercury sphygmomanometer, an air pump (manual or automatic inflation), a deflation valve, and a stethoscope. The cuff is typically wrapped around the upper arm and inflated to fit snugly against the brachial artery. The sphygmomanometer displays changes in cuff pressure; the stethoscope listens for Korotkoff sounds produced by blood flow; and the air pump and deflation valve control the cuff pressure. During measurement, the cuff is inflated to a level higher than the expected systolic pressure, completely blocking the brachial artery; then it is slowly deflated. When the cuff pressure drops slightly below the systolic pressure, blood begins to flow through the artery, producing the first clear Korotkoff sound, corresponding to systolic blood pressure (SBP). As the pressure decreases further, the Korotkoff sound gradually weakens until it disappears; the point of disappearance corresponds to diastolic blood pressure (DBP). Mercury sphygmomanometers display the pressure value through the mercury level, while mechanical sphygmomanometers indicate the pressure value through a pointer; the reading depends on the operator's judgment of the start and end points of the Korotkoff sounds.

[0006] An oscillometric blood pressure monitor mainly consists of a cuff, a pressure sensor, a microprocessor (MCU), an air pump and deflation valve, and a display or communication interface. The cuff is used to wrap around the upper arm and apply pressure. The pressure sensor detects changes in pressure and oscillation signals within the cuff in real time. The microprocessor is responsible for data acquisition, signal processing, and blood pressure calculation. The display or interface outputs the measurement results. Electronic blood pressure monitors are typically fully automatic or semi-automatic, controlling the inflation and deflation of the cuff via an air pump. In the oscillometric method, the cuff is first inflated to a level higher than the systolic pressure, temporarily blocking the artery. It is then slowly deflated, and the pressure sensor detects changes in the cuff pressure oscillation signal as blood flows again. The point of maximum amplitude corresponds to the mean arterial pressure (MAP). Using a pre-set empirical algorithm, the system calculates the systolic and diastolic blood pressure based on the MAP and amplitude changes. This method does not rely on a stethoscope; the measurement results are automatically calculated by an electronic algorithm. However, its accuracy is affected by individual vascular compliance, hemodynamic differences, and motion interference.

[0007] Existing technologies also include electronic blood pressure measurement methods that do not rely on manual interpretation of Korotkoff sounds. For example, the Hanwang Korotkoff sound electronic blood pressure monitor mainly consists of an upper arm cuff, a pressure sensor, a piezoelectric sensor, an inflation and deflation system, a control and signal processing unit, and a data transmission module. The cuff wraps around the upper arm, temporarily blocking blood flow to the brachial artery through inflation. Simultaneously, the pressure sensor monitors the pressure changes within the cuff in real time, while the piezoelectric sensor detects Korotkoff sound signals during deflation—the sound or micro-vibration produced by the reflow of blood. The inflation and deflation system controls the increase and decrease of cuff pressure, ensuring a slow and stable deflation process for accurate Korotkoff sound capture. The control and signal processing unit uses an internal microprocessor and artificial intelligence algorithms to analyze the signals collected by the pressure and piezoelectric sensors in real time, identifying the onset and disappearance times of Korotkoff sounds and calculating the corresponding systolic and diastolic blood pressure values. The data transmission module supports the display, storage, and uploading of measurement results, enabling remote monitoring and management.

[0008] Through electronic and algorithmic processing, the device can automatically identify the start and end points of Korotkoff sounds and match the cuff pressure to complete blood pressure measurement. Compared with traditional oscillometric electronic blood pressure monitors, the Hanvon Korotkoff sound electronic blood pressure monitor directly relies on the sound signal generated by blood flow to measure blood pressure, rather than indirectly calculating it through oscillation signals. This can improve measurement accuracy and adaptability to a certain extent, while also having automation, portability, and data management functions.

[0009] However, this blood pressure monitor relies on a piezoacoustic sensor to detect Korotkoff sounds, and the measurement results are easily affected by environmental noise and external vibrations, thus affecting the determination of systolic and diastolic blood pressure. Secondly, this blood pressure monitor relies on a single-channel signal for blood pressure calculation, and for individuals with differences in vascular compliance or weak pulses, signal recognition may be unstable, leading to errors in blood pressure estimation.

[0010] Therefore, the field of blood pressure measurement urgently needs a new technical solution that can combine high accuracy, simple operation, and non-invasiveness to overcome the limitations of traditional methods. Summary of the Invention

[0011] The main objective of this invention is to provide a high-precision blood pressure measurement device and a blood pressure value estimation method with dual-channel fusion, so as to overcome the shortcomings of the prior art.

[0012] To achieve the above-mentioned objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a high-precision blood pressure measurement device that integrates air cuff pressure and pulse signal dual channels, comprising: an air cuff, an air cuff pressure sensor, an array of skin pressure sensors, and a microprocessor unit. The airbag pressure sensor is connected to the airbag in the airbag cuff and is used to acquire air pressure data in the airbag. The array-type skin pressure sensor is fixed on the skin-contacting surface of the airbag cuff and is used to acquire pulse pressure data. The microprocessor unit is used to receive the air pressure data and pulse pressure data, and fuse them to obtain the final blood pressure estimate.

[0013] Secondly, the present invention also provides a method for estimating blood pressure values ​​for the aforementioned high-precision blood pressure measuring device, comprising: Acquire air pressure data and pulse pressure data in the airbag; Calculate a first data quality estimate of the air pressure data and a second data quality estimate of the pulse pressure data; Confidence weights are generated based on the first data quality estimate and the second data quality estimate; Based on the confidence weights, the final blood pressure estimate is calculated by combining the blood pressure estimates obtained from the air pressure data and pulse pressure data respectively.

[0014] Compared with the prior art, the beneficial effects of the present invention include at least the following: In the technical solution provided by this invention, by introducing a skin pulse signal pressure sensing array on the basis of cuff pressure detection, local pulse waveforms can be directly acquired, reducing external noise interference and improving signal stability. In addition, by dual-channel fusion, the cuff pressure signal and the skin pulse array signal are combined, so that blood pressure estimation not only refers to cuff pressure changes but also combines continuous pulse waveforms, achieving adaptability to different individuals and high-precision blood pressure measurement, while improving the ability to identify weak pulse signals and reducing the probability of measurement failure or error.

[0015] The above description is merely an overview of the technical solution of the present invention. In order to enable those skilled in the art to better understand the technical means of this application and to implement it in accordance with the contents of the specification, the preferred embodiments of the present invention are described below in conjunction with detailed drawings. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application 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 only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a cross-sectional structural schematic diagram of a high-precision blood pressure measuring device provided in one embodiment of the present invention; Figure 2 This is a schematic diagram of the unfolded planar structure of a high-precision blood pressure measuring device provided in one embodiment of the present invention. Detailed Implementation

[0018] In view of the shortcomings of the prior art, the inventors of this invention, through long-term research and extensive practice, have proposed the technical solution of this invention. The following will further explain and illustrate the technical solution, its implementation process, and its principles.

[0019] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0020] This invention provides a high-precision blood pressure measurement device that integrates cuff pressure and pulse signal in a dual-channel manner. The device includes a cuff, a cuff pressure sensor, an array of skin pressure sensors, and a microprocessor unit. The cuff pressure sensor is connected to the cuff in the cuff to acquire air pressure data from the cuff. The array of skin pressure sensors is fixed to the skin-contacting surface of the cuff to acquire pulse pressure data. The microprocessor unit receives the air pressure data and the pulse pressure data, and integrates them to obtain a final blood pressure estimate.

[0021] As a typical example of the above technical solution, the overall architecture of the high-precision blood pressure measuring device provided by the present invention is shown below.

[0022] A. Hardware layer: It consists of a cuff airbag and its pressure sensor channel (airbag channel) and a multi-point flexible skin-fitting pressure sensor array (skin array channel); and is equipped with a data acquisition front end, an analog-to-digital conversion module, a clock synchronization module, a microprocessor / FPGA and wireless / wired data interface, power supply and packaging structure.

[0023] B. Software / Algorithm Layer: Divided into in-array preprocessing and fusion module, airbag channel processing module, signal quality assessment module, dynamic weight adjustment module, fusion regression / estimation module, and calibration / calibration module.

[0024] C. System Workflow: First, a high-quality skin pulse channel is obtained through in-array fusion, then it is fused with the airbag channel at a higher level in the time domain / feature layer, and the weights of the two channels are adaptively adjusted according to the signal quality to output the final blood pressure estimate; the whole process supports online quality assessment and retest recommendations.

[0025] In some implementations, the array-type skin pressure sensor includes multiple sensor units arranged in an array, each of which outputs a unit pressure signal, and the pulse pressure data is obtained by preprocessing and primary fusion of the multiple unit pressure signals.

[0026] In some implementation schemes, see Figure 1 and Figure 2 As shown, the airbag cuff includes, from the outside to the inside, a restraining layer, an airbag interlayer, and a flexible support layer. The inner wall of the flexible support layer is provided with a fitting area, and the array-type skin pressure sensor is disposed in the fitting area; around the fitting area, the flexible support layer is also provided with an annular pressure isolation groove, the depth of which is more than half the thickness of the flexible support layer.

[0027] In some embodiments, the contact area is a concave structure with a depth of 1 / 5 to 1 / 3 of the array-type skin pressure sensor. Part of the thickness of the array-type skin pressure sensor is embedded in the concave structure, and the remaining thickness protrudes from the concave structure.

[0028] In some embodiments, the diameter of the distal end of the airbag cuff is smaller than the diameter of the proximal end, so that the airbag pressure at the distal end is 1.05–1.15 times that at the proximal end.

[0029] In some implementations, the array-type skin pressure sensor is elongated, with its length aligned with the axial direction of the airbag cuff.

[0030] In some implementations, the array-type skin pressure sensor is divided into a central sensitive area and two pressure compensation areas on both sides along its width. The central sensitive area is used to acquire the main pulse wave, and the pressure compensation areas are used to acquire the lateral stress gradient caused by the airbag pressure, thereby decoupling and compensating for the main pulse wave to obtain the pulse pressure data.

[0031] Embodiments of the present invention also provide a method for estimating blood pressure values ​​using a high-precision blood pressure measuring device provided in any of the above embodiments, comprising the following steps: Acquire air pressure data and pulse pressure data in the airbag; Calculate a first data quality estimate of the air pressure data and a second data quality estimate of the pulse pressure data; Confidence weights are generated based on the first data quality estimate and the second data quality estimate; Based on the confidence weights, the final blood pressure estimate is calculated by combining the blood pressure estimates obtained from the air pressure data and pulse pressure data respectively.

[0032] In some implementation schemes, the generation of the pulse pressure data specifically includes the following process: Acquire unit signals from multiple sensor units; The unit signal is subjected to bandpass filtering and baseline drift removal processing; The unit quality index is obtained by calculating the ratio of signal amplitude to noise power for each unit signal. Using the reference channel with the highest signal-to-noise ratio as a benchmark, cross-correlation is used to estimate the time delay at each point and align the waveforms. The relative magnitude of the unit quality index is used as the weight corresponding to each unit signal, and unit signals with unit quality indices below a preset quality threshold are removed to synthesize the pulse pressure data. The unit quality indices are then fused according to their weights to obtain a second data quality estimate of the pulse pressure data.

[0033] In some implementations, the final blood pressure estimate is calculated as follows: ; ; P final = W bag ·P bag_est + W skin ·P skin_est ; Wherein, represents the confidence weight of the air pressure data; represents the confidence weight of the pulse pressure data; is the first data quality estimate of the air pressure data; is the second data quality estimate of the pulse pressure data; represents the final blood pressure estimate; represents the blood pressure estimate obtained based on the air pressure data analysis; represents the blood pressure estimate obtained based on the pulse pressure data analysis; α, β, and γ are adjustment coefficients.

[0034] The technical solution of the present invention will be further described in detail below through several embodiments and in conjunction with the accompanying drawings. However, the selected embodiments are only for illustrating the present invention and do not limit the scope of the present invention.

[0035] Example 1 This embodiment illustrates the specific structure of the aforementioned high-precision blood pressure measuring device, as shown below.

[0036] Airbag (cuff) system

[0037] Cuff: The upper arm features an inflatable cuff with a built-in standard medical-grade airbag. The material is polyurethane or polyester composite, and the arm circumference fits 22–42 cm (multiple sizes are available).

[0038] Air source and control: Miniature air pump and electromagnetic venting valve, solenoid valve or controllable venting port, supporting constant pressure inflation, constant speed deflation and step deflation strategies.

[0039] Airbag pressure sensor: Medical-grade pressure sensor (piezoresistive or capacitive), range 0–300 mmHg, accuracy ±1 mmHg, resolution ≤0.1 mmHg, sampling frequency ≥200 Hz (to ensure waveform details).

[0040] Skin pulse sensor array (skin patch)

[0041] Sensing unit type (preferably one or a combination): flexible capacitive / piezoresistive / piezoelectric thin film unit, unit size 3–8 mm, thickness <1 mm, sensitivity ≥ 0.1 kPa-1 or equivalent micro-displacement detection sensitivity.

[0042] Array geometry: Common implementations use a 3×3 or 4×4 two-dimensional grid; a 1×N linear array can also be used for measurements along the wrist or brachial artery. Cell spacing is 5–15 mm (5–10 mm is optional for the wrist side).

[0043] Encapsulation and bonding: Flexible substrate (such as PI, PDMS) + medical pressure-sensitive adhesive; a micro-pressure-conducting membrane is provided on the back of the unit to improve skin coupling consistency. The surface is coated with a hydrophobic / breathable coating to reduce the impact of sweat.

[0044] Signal link: front-end low-noise amplifier (input noise density ≤ 10 nV / √Hz), anti-power frequency filtering, differential acquisition, ADC 16–24 bit, sampling rate ≥250 Hz per channel (array parallel or high-speed multiplexed sampling).

[0045] Control and Synchronization

[0046] Main control unit: An embedded microprocessor (such as the ARM Cortex-M series) or SoC, responsible for controlling sampling, real-time processing, and wireless transmission. Heavy computational tasks (deep learning / model inference) can be offloaded to the Edge TPU / NN accelerator or computed via mobile cloud computing.

[0047] Clock synchronization: The airbag channel and all array channels use the same system clock or use hardware timestamps to ensure timing error <1 ms, in order to meet PTT estimation requirements.

[0048] Power and Communications

[0049] Power supply: Rechargeable lithium battery, capacity designed according to power consumption (e.g., 300–1000 mAh); Low power consumption strategies: on-demand sampling, event-triggered mode, local edge preprocessing.

[0050] Communication: BLE / Wi-Fi / USB are used for data upload and calibration interaction.

[0051] Example 2 The structural optimization of the high-precision blood pressure measuring device provided in Example 1 of this embodiment is shown below.

[0052] The cuff system provided in this embodiment has been structurally optimized based on the traditional airbag pressurization structure. Through the design of a double-layer flexible support layer, a local deformable pressure relief groove, and a central sensor embedding area, the cuff can achieve a uniform and controllable pressure distribution during the pressurization process, and provide a stable working environment for the pulse sensor array surface that avoids the influence of shear force.

[0053] (1) Key Structure: Double-layered flexible support: The outer layer has high tensile strength for overall tightness, while the inner layer uses a flexible and elastic material to automatically conform to the curvature of the arm.

[0054] Locally deformable pressure relief grooves: distributed in a circumferential direction, they can actively release excess local stress when the cuff is pressurized, preventing stress concentration in the sensor array area.

[0055] Central embedded sensor window: Reserves an independent area for the sensor array, so that the array can maintain a stable fit and is not affected by the deformation of the airbag.

[0056] (2) Design advantages: 1. Avoids sensor shear deformation caused by traditional cuff pressure, significantly reducing pulse waveform drift and distortion.

[0057] 2. Maintain consistent fit under different arm sizes and skin elasticity conditions to improve measurement repeatability.

[0058] 3. It can achieve effective fixation at lower inflation pressure, reducing discomfort for the test subject and improving the ability to monitor for a long time.

[0059] 4. It implements a zoning strategy of "local stability + circumferential pressure" to simultaneously take into account blood pressure measurement and high-fidelity pulse wave acquisition.

[0060] 2. Structural optimization of the skin pulse sensor array

[0061] To achieve high sensitivity and high stability, this invention designs a flexible biomimetic multi-point pulse sensor array, which is constructed using a flexible substrate, an independent buffer island structure, and micro-bumps to enhance bonding technology.

[0062] (1) Key Structure: Flexible thin-film substrate: ensures the entire array can bend naturally with the curvature of the wrist.

[0063] "Island-bridge" architecture: Each array element is interconnected by a flexible connecting bridge, which maintains the integrity of the whole while allowing local independent deformation to adapt to the microscopic bumps and depressions of the skin.

[0064] Bionic micro-bumps: mimicking the microstructure of fingertips or octopus suckers to increase the actual contact area and reduce slippage and false signals.

[0065] Differential compensation channels are formed between array elements: sampling combinations of adjacent array elements are used to offset uniform interference from wrist movement or cuff pressure.

[0066] (2) Design advantages: 1. The multi-element structure can maintain effective measurement even in areas with local skin damage (such as burns or rashes) or poor skin quality, significantly improving versatility.

[0067] 2. The "island-bridge" structure makes the array insensitive to local stress, but highly sensitive to the weak normal pulsations caused by the pulse, which greatly improves the clarity of the pulse waveform.

[0068] 3. The micro-protrusion enhanced fit technology can significantly reduce lateral wrist slippage (especially when the cuff is compressed), improving waveform stability.

[0069] 4. The array configuration allows for spatial redundancy. When an array element is affected by sweat, dander, or skin damage, the system can automatically select the optimal array element combination for stable output.

[0070] 3. Specific combination method of cuff and pulse sensor array

[0071] This invention proposes a "non-isobaric embedded combination structure" that integrates a sensor array inside the cuff in a specific way. Through a three-layer isolation structure, pressure guide groove, and low-resistance isolation film, it achieves high synchronization between blood pressure measurement and pulse wave measurement without interference.

[0072] (1) Key Structure: Three-layer isolation structure: Top layer: cuff airbag; Middle layer: soft isolation film, blocking airbag deformation from directly transmitting to the array; Bottom layer: array bonding layer, in direct contact with the skin. Pressure guide groove: Laterally guides the cuff pressure around the sensor area, making the sensor area controlled but not the "highest pressure point".

[0073] Independent Fitting Zone Fixation Strap: An "independent fitting zone" is formed inside the cuff, where there is almost no relative displacement when the airbag is inflated.

[0074] (2) Design advantages: This design addresses a critical issue with traditional cuff and patch sensors: inflation causes patch displacement or waveform distortion. The new structure allows the pulse array to maintain a "fixed displacement" during pressurization, avoiding shear disturbances.

[0075] The cuff compression and pulse detection are truly synchronized, rather than interfering with each other: the external pressure required for blood pressure detection changes over time; an array is used to capture the pulse wave inside the skin; and an isolation structure ensures that the two are orthogonal to each other and do not contaminate each other's signals.

[0076] The low-pressure, fixable binding method reduces the pressure burden on the skin and can still be used in cases of skin injury, burns, rashes, or in special populations (the elderly, infants).

[0077] Achieving "pseudo-motion compensation": The array and cuff are treated as a whole, rather than two separate attachments, so that the overall offset caused by motion has the same effect on the two types of signals, making it easier for the algorithm to cancel them out.

[0078] No additional straps or patches are needed; the integrated structure enhances ease of use.

[0079] Through the collaboration of the above systems, the following was achieved: This enables true synchronous calibration between cuff-type blood pressure measurement and array-type pulse measurement, which can be used for higher precision detection of systolic / diastolic phases or new algorithms (such as array-based maximum pulse point).

[0080] To avoid the common problem of cuffs "suppressing or deforming" the skin pulse wave signal in blood pressure measurement and improve the quality of early pulse signal detection, this method can be extended to the calibration of novel cuffless algorithms.

[0081] It can be applied to patients with damaged skin conditions, such as burns, severe dermatitis, and sensitive skin, because the array can work by relying on multi-point redundancy and light pressure bonding.

[0082] Enhanced measurement robustness: The array can automatically find the optimal combination of array elements to achieve "adaptive fitting".

[0083] The integrated system design reduces assembly and usage errors (such as wristband misalignment and patch misalignment), ensuring that it is usable by everyone and reproducible every time.

[0084] II. The product examples obtained based on the above structural design are shown below.

[0085] 1. Cuff structure

[0086] The cuff has a double-layer structure inside: (1) The inner layer is a controllable flexibility support layer made of an elastomer with 20–40 Shore A to avoid local stress peaks during inflation; (2) The outer layer is a confinement layer, which is used to control the direction of airbag pressurization so that the pressure is mainly applied to the arterial area in the radial direction.

[0087] (3) A main airbag channel is set inside the cuff. Its geometry is an asymmetrical semi-arc structure, so that the pressure distribution presents a gradient of 1.05–1.15 along the distal end > proximal end, in order to enhance the occlusion effect on the radial artery and at the same time reduce the compression interference on the skin pulse sensor array area.

[0088] 2. Skin pulse sensor array structure

[0089] The skin pulse sensor array is a 4×6 matrix of capacitive or strain gauge sensing units, with each unit preferably 3–4 mm in size and spaced 1.5–2.0 mm apart. The array has an elongated shape to geometrically match the course of the radial artery.

[0090] The array employs a central high-sensitivity channel for acquiring the main pulse wave.

[0091] Compensation channels on both sides: used to monitor the lateral stress gradient caused by the compression of the cuff, and for subsequent decoupling compensation.

[0092] The sensor array is externally encapsulated with a compliant matching layer (softness 10–20 Shore A), which can maintain a stable skin contact area under high voltage and avoid waveform flattening and mechanical low-pass filtering effects.

[0093] 3. The combination of array and cuffs

[0094] The present invention preferably adopts a structure of "partial window opening + partial embedding + pressure isolation groove": A structural window area with a length of 40–50 mm and a width of 12–18 mm is opened on the inside of the cuff.

[0095] The sensor array is installed in the window area with 1 / 3 of its thickness embedded and 2 / 3 of its thickness exposed, allowing it to directly contact the skin.

[0096] Pressure isolation grooves (1–2 mm wide and 0.5–1.0 mm deep) are set around the window area to prevent the airbag pressurization from directly affecting the sensor array.

[0097] In the aforementioned products, the cuff pressure does not directly "flatten" the array, ensuring that the pulse wave P_skin(t) does not deform or distort. The cuff pressure P_cuff(t) and P_skin(t) can be acquired simultaneously and with high synchronization. The array is not sensitive to skin condition and can be used for individuals with burns, ulcers, or other local tissue defects. Scene self-alignment feature: the array's shape matches the radial artery's course, eliminating the need for manual user positioning.

[0098] Example 3 The key algorithms and processing flow of the high-precision blood pressure measurement device provided in Examples 1 and 2 are as follows.

[0099] (1) Data acquisition and synchronization

[0100] Start measurement process: The system inflates the airbag to the preset starting pressure (usually 180-200 mmHg or as configured by the user) and records the starting time of the airbag pressure.

[0101] Synchronous acquisition: The airbag pressure sensor and the skin array sample in parallel, with a recommended sampling rate of ≥250 Hz, and all channels are stamped with the same timestamp.

[0102] In-array preprocessing (low-level fusion)

[0103] Perform the following processing on each cell in the array: Bandpass filtering: 0.5-15 Hz (or 0.3-20 Hz, depending on unit response) to preserve the pulse band and remove low-frequency drift and high-frequency noise.

[0104] Baseline drift removal: Use filtering or polynomial fitting to remove slow drift.

[0105] SNR estimation: Calculate the ratio of signal amplitude to noise power per unit (e.g., using power within the heart rate bandwidth / power outside the band) to obtain the quality index Qi.

[0106] Cross-correlation / alignment: Using the reference channel with the highest signal-to-noise ratio in the array as a benchmark, cross-correlation is used to estimate the time delay at each point and align the waveforms.

[0107] Spatial fusion: A single high-quality pulse channel P_skin(t) is synthesized using a weighted average or a PPA (pairwise phase alignment) / PCA method. The weights are adjusted according to Q_i, and outlier cells with Q_i less than a threshold are removed. Beamforming alignment and SNR enhancement can be optionally used.

[0108] Airbag channel treatment

[0109] Calculate the amplitude curve A(p) (the amplitude corresponding to the cuff pressure p) during the deflation process.

[0110] Extract mean arterial pressure (MAP): the cuff pressure p_max at the point of maximum amplitude.

[0111] Based on traditional oscillation / Korchinski judgment, the systolic / diastolic blood pressure estimates P_bag_SBP and P_bag_DBP of the airbag channel are calculated (based on amplitude threshold or model regression). At the same time, the quality estimate Q_bag of the airbag channel is performed (e.g., amplitude curve smoothness, noise level, and mixed pulse loss rate).

[0112] Signal quality assessment and dynamic weight adjustment

[0113] The set of quality metrics includes Q_bag (airbag quality) and Q_skin (skin channel quality obtained from the fusion results within the array), as well as the array consistency metric C_array (reflecting spatial consistency).

[0114] Dynamic weight calculation: Weights W_bag and W_skin are given by the following formula: ; ; α, β, and γ are adjustment coefficients that can be preset or adjusted through online learning. When airbag fluctuations are large or Q_bag decreases, W_skin is automatically increased; conversely, when Q_bag decreases, W_skin is increased. Array consistency C_array can be used as an amplification factor or threshold trigger for Q_skin.

[0115] Abnormal policy: If the quality of a channel is lower than the safety threshold (e.g., Q < Q_min), the channel is temporarily removed and a retest or delayed result output is prompted.

[0116] Feature extraction and blood pressure estimation model

[0117] Features extracted from P_skin(t) include: pulse peak, rise time, rise rate, pulse width, second wave amplitude, and interpulse delay (PTT / PWV obtained from an array).

[0118] Features extracted from the airbag channel include: amplitude peak position (p_max), amplitude curve morphology, and pressure value corresponding to the amplitude threshold.

[0119] Fusion Regression Model: This model utilizes a hybrid approach (physical + data-driven) for blood pressure estimation. Example solution: Physical priors: Initial estimates are provided by utilizing the physical relationship (linear or nonlinear) between PTT and blood pressure.

[0120] Data-driven regression: Gradient boosting trees (XGBoost / LightGBM) or neural networks (lightweight CNN+LSTM) are used to jointly regress physical features and waveform features to obtain the final systolic / diastolic blood pressure estimate.

[0121] Uncertainty output: The regression model also outputs the uncertainty σ. If σ exceeds the threshold, it will prompt that a retest is needed or provide a confidence interval.

[0122] Fusion output: Final blood pressure value P final = W bag ·P bag_est + W skin ·P skin_est (Or use a more complex weighted regression fusion machine).

[0123] Calibration and Self-Learning

[0124] Initial calibration: Each user performs N (e.g., 3-5) synchronous measurements with the reference cuff / hospital equipment upon first use for individualized model fine-tuning / calibration coefficients.

[0125] Online self-calibration: The system can gradually update individualized parameters (incremental learning) using low uncertainty data during daily use, combined with regular reference calibration recommendations.

[0126] Based on the above embodiments, it is clear that the key points of the technical solution proposed by the present invention include: 1. Dual-channel sensor fusion structure This invention achieves collaborative operation of two types of sensors in the same measurement system by constructing a composite structure of an airbag pressure sensor and a flexible skin pulse sensor. Its design includes the sensor installation method and relative position layout, the synchronous signal acquisition mechanism between the two, and the structural configuration to enhance the mechanical coupling effect between the sensor and the brachial artery (or other target artery), thereby ensuring the stability and consistency of dual-channel signal acquisition.

[0127] 2. Flexible Skin Pulse Sensor Array Solution

[0128] This invention employs a flexible array structure composed of multiple micro pressure-sensitive units. Its design encompasses the arrangement, density setting, and flexible packaging method of the array units. Simultaneously, by performing spatial filtering, weighted processing, and redundancy fusion on the array output, it achieves enhancement of effective pulse components, suppression of noise interference, and higher precision capture of skin micro-deformation details, thereby significantly improving the stability and quality of the pulse signal.

[0129] 3. Multi-source signal synchronous acquisition and preprocessing mechanism

[0130] This invention constructs a synchronous triggering mechanism for airbag pressure vibration signals and pulse array signals, thereby achieving timestamp alignment and phase correction of multi-source signals. At the same time, it combines multi-level denoising and filtering techniques to improve the usability of the original signals. In addition, by introducing quality evaluation indicators including signal-to-noise ratio (SNR), waveform stability and signal integrity, it achieves real-time quantification of the reliability of various signals, providing basic support for subsequent adaptive fusion.

[0131] 4. Adaptive Fusion Algorithm System with Dynamic Weight Adjustment

[0132] This invention dynamically adjusts the fusion weights of the airbag pressure signal and the skin pulse signal based on signal quality, enabling the system to adopt corresponding adaptive weight strategies at different measurement stages such as inflation, stabilization, initial deflation, and pressure fluctuations. Based on this, a fusion feature system and a dual-signal collaborative modeling method are constructed, thereby achieving the optimal combination of multi-source information and robust blood pressure estimation.

[0133] 5. Blood Pressure Estimation Model (Solution Method for Systolic / Diastolic Blood Pressure)

[0134] This invention proposes a method for constructing an oscillation envelope based on fused signals, and on this basis, it realizes the extraction of key feature points, including the maximum amplitude point and the feature ratio point, and the corresponding solution strategy. By simultaneously utilizing multi-source information from airbag pressure signals and skin pulse signals, the robustness and feature recognition accuracy of the oscillation method in blood pressure calculation are effectively improved. Furthermore, by combining an outlier removal mechanism based on fused data and model optimization methods, the reliability and anti-interference ability of the overall blood pressure solution process are further enhanced.

[0135] 6. A comprehensive mechanism to improve noise immunity and measurement stability

[0136] This invention utilizes the redundancy mechanism of arrayed signals to effectively suppress the influence of environmental noise, body movement interference, and weak pulse signals on the measurement. At the same time, it improves signal stability through tight coupling between the sensor and the skin, and introduces robust strategies for low signal-to-noise ratio or signal distortion stages into the fusion algorithm, thereby ensuring the accuracy and reliability of blood pressure measurement in complex environments.

[0137] 7. Complete blood pressure measuring device and its system structure

[0138] The blood pressure measuring device of the present invention includes an airbag system, a dual-channel sensing module, an array flexible pulse sensor, a signal processing circuit and a processor module. Its hardware layout has been optimized to enable the modules to work together and realize the synchronous acquisition, processing and fusion of multi-source signals, thereby ensuring the high accuracy and stability of the system in the blood pressure measurement process.

[0139] 8. Blood pressure measurement method based on the above fusion structure and algorithm

[0140] This invention provides a complete measurement process from signal acquisition, preprocessing, fusion, feature extraction to blood pressure calculation, and introduces a dynamic adaptive adjustment step in this process to optimize the weight allocation of multi-source signals in real time. This method is applicable to both continuous patch monitoring and traditional cuff measurement modes, enabling reliable acquisition of high-precision blood pressure data.

[0141] Therefore, the embodiments of the present invention significantly reduce the dependence on single-point alignment by using arrayed skin-touch sensors, thereby improving measurement robustness; the proposed dynamic weight adjustment mechanism automatically increases the skin channel contribution when the airbag signal is unstable or there is strong noise, thereby improving the overall measurement accuracy and stability; the PTT / PWV and high-resolution pulse characteristics provided by the array signal can be used as regression features to improve the accuracy of blood pressure estimation; in addition, the system provided by the embodiments of the present invention supports partial continuous monitoring and online self-calibration, improving its applicability in clinical and remote health management.

[0142] It should be understood that the above embodiments are merely illustrative of the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made in accordance with the spirit and essence of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A high-precision blood pressure measuring device with airbag pressure and pulse signal dual-channel fusion, characterized in that, The application relates to a blood pressure estimation device, comprising: an airbag cuff, an airbag pressure sensor, an arrayed skin pressure sensor and a micro-processing unit; the airbag pressure sensor is in communication with an airbag in the airbag cuff and is used for acquiring air pressure data in the airbag, and the arrayed skin pressure sensor is fixed on a skin-adhering surface of the airbag cuff and is used for acquiring pulse pressure data; the micro-processing unit is used for receiving the air pressure data and the pulse pressure data and fusing and processing to obtain a final blood pressure estimation value.

2. The high precision blood pressure measurement device of claim 1, wherein, The arrayed skin pressure sensor comprises a plurality of arrayed sensor units, each of which outputs a unit pressure signal, and the pulse pressure data is obtained by preprocessing and primary fusion of the unit pressure signals.

3. The high precision blood pressure measurement device of claim 1, wherein, The airbag cuff comprises, from outside to inside, a limiting layer, an airbag interlayer and a flexible support layer. An inner wall of the flexible support layer is provided with a fitting area, and the arrayed skin pressure sensor is arranged in the fitting area; surrounding the fitting area, the flexible support layer is further provided with an annular pressure isolation groove, and the depth of the pressure isolation groove is more than half the thickness of the flexible support layer.

4. The high precision blood pressure measurement device of claim 3, wherein, The fitting area is an inner recess structure, and the recess depth is 1 / 5-1 / 3 of the arrayed skin pressure sensor; part of the thickness of the arrayed skin pressure sensor is embedded in the inner recess structure, and the remaining part of the thickness protrudes from the inner recess structure.

5. The high precision blood pressure measurement device of claim 3, wherein, The diameter of a distal end of the airbag cuff is smaller than the diameter of a proximal end, so that the airbag pressure of the distal end is 1.05-1.15 times the airbag pressure of the proximal end.

6. The high precision blood pressure measurement device of claim 5, wherein, The arrayed skin pressure sensor has an outer shape of a long strip, and the length direction is consistent with the axial direction of the airbag cuff.

7. The high precision blood pressure measurement device of claim 6, wherein, In the width direction of the arrayed skin pressure sensor, the arrayed skin pressure sensor is divided into a central sensitive area and two side pressure compensation areas; the central sensitive area is used for acquiring a main pulse wave, and the pressure compensation areas are used for acquiring a transverse stress gradient caused by the airbag pressure, for decoupling compensation of the main pulse wave to obtain the pulse pressure data.

8. A blood pressure value estimation method for the high-precision blood pressure measuring apparatus according to any one of claims 1 to 7, characterized by, The application relates to a blood pressure estimation device, comprising: acquiring air pressure data in an airbag and pulse pressure data; calculating a first data quality estimation value of the air pressure data and a second data quality estimation value of the pulse pressure data; generating a confidence weight based on the first data quality estimation value and the second data quality estimation value; calculating a final blood pressure estimation value based on the confidence weight and blood pressure estimation values obtained by analyzing the air pressure data and the pulse pressure data respectively.

9. The blood pressure value estimation method according to claim 8, characterized by, The generation mode of the pulse pressure data comprises the following steps: acquiring unit signals of a plurality of sensor units; performing band-pass filtering and baseline drift processing on the unit signals; calculating a signal amplitude-to-noise power ratio of each unit signal to obtain a unit quality index; taking a reference channel with the highest signal-to-noise ratio as a reference, using cross-correlation to estimate time delays of each point and aligning waveforms; taking the relative size of the unit quality index as a weight corresponding to each unit signal, removing unit signals with a unit quality index below a preset quality threshold, synthesizing the pulse pressure data, and fusing the unit quality index as a second data quality estimation value of the pulse pressure data according to the weight.

10. The blood pressure value estimation method according to claim 8, characterized by, The final blood pressure estimation value is calculated in the following manner: ； ； P final = W bag ·P bag_est + W skin ·P skin_est ; Wherein, represents the confidence weight of the air pressure data; represents the confidence weight of the pulse pressure data; is the first data quality estimation value of the air pressure data; is the second data quality estimation value of the pulse pressure data; represents the final blood pressure estimation value; represents the blood pressure estimation value obtained based on the air pressure data; represents the blood pressure estimation value obtained based on the pulse pressure data; and α, β, γ are adjustment coefficients.

Images