Blood pressure measurement device and method thereof
The integration of a piezopump and neural network-based audio analysis in a blood pressure measurement device addresses noise and variability issues, ensuring accurate and comfortable blood pressure determination through dual-phase measurement.
Patent Information
- Application Number
- PCT/US2026/026779
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-06
- Filing Date
- 2026-05-05
- Publication Date
- 2026-11-12
AI Technical Summary
Existing automated blood pressure measurement devices face challenges with noise interference, operator variability, and measurement comfort, particularly in accurately detecting Korotkoff sounds for systolic and diastolic blood pressure readings.
A blood pressure measurement device combining a piezopump-based pneumatic system with neural network-based audio analysis to detect Korotkoff sounds, using a microcomputer for real-time neural network inference, solenoid valves for precise pressure control, and a digital stethoscope for audio signal capture, enabling dual-phase measurement for accurate blood pressure determination.
The device provides accurate and consistent blood pressure readings by reducing acoustic interference and operator variability, ensuring user comfort through precise pressure control and refined measurement phases, with improved signal-to-noise ratio and reduced false positives.
Smart Images

Figure US2026026779_12112026_PF_FP_ABST
Abstract
Description
Atorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCTBLOOD PRESSURE MEASUREMENT DEVICE AND METHOD THEREOFCROSS-REFERENCE TO RELATED APPLICATIONS)
[0001] The present international application claims priority to United States Provisional Patent Application No. 63 / 800,987, filed on May 6, 2025. The entire contents of the above noted provisional application are incorporated by reference as part of the disclosure of this document.TECHNICAL FIELD
[0002] The present disclosure relates to blood pressure measurement devices and methods, and more particularly to an automated blood pressure measurement device that uses neural network-based audio analysis to detect Korotkoff sounds for determining systolic and diastolic blood pressure values.BACKGROUND
[0003] Blood pressure measurement is a routine diagnostic procedure performed in clinical settings and increasingly in home environments. Accurate blood pressure readings provide valuable information for assessing cardiovascular health and managing conditions such as hypertension. Non-invasive blood pressure measurement techniques have been developed to allow for convenient monitoring without the risks associated with invasive arterial catheterization. For example, the auscultatory method, which involves listening for Korotkoff sounds using a stethoscope while manually controlling cuff pressure, has long been recognized as a reliable approach for non-invasive blood pressure measurement. Korotkoff sounds are the audible sounds produced by blood flow through an artery as external cuff pressure transitions between complete occlusion and unrestricted flow. The appearance and disappearance of these sounds correspond to systolic and diastolic blood pressure values, respectively. However, traditional auscultatory methods rely on trained human operators to control cuff pressure and manually detect Korotkoff sounds, which can introduce variability in measurements.
[0004] Automated blood pressure measurement devices typically employ pumps to inflate and deflate the arm cuff. Conventional pump designs, such as peristaltic pumps, can generate audible noise during operation. This noise may interfere with audio-based measurement approaches that rely on detecting subtle sounds during the measurement process. The integration of digital signal processing and machine learning techniques into medical devicesAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT has opened new possibilities for analyzing physiological signals. For example, neural networkbased approaches can be trained to recognize patterns in audio and other sensor data, potentially enabling automated classification of sounds that were previously identified through manual observation.
[0005] Accordingly, there remains an opportunity for improvements in automated blood pressure measurement devices that can provide accurate and consistent readings while addressing challenges related to noise, operator variability, and measurement comfort.SUMMARY
[0006] The disclosed embodiments describe automated blood pressure measurement devices and methods that, among other features and benefits, combine piezopump-based pneumatic systems with neural network-based audio analysis to accurately detect Korotkoff sounds in determining blood pressure values.
[0007] One example blood pressure measurement device includes a pump configured to inflate and deflate a cuff, and a pump control system coupled to the pump and configured to manage operations of the pump, the pump control system including a microcomputer configured for neutral network inference and control of the pump. The example blood pressure measurement device also includes a plurality of solenoid valves coupled to the pump control system and controlled by the microcomputer to regulate cuff pressure, and an audio sensor configured to detect Korotkoff sounds. Furthermore, the example blood pressure measurement device includes a neural network implemented on the microcomputer, wherein the neural network is configured to process audio data from the audio sensor to classify detected sounds as Korotkoff sounds or non- Korotkoff sounds during both inflation and deflation of the cuff. BRIEF DESCRIPTION OF THE DRAWINGS
[0008] FIG. 1 illustrates a block diagram of a blood pressure measurement device according to one or more embodiments of the present technology.
[0009] FIG. 2 illustrates a block diagram of an automated blood pressure measurement system according to one or more embodiments of the present technology.
[0010] FIG. 3A illustrates an audio signal waveform with Korotkoff sound classifications according to one or more embodiments of the present technology.
[0011] FIG. 3B illustrates an audio signal waveform with non-Korotkoff sound classifications according to one or more embodiments of the present technology.Atorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT
[0012] FIG. 3C illustrates pump pressure data with detection markers during a blood pressure measurement according to one or more embodiments of the present technology.
[0013] FIG. 4 illustrates a user interface screen example according to one or more embodiments of the present technology.
[0014] FIG. 5A illustrates a pneumatic flow configuration during an inflation mode of the blood pressure measurement device according to one or more embodiments of the present technology.
[0015] FIG. 5B illustrates a pneumatic flow configuration during a slow deflation mode of the blood pressure measurement device according to one or more embodiments of the present technology.
[0016] FIG. 5C illustrates a pneumatic flow configuration during a rapid deflation mode of the blood pressure measurement device according to one or more embodiments of the present technology.
[0017] FIG. 6 illustrates a multi-curve diagram depicting a relationship between cuff pressure, oscillogram peaks, and blood pressure estimates across four measurement phases according to one or more embodiments of the present technology.
[0018] FIG. 7 illustrates a comparative analysis diagram depicting SBP evaluation results according to one or more embodiments of the present technology.
[0019] FIG. 8 illustrates a comparative analysis diagram depicting DBP evaluation results according to one or more embodiments of the present technology
[0020] FIG. 9 illustrates a flowchart for a method of measuring blood pressure according to one or more embodiments of the present technology.
[0021] FIG. 10 illustrates a flowchart for another method of measuring blood pressure according to one or more embodiments of the present technology.
[0022] FIG. 11 illustrates a flowchart for another method of measuring blood pressure according to one or more embodiments of the present technology.DETAILED DESCRIPTION
[0023] The following description sets forth example aspects of the present disclosure. It should be recognized, however, that such description is not intended as a limitation on theAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT scope of the present disclosure. Rather, the description also encompasses combinations and modifications to those exemplary aspects described herein.
[0024] The present disclosure provides an automated blood pressure measurement device and method that combines a piezopump-based pneumatic system with neural network-based audio analysis to accurately detect Korotkoff sounds and determine blood pressure values. The device includes a housing containing a microcomputer configured for real-time neural network inference, a pump-control board, a piezopump, solenoid valves for controlling inflation and deflation phases, and a pressure sensor. An arm cuff assembly connected to the housing includes an inflatable bladder and a digital stethoscope that captures audio signals during measurement. The microcomputer processes audio data from the digital stethoscope using a trained neural network to classify detected sounds as Korotkoff sounds or non-Korotkoff sounds, and correlates these classifications with corresponding cuff pressure values to determine systolic and diastolic blood pressure.
[0025] Key features of the present technology include a dual-phase measurement approach that performs neural network-based Korotkoff sound detection during both inflation and deflation phases, enabling rough blood pressure estimation during inflation followed by refined measurements during controlled deflation. The system employs oscillogram analysis to detect pressure peaks corresponding to heartbeats, and for each detected peak, extracts a corresponding audio window for neural network classification. The piezopump adopted in the present technology provides quiet operation compared to conventional pump designs, reducing acoustic interference with the audio-based detection approach. The blood pressure measurement device further includes multiple solenoid valves that enable precise control over cuff pressure transitions, including rapid deflation for user comfort and slow controlled deflation for accurate measurement. A terminating process is included in the operations of blood pressure measurement device of the present technology to ensure safe shutdown of hardware components and software threads in response to various interrupt conditions, including user abort inputs, successful blood pressure determination, or failure to detect significant oscillogram peaks.
[0026] FIG. 1 illustrates a block diagram of a blood pressure measurement device according to one or more embodiments of the present technology. As shown in FIG. 1, the blood pressure measurement device comprises a blood pressure measurement device housing 100 and an arm cuff assembly 200. The blood pressure measurement device housing 100 contains components configured to perform automated blood pressure measurements, and theAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT arm cuff assembly 200 is pneumatically and electrically connected to the blood pressure measurement device housing 100. In some embodiments, the blood pressure measurement device of the present technology performs automated blood pressure measurement using neural network-based audio analysis to detect Korotkoff sounds. Specifically, the blood pressure measurement device housing 100 includes processing components configured to receive audio data from the arm cuff assembly 200 and to classify detected sounds as Korotkoff sounds or non-Korotkoff sounds using a neural network. Based on the classification of detected sounds and corresponding cuff pressure data, the blood pressure measurement device of the present technology determines systolic blood pressure (SBP) and diastolic blood pressure (DBP) values.
[0027] As shown in FIG. 1, the blood pressure measurement device housing 100 includes a touchscreen interface 120 operatively connected to a blood pressure measurement system / microcomputer 130. The touchscreen interface 120 facilitates user interaction and system control, allowing a user to interact with the blood pressure measurement device through a graphical interface. In some embodiments, the touchscreen interface 120 receives user inputs for initiating a blood pressure measurement process, setting pressure parameters, and selecting a measurement mode. For example, the touchscreen interface 120 can be configured to allow a user to select between different measurement modes, such as a “Sweep” mode or a “Ramp” mode, depending on the desired measurement approach. The touchscreen interface 120 can also be configured to display blood pressure measurement results to the user, including SBP and DBP values determined by the blood pressure measurement device of the present technology. In some other embodiments, the touchscreen interface 120 displays a visualization summarizing trial data from a blood pressure measurement session. The touchscreen interface 120 can be further configured to include one or more user interface controls, such as buttons, that may be enabled or disabled during different phases of the measurement process. For example, an abort button can be implemented on the touchscreen interface 120 to allow a user to terminate a measurement session before completion.
[0028] The blood pressure measurement system / microcomputer 130 serves as a central processing unit for the blood pressure measurement device and is configured to handle realtime neural network inference and overall system control. In some embodiments, the blood pressure measurement system / microcomputer 130 may comprise a Jetson Orin NX microcomputer or similar edge artificial intelligent computing device capable of performing neural network inference operations. The blood pressure measurement system / microcomputerAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT 130 may also include a processor, a memory unit, and an interface that are communicatively connected to each other. In the present technology, the blood pressure measurement system / microcomputer 130 is configured to perform various operations based on execution of program code stored on a storage medium. In some embodiments, the blood pressure measurement system / microcomputer 130 may be in communication with at least one memory and with at least one communication unit that enables the exchange of data and information with other entities, devices, and networks.
[0029] The blood pressure measurement system / microcomputer 130 is configured to control the operation of pneumatic components within the blood pressure measurement device housing 100, including a pump and solenoid valves, to inflate and deflate the arm cuff assembly 200. The blood pressure measurement system I microcomputer 130 is further configured to receive audio data from the arm cuff assembly 200 and to analyze the audio data using a neural network to detect Korotkoff sounds. Based on the detected Korotkoff sounds and corresponding cuff pressure data, the blood pressure measurement system / microcomputer 130 determines blood pressure values. In some embodiments, the blood pressure measurement system / microcomputer 130 manages multiple background threads that operate concurrently during a blood pressure measurement session, including threads for audio recording, pressure monitoring, oscillogram processing, neural network inference, and pump control. Upon completion of a measurement session, the blood pressure measurement system / microcomputer 130 transmits final blood pressure results to the touchscreen interface 120 for display to the user.
[0030] As shown in FIG. 1, the blood pressure measurement device housing 100 includes a pump control system, such as a pump-control board 140, coupled to the blood pressure measurement system I microcomputer 130. The pump-control board 140 is configured to manage the operation of pneumatic components within the blood pressure measurement device housing 100. In some embodiments, the pump-control board 140 receives control signals from the blood pressure measurement system I microcomputer 130 and translates the control signals into drive signals for pneumatic components. In particular, the pump-control board 140 interfaces with a pump and solenoid valves to regulate cuff pressure during blood pressure measurement. In some other embodiments, the pump-control board 140 is a pump control system and includes a microcomputer configured for neutral network processing (such as inference) and control of the piezopump.Atorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT
[0031] The blood pressure measurement device housing 100 further includes a pump, such as a piezopump 150, connected to the pump-control board 140, as shown in FIG. 1. The piezopump 150 is configured to inflate and deflate the arm cuff assembly 200 through pneumatic connections (shown in dotted line). In general, the piezopump 150 generates reduced audio noise compared to conventional pump designs, such as peristaltic pumps. The reduced audio noise generated by the piezopump 150 improves the signal-to-noise ratio for audio signals captured during blood pressure measurement, allowing for detection of Korotkoff sounds during both inflation and deflation phases. The pump-control board 140 drives the piezopump 150 according to control signals received from the blood pressure measurement system / microcomputer 130, enabling the blood pressure measurement device to inflate the arm cuff assembly 200 during an inflation phase and to deflate the arm cuff assembly 200 during a deflation phase. In some other embodiments, the pump included in the blood pressure measurement device housing 100 can be an air pump, a micro pump, a bulb pump, or any other types of pumps that can be adopted for blood pressure measurement application.
[0032] The blood pressure measurement device housing 100 also includes a first solenoid valve 160 and a second solenoid valve 170, both connected to the pump-control board 140. The first solenoid valve 160 is configured to control the transition between inflation and deflation phases of the arm cuff assembly 200. In some embodiments, the first solenoid valve 160 is actuated by the pump-control board 140 to switch the pneumatic system from an inflation state to a deflation state. The first solenoid valve 160 may be positioned in a pneumatic pathway between the piezopump 150 and the arm cuff assembly 200 to regulate airflow direction. The second solenoid valve 170 is configured to regulate the rate of deflation of the arm cuff assembly 200. In some embodiments, the second solenoid valve 170 allows for both rapid pressure release and slow controlled pressure release as required by the measurement protocol. For example, the second solenoid valve 170 may be controlled to provide rapid deflation for user comfort at the end of a measurement session, releasing cuff pressure to atmospheric pressure. In some other embodiments, the second solenoid valve 170 may be controlled to provide slow deflation during a measurement phase, allowing the blood pressure measurement system / microcomputer 130 to detect Korotkoff sounds and determine blood pressure values with increased accuracy. The slow deflation rate may be configured to release pressure at a rate suitable for detecting individual heartbeats and corresponding Korotkoff sounds, such as a rate in a range of approximately 2 to 5 mmHg per second. The rapid deflation rate may beAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT configured to release pressure at a higher rate, such as a rate in a range of approximately 10 to 50 mmHg per second, to quickly equalize cuff pressure to atmospheric pressure.
[0033] In the present technology, the pump-control board 140 controls the first solenoid valve 160 and the second solenoid valve 170 in coordination with the piezopump 150 to implement different measurement modes. For example, in the “Sweep” mode, the pumpcontrol board 140 can be configured to control the piezopump 150 to inflate the arm cuff assembly 200 while the blood pressure measurement system / microcomputer 130 performs preliminary audio analysis to estimate rough blood pressure values. In the “Ramp” mode, the pump-control board 140 may control the second solenoid valve 170 to provide slow deflation within a pressure region identified during inflation, allowing for refined blood pressure measurements. The coordination of the piezopump 150, the first solenoid valve 160, and the second solenoid valve 170 by the pump-control board 140 enables the blood pressure measurement device to perform a dual-phase measurement approach that includes estimated rough blood pressure estimation during inflation followed by refined measurements during controlled deflation. In some other embodiments, the blood pressure measurement system / microcomputer 130 is configured to analyze a sequence of annotated peaks to determine the rough DBF and SBP values during the inflation and to determine the refined DBP and SBP values during deflation, based on a pattern of Korotkoff sound classifications.
[0034] In the present technology, the blood pressure measurement device housing 100 further includes a pressure sensor 180 pneumatically connected to the arm cuff assembly 200. The pressure sensor 180 is configured to monitor cuff pressure within the inflatable arm cuff bladder 210 during blood pressure measurement. In some embodiments, the pressure sensor 180 is positioned within the blood pressure measurement device housing 100 and is connected to the arm cuff assembly 200 through a pneumatic pathway that allows the pressure sensor 180 to detect pressure changes within the inflatable arm cuff bladder 210. The pressure sensor 180 provides pressure data to the blood pressure measurement system / microcomputer 130 for processing. In some embodiments, the pressure sensor 180 generates timestamped pressure data that is transmitted to the blood pressure measurement system I microcomputer 130 during both inflation and deflation phases of the arm cuff assembly 200. The blood pressure measurement system / microcomputer 130 may continuously monitor the timestamped pressure data from the pressure sensor 180 throughout a blood pressure measurement session. The timestamped pressure data allows the blood pressure measurement system / microcomputer 130 to correlate detected audio signals with corresponding cuff pressure values,Atorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT enabling determination of systolic and diastolic blood pressure values based on the timing of Korotkoff sound classifications.
[0035] As shown in FIG. 1, the arm cuff assembly 200 includes the inflatable arm cuff / bladder 210 and a digital stethoscope I audio sensor 220. The inflatable arm cuff I bladder 210 receives air from the piezopump 150 through the first solenoid valve 160 and the second solenoid valve 170 via pneumatic connections (dotted arrows). During a blood pressure measurement session, the inflatable arm cuff bladder 210 is inflated by the piezopump 150 to apply pressure to a user's arm, and the inflatable arm cuff bladder 210 is deflated through the solenoid valves to release pressure according to the measurement protocol. The digital stethoscope I audio sensor 220 is integrated within the arm cuff assembly 200 and is configured to detect Korotkoff sounds during blood pressure measurement. In some embodiments, the digital stethoscope / audio sensor 220 comprises a microphone configured to capture audio signals produced by blood flow through an artery as external cuff pressure transitions between complete occlusion and unrestricted flow. The digital stethoscope I audio sensor 220 is configured to transmit audio data to the blood pressure measurement system / microcomputer 130 via an electrical data connection for real-time analysis using neural network inference. In some embodiments, the blood pressure measurement system / microcomputer 130 processes the audio data from the digital stethoscope / audio sensor 220 to classify detected sounds as Korotkoff sounds or non-Korotkoff sounds. For example, the neural network may analyze audio windows corresponding to detected pressure peaks in cuff oscillations, and the neural network may classify each audio window based on audio signal characteristics such as amplitude variations and waveform paterns. The first appearance of Korotkoff sounds during deflation corresponds to systolic blood pressure, and the disappearance of Korotkoff sounds corresponds to diastolic blood pressure. In some embodiments, the blood pressure measurement system / microcomputer 130 may use pressure fluctuations within the inflatable arm cuff / bladder 210 to validate the timing of detected Korotkoff sounds, reducing the likelihood of false positive classifications.
[0036] In some embodiments, the neural network implemented on the blood pressure measurement system / microcomputer 130 may comprise various neural network architectures suitable for audio classification tasks. For example, the neural network may include convolutional neural networks (CNNs) configured to extract features from audio spectrograms or waveform representations, recurrent neural networks (RNNs) or long short-term memory (LSTM) networks configured to capture temporal dependencies in audio sequences, orAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT transformer-based architectures configured to process audio data using attention mechanisms. In some embodiments, the neural network may comprise a lightweight model optimized for edge deployment, such as a MobileNet or EfficientNet variant, to enable real-time inference on the microcomputer with limited computational resources. The neural network may be trained on labeled audio data comprising Korotkoff sound samples and non-Korotkoff sound samples to leam discriminative features for classification.
[0037] FIG. 2 illustrates a block diagram of an automated blood pressure measurement system according to another embodiment of the present technology. As shown in FIG. 2, the automated blood pressure measurement system includes an Edge Al Device (such as a Jetson Orin NX microcomputer) that serves as a central processing unit for the system. The Edge Al Device is configured to perform real-time neural network inference and to estimate SBP and DBP based on cuff pressure data and Korotkoff sound data received from an arm cuff with an integrated digital stethoscope. The automated blood pressure measurement system further includes a Touch Screen for User Interface connected to the Edge Al Device, allowing users to interact with the system and view measurement results. The Edge Al Device controls cuff pressure by sending control signals to a piezopump development board, which is configured to drive a pump and drive valves that regulate cuff inflation and deflation. The piezopump development board is connected to an Arm Cuff with Digital Stethoscope, and the piezopump development board is configured to inflate and deflate the arm cuff as directed by the Edge Al Device. During operation, the Arm Cuff with Digital Stethoscope captures Korotkoff sound audio signals during the measurement process and transmits the audio data to the Edge Al Device for analysis. The neural network implemented on the Edge Al Device processes the audio data to classify detected sounds as Korotkoff sounds or non-Korotkoff sounds, and based on the classification and corresponding cuff pressure data, the automated blood pressure measurement system determines blood pressure values. The Touch Screen for User Interface displays the measurement results to the user and provides controls for initiating and managing the measurement process.
[0038] In the automated blood pressure measurement system of FIG. 2, the Edge Al Device receives Korotkoff sound audio from the digital stethoscope integrated within the Arm Cuff with Digital Stethoscope. During a blood pressure measurement session, the digital stethoscope continuously captures audio signals produced by blood flow through an artery as cuff pressure changes. The captured audio data is transmitted from the Arm Cuff with Digital Stethoscope to the Edge Al Device via an electrical data connection for real-time analysis. InAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT the present technology, the Edge Al Device implements a neural network configured to process the received audio data and classify detected sounds as Korotkoff sounds or non-Korotkoff sounds. In some embodiments, the neural network processing on the Edge Al Device operates in conjunction with oscillogram analysis. The Edge Al Device may process timestamped pressure data to extract cuff oscillations and detect peaks in the cuff oscillations. For each detected peak in the oscillogram, the Edge Al Device may extract a corresponding audio window from the recorded audio data. The extracted audio window is then processed using the neural network to classify the audio window as a Korotkoff sound or a non-Korotkoff sound. The Edge Al Device may annotate each corresponding oscillogram peak with the classification result from the neural network.
[0039] In some embodiments, the neural network implemented on the Edge Al Device classifies audio signals to detect the first appearance and subsequent disappearance of Korotkoff sounds. The first appearance of Korotkoff sounds during deflation corresponds to SBP, and the disappearance of Korotkoff sounds corresponds to DBP. The Edge Al Device may analyze a sequence of annotated oscillogram peaks to determine blood pressure values based on a pattern of Korotkoff sound classifications. In some embodiments, the Edge Al Device determines estimated rough DBP and SBP values during an inflation phase and determines refined DBP and SBP values during a deflation phase, based on the pattern of Korotkoff sound classifications and corresponding cuff pressure data.
[0040] In some embodiments, the Edge Al Device may use pressure fluctuations within the arm cuff to validate the timing of detected Korotkoff sounds. The validation using pressure fluctuations may reduce the likelihood of false positive classifications by correlating detected audio signals with expected pressure variations corresponding to heartbeats. Upon completion of a blood pressure measurement session, the Edge Al Device transmits the determined blood pressure values to the Touch Screen for User Interface. The Touch Screen for User Interface displays the measurement results, including SBP and DBP values, to the user. In some embodiments, the Touch Screen for User Interface may display a visualization summarizing trial data from the blood pressure measurement session. The Touch Screen for User Interface may also provide controls for initiating subsequent measurement sessions and managing measurement settings.
[0041] FIGS. 3A-3C illustrate time-synchronized graphs depicting audio signal analysis and pump pressure data during a blood pressure measurement process according to one or more embodiments of the present technology. In particular, FIG. 3A displays an audio signalAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT waveform captured by a digital stethoscope during blood pressure measurement, with regions highlighted in shaded segments where a neural network has classified the audio as Korotkoff sounds. The audio signal waveform in FIG. 3A shows amplitude variations over time as the digital stethoscope captures sounds produced by blood flow through an artery during cuff pressure changes. Here, the green highlighted regions indicate classification results from neural network inference, where the neural network has determined that the audio within the highlighted segments corresponds to Korotkoff sounds. The neural network processes audio windows extracted from the recorded audio data and classifies each audio window based on audio signal characteristics. In some embodiments, the neural network may analyze features such as amplitude variations, waveform patterns, frequency content, and temporal characteristics to distinguish Korotkoff sounds from other sounds. As shown, the shaded highlighted regions in FIG. 3A appear at intervals corresponding to individual heartbeats during the blood pressure measurement process. Each shaded highlighted region represents an audio window that the neural network has classified as containing a Korotkoff sound. In some embodiments, the neural network classification operates in conjunction with oscillogram analysis, where detected peaks in cuff pressure oscillations trigger extraction of corresponding audio windows for neural network inference. The classification results shown in FIG. 3A can be used by the blood pressure measurement system to determine systolic and diastolic blood pressure values based on the first appearance and disappearance of Korotkoff sounds during deflation of the cuff.
[0042] FIG. 3B illustrates the same audio signal waveform as FIG. 3A, with regions highlighted in red representing segments where the neural network has classified the audio as non-Korotkoff sounds (also referred to as Class 1 or other regions). The shaded highlighted regions in FIG. 3B indicate audio windows that the neural network has determined do not correspond to arterial blood flow sounds associated with Korotkoff sounds. In some embodiments, the non-Korotkoff sound classification may include ambient noise captured by the digital stethoscope, device-generated noise from pneumatic components, user movement artifacts, or other acoustic signals that do not exhibit the characteristic waveform patterns and frequency content of Korotkoff sounds. The neural network adopted in present technology can distinguish between Korotkoff sounds and non-Korotkoff sounds based on learned discriminative features, such as differences in amplitude envelope, spectral composition, and temporal structure. By classifying audio segments as non-Korotkoff sounds, the blood pressureAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT measurement system may filter out acoustic interference and reduce false positive detections that could otherwise affect the accuracy of blood pressure determination.
[0043] FIG. 3C illustrates pump pressure data with detection markers and oscillation segment highlighting during a blood pressure measurement process. The graph in FIG. 3C includes a legend indicating pressure values in mmHg, a start pressure threshold of 60 mmHg shown as a dashed horizontal line, and an oscillation segment shown as a shaded region. The pressure curve in FIG. 3C depicts an initial inflation phase where cuff pressure increases as the piezopump inflates the arm cuff, followed by a transition to a measurement phase where the pressure stabilizes and oscillations are detected. The vertical dashed line in FIG. 3C marks the transition point between the inflation phase and the measurement phase. The shaded region in FIG. 3C corresponds to the pressure range where oscillogram signals are analyzed to detect heartbeats, and the detected heartbeats subsequently trigger neural network inference for Korotkoff sound classification as shown in the upper two graphs of FIGS. 3A and 3B. For each detected pressure peak within the oscillation segment, the blood pressure measurement system may extract a corresponding audio window from the recorded audio data and process the audio window using the neural network to classify the audio window as a Korotkoff sound or a non-Korotkoff sound. The time synchronization between the pump pressure data in FIG. 3C and the audio signal analysis in FIGS. 3 A and 3B allows the blood pressure measurement system to correlate detected Korotkoff sounds with corresponding cuff pressure values for determination of systolic and diastolic blood pressure.
[0044] In some embodiments, the blood pressure measurement device of the present technology includes a graphical user interface displayed on a touchscreen interface for controlling and monitoring blood pressure measurement sessions. FIG. 4 illustrates an example user interface screen according to one or more embodiments of the present technology. As shown in FIG. 4, the user interface includes a main display area presenting real-time blood pressure measurement data and controls for managing the blood pressure measurement process. The user interface includes a pressure gauge or numerical display showing current cuff pressure in mmHg, allowing the user to monitor inflation and deflation progress during a measurement session. The user interface further includes measurement result fields for displaying determined SBP and DBP values upon completion of a measurement. Control buttons are provided on the user interface for initiating a new measurement session, aborting an ongoing measurement, and accessing measurement settings or historical data. In some embodiments, the user interface includes a mode selection control allowing the user to choose betweenAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT different measurement modes, such as a “Sweep” mode for rapid estimation or a “Ramp” mode for refined measurements. The user interface may also display status indicators showing the current phase of the measurement process, such as inflation, measurement, or deflation phases, and may provide visual feedback indicating successful detection of Korotkoff sounds during the measurement session.
[0045] FIGS. 5A-5C illustrate pneumatic flow diagrams depicting different operational modes of the blood pressure measurement device according to one or more embodiments of the present technology. The pneumatic flow diagrams show the configuration of airflow paths among the piezopump, the inflatable arm cuff bladder, solenoid valves, and the atmosphere during inflation mode, slow deflation mode, and rapid deflation mode. Each diagram illustrates how the solenoid valves are controlled to direct airflow for achieving the desired pressure behavior in the arm cuff during different phases of the blood pressure measurement process.
[0046] FIG. 5A illustrates the pneumatic flow configuration during an inflation mode of the blood pressure measurement device. In the inflation mode, the piezopump is activated to generate airflow that is directed toward the inflatable arm cuff bladder. The first solenoid valve is configured to establish a pneumatic pathway between the piezopump and the inflatable arm cuff bladder, allowing air to flow from the piezopump into the cuff. The second solenoid valve is configured to block the pathway to the atmosphere, preventing air from escaping the pneumatic system. As a result, the air generated by the piezopump accumulates within the inflatable arm cuff bladder, causing the cuff pressure to increase progressively until a target inflation pressure is reached.
[0047] FIG. 5B illustrates the pneumatic flow configuration during a slow deflation mode of the blood pressure measurement device. In the slow deflation mode, the piezopump is deactivated, and the first solenoid valve is configured to isolate the piezopump from the pneumatic pathway. The second solenoid valve is configured to establish a restricted pathway between the inflatable arm cuff bladder and the atmosphere, allowing air to escape from the cuff and through the orifice aperture at a controlled rate. The restricted airflow pathway enables slow pressure release at a rate suitable for detecting individual heartbeats and corresponding Korotkoff sounds during the measurement phase, such as a rate in a range of approximately 2 to 5 mmHg per second.
[0048] FIG. 5C illustrates the pneumatic flow configuration during a rapid deflation mode of the blood pressure measurement device. In the rapid deflation mode, the piezopump isAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT deactivated, and both the first solenoid valve and the second solenoid valve are configured to establish an unrestricted pathway between the inflatable ami cuff bladder and the atmosphere. The unrestricted airflow pathway allows air to escape rapidly from the cuff, equalizing the cuff pressure to atmospheric pressure for user comfort and safety. The rapid deflation mode is typically employed at the end of a measurement session or in response to an abort command from the user, releasing any remaining pressure in the inflatable arm cuff bladder.
[0049] FIG. 6 illustrates a multi-curve diagram depicting the relationship between cuff pressure, oscillogram peaks, and blood pressure estimates across four measurement phases according to one or more embodiments of the present technology. The diagram includes four vertically aligned curves that are time-synchronized to show the progression of a blood pressure measurement session. The top curve displays cuff pressure in mmHg over time, showing an initial inflation phase (Phase 1) where pressure increases, followed by a plateau region (Phase 2), and then a controlled deflation phase where pressure gradually decreases through Phases 3, and 4. The second curve illustrates oscillogram peaks detected from pressure fluctuations within the arm cuff, with vertical bars indicating the amplitude of detected oscillations corresponding to individual heartbeats. The third curve shows Korotkoff sound detection results, where detected Korotkoff sounds are indicated by markers at corresponding time points during the measurement process. The bottom curve presents blood pressure estimates, showing how rough SBP and DBP values are estimated during the inflation phase (Phases 1 and 2) and subsequently refined during the deflation phases (Phases 3-4) as additional Korotkoff sound data is collected. In the present technology, Phase 1 corresponds to the inflation and initial sweep measurement, Phase 2 corresponds to the transition to controlled deflation, Phase 3 corresponds to the refined measurement region where Korotkoff sounds are actively detected, and Phase 4 corresponds to the final deflation and measurement completion. The multi-panel visualization demonstrates how the blood pressure measurement system correlates cuff pressure data, oscillogram analysis, and neural network-based Korotkoff sound classification to determine accurate systolic and diastolic blood pressure values through the dual-phase measurement approach.
[0050] FIG. 7 illustrates a comparative analysis diagram depicting SBP evaluation results according to one or more embodiments of the present technology. The diagram presents two scatter plots that compare SBP values measured by the blood pressure measurement device of the present technology (prototype) against reference measurement methods. The left scatter plot compares prototype SBP measurements against arterial line (A-line) SBP measurements,Atorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT which represent an invasive gold standard for blood pressure measurement. The right scatter plot compares prototype SBP measurements against oscillometry SBP measurements, which represent a conventional non-invasive automated measurement approach. In each scatter plot, data points represent individual measurement comparisons, with the horizontal axis indicating the reference measurement value and the vertical axis indicating the difference of prototype measurement value and reference value. A diagonal reference line in each plot represents perfect agreement between the prototype and reference measurements, where data points falling on or near this line indicate high measurement accuracy. The comparative analysis demonstrates the performance of the neural network-based Korotkoff sound detection approach of the present technology relative to established blood pressure measurement techniques, allowing evaluation of measurement accuracy and consistency across different reference standards.
[0051] FIG. 8 illustrates a comparative analysis diagram depicting DBP evaluation results according to one or more embodiments of the present technology. The diagram presents two scatter plots that compare DBP values measured by the blood pressure measurement device of the present technology (prototype) against reference measurement methods. The left scatter plot compares prototype DBP measurements against arterial line (A-line) DBP measurements, which represent an invasive gold standard for blood pressure measurement. The right scatter plot compares prototype DBP measurements against oscillometry DBP measurements, which represent a conventional non-invasive automated measurement approach. In each scatter plot, data points represent individual measurement comparisons, with the horizontal axis indicating the reference measurement value and the vertical axis indicating the difference of prototype measurement value and reference value. A diagonal reference line in each plot represents perfect agreement between the prototype and reference measurements, where data points falling on or near this line indicate high measurement accuracy. Similar to the SBP evaluation results of FIG. 7, The comparative analysis of FIG. 8 demonstrates the performance of the neural network-based Korotkoff sound detection approach of the present technology for diastolic blood pressure determination relative to established blood pressure measurement techniques.
[0052] FIG. 9 illustrates a flowchart for a method 900 of measuring blood pressure according to one or more embodiments of the present technology. The method 900 employs a dual-phase measurement strategy that leverages neural network-based audio analysis during both an inflation phase and a deflation phase. In some embodiments, the dual-phaseAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT measurement strategy enables estimated rough blood pressure estimation during inflation of a cuff, followed by refined measurements during controlled deflation of the cuff. The method 900 may be performed by a blood pressure measurement device, such as the blood pressure measurement device described with reference to FIG. 1 or the automated blood pressure measurement system described with reference to FIGS. 2, 5A, 5B, and 5C.
[0053] As shown in FIG. 9, the method 900 begins with a step 910, where a cuff is inflated using a pump. For example, the piezopump 150 described with reference to FIG. 1 inflates an inflatable arm cuff bladder 210 to apply pressure to a user's arm. The method 900 also includes recording audio data during inflation of the cuff, at step 920. For example, a digital stethoscope / audio sensor 220 integrated within an arm cuff assembly 200 continuously captures audio signals produced by blood flow through an artery as cuff pressure increases. The method 900 continues to a step 930, where the recorded audio data is analyzed using a neural network to detect Korotkoff sounds during the inflation. For example, the blood pressure measurement system / microcomputer 130 implements a neural network configured to process the recorded audio data and classify detected sounds as Korotkoff sounds or non-Korotkoff sounds. The neural network may analyze audio windows extracted from the recorded audio data based on detected peaks in cuff pressure oscillations. The method 900 then proceeds to a step 940, where estimated rough SBP and DBP values are determined based on the detected Korotkoff sounds during inflation. For example, the blood pressure measurement system I microcomputer 130 analyzes a sequence of annotated oscillogram peaks to estimate rough SBP and DBP values based on a pattern of Korotkoff sound classifications and corresponding cuff pressure data. The rough SBP and DBP values estimated during the inflation phase may be used to identify a pressure region for refined measurements during a subsequent deflation phase.
[0054] Following the estimation of rough SBP and DBP values during the inflation phase, the method 900 proceeds to a step 950, where the cuff is deflated at a controlled rate. For example, deflating the cuff at a controlled rate can be done by controlling the first solenoid valve 160 to transition from the inflation phase to the deflation phase, and controlling the second solenoid valve 170 to regulate the rate of deflation. The method 900 continues to a step 960, where audio data is recorded during the deflation of the cuff. For example, the digital stethoscope I audio sensor 220 continuously captures audio signals produced by blood flow through an artery as cuff pressure decreases during the deflation phase. The method 900 then proceeds to a step 970, where the recorded audio data during the deflation is analyzed using the neural network to detect Korotkoff sounds. For example, the blood pressure measurementAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT system / microcomputer 130 processes audio windows extracted from the recorded audio data based on detected peaks in cuff pressure oscillations and classifies each audio window as a Korotkoff sound or a non-Korotkoff sound. The method 900 concludes with a step 980, where final SBP and DBP values are determined based on the detected Korotkoff sounds during deflation and corresponding cuff pressures. For example, the blood pressure measurement system / microcomputer 130 analyzes a sequence of annotated oscillogram peaks to determine the final SBP and DBP values based on a pattern of Korotkoff sound classifications, where the first appearance of Korotkoff sounds during deflation corresponds to SBP and the disappearance of Korotkoff sounds corresponds to DBP.
[0055] FIG. 10 illustrates a flowchart for a method 1000 of processing pressure and audio data during blood pressure measurement according to one or more embodiments of the present technology. The method 1000 describes sequential processing of pressure oscillations and audio data to identify and classify Korotkoff sounds, enabling correlation between detected pressure peaks and audio signals. As shown in FIG. 10, the method 1000 includes continuously monitoring timestamped pressure data from a pump during inflation and deflation of a cuff. For example, the blood pressure measurement system / microcomputer 130 continuously receives pressure data from the pressure sensor 180 during a blood pressure measurement session.
[0056] The method 1000 proceeds to a step 1020, where the timestamped pressure data is processed to extract cuff oscillations. For example, the blood pressure measurement system / microcomputer 130 filters the pressure signal to isolate oscillations caused by arterial pulsations within the inflatable arm cuff I bladder 210. The method 1000 continues to a step 1030, where peaks in the cuff oscillations are detected. For example, the blood pressure measurement system / microcomputer 130 analyzes the oscillogram to identify local maxima corresponding to individual heartbeats. In some embodiments, the blood pressure measurement system / microcomputer 130 may apply a peak detection algorithm that identifies peaks based on amplitude thresholds and minimum inter-peak intervals.
[0057] The method 1000 proceeds to a step 1040, where for each detected peak, a corresponding audio window is extracted from the recorded audio data. For example, the blood pressure measurement system I microcomputer 130 identifies the timestamp associated with each detected oscillogram peak and extracts a segment of audio data centered on or corresponding to the timestamp. The method 1000 continues to a step 1050, where the audio window is processed using the neural network to classify the audio window as a KorotkoffAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT sound or a non-Korotkoff sound. For example, the blood pressure measurement system / microcomputer 130 passes the extracted audio window to the neural network, and the neural network analyzes audio signal characteristics within the audio window to generate a classification output. The method 1000 concludes with a step 1060, where the corresponding peak is associated with the classification result. For example, the blood pressure measurement system / microcomputer 130 updates the oscillogram peak entry in memory with the classification result from the neural network, associating each detected pressure peak with a corresponding Korotkoff sound classification or non-Korotkoff sound classification for subsequent analysis.
[0058] In some embodiments, the method 1000 further includes analyzing a sequence of peaks having an associated classification result based on a pattern of Korotkoff sound classifications. The method 1000 also includes controlling a first solenoid valve to transition from the inflation to the deflation and controlling a second solenoid valve to regulate the rate of the deflation. Further, the method 1000 includes focusing on a pressure region identified during the inflation, and performing a slow deflation phase within the identified pressure region to refine determining of the final SBP and DBP values.
[0059] FIG. 11 illustrates a flowchart for a method 1100 for performing a terminating process in an automated blood pressure measurement system according to one or more embodiments of the present technology. The method 1100 addresses scenarios where the measurement process is stopped before a cuff pressure reaches an end pressure in response to detecting an interrupt condition. In some embodiments, the interrupt condition may include receiving a user abort input through a user interface, detennining a systolic blood pressure value during a sweep measurement mode, determining an accurate diastolic blood pressure value during a ramp measurement mode, or failing to detect significant oscillogram peaks for a predetermined duration during a measurement phase.
[0060] As shown in FIG. 11 , the method 1100 begins with a step 1110, where the blood pressure measurement system performs a terminating process before a cuff pressure reaches an end pressure in response to detecting an interrupt condition. For example, the intermpt condition may comprise one or more of several conditions that trigger early termination of the blood pressure measurement session. The method 1100 also includes a step 1120, where the blood pressure measurement system transmits a stop signal to a plurality of background processing threads. For example, upon detecting the intermpt condition at step 1110, the blood pressure measurement system / microcomputer 130 transmits a stop signal to each of theAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT plurality of background processing threads to initiate an orderly shutdown of concurrent operations.
[0061] The method 1100 further includes a step 1130, where the blood pressure measurement system stops the piezopump 150 to halt inflation or deflation operations of the arm cuff assembly 200. The method 1100 continues to a step 1140, where the blood pressure measurement system equalizes pressure in the inflatable arm cuff bladder 210 to atmospheric pressure, releasing any remaining pressure in the inflatable arm cuff bladder 210 for user comfort and safety. The method 1100 concludes with a step 1150, where the blood pressure measurement system waits for the plurality of background processing threads to complete cleanup operations and exit, ensuring that all concurrent processes are properly terminated and that collected measurement data is preserved before the blood pressure measurement system returns to an idle state.
[0062] In some embodiments, the interrupt condition of the method 1100 includes at least one of receiving a user abort input through a user interface, determining the SBP value during a sweep measurement mode, determining an accurate final DBP value during a ramp measurement mode, and failing to detect significant oscillogram peaks for a predetermined duration during a measurement phase. In some other embodiments, the method 1100 further includes saving collected measurement data to one or more timestamped files, the collected measurement data comprising audio data, pressure data, neural network output data, oscillogram peak information, final blood pressure values, and measurement settings. The method 1100 may also include generating and saving a visualization summarizing the blood pressure measurement, and closing a debug log file. In addition, the method 1100 may include displaying final blood pressure results in a user interface and re-enabling one or more user interface controls.
[0063] It is understood that the various disclosed embodiments may be implemented individually, or collectively, using devices comprised of various optical components, electronics hardware and / or software modules and components. These devices, for example, may comprise a processor, a memory unit, an interface that are communicatively connected to each other, and may range from desktop and / or laptop computers, to mobile devices and the like. The processor and / or controller can perform various disclosed operations based on execution of program code that is stored on a storage medium. The processor and / or controller can, for example, be in communication with at least one memory and with at least one communication unit that enables the exchange of data and information, directly or indirectly,Atorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT through the communication link with other entities, devices and networks. The communication unit may provide wired and / or wireless communication capabilities in accordance with one or more communication protocols, and therefore it may comprise the proper transmitter / receiver antennas, circuitry and ports, as well as the encoding / decoding capabilities that may be necessary for proper transmission and / or reception of data and other information.
[0064] Various information and data processing operations described herein may be implemented in one embodiment by a computer program product, embodied in a computer-readable medium, including computer-executable instructions, such as program code, executed by computers in networked environments. A computer-readable medium may include removable and non-removable storage devices including, but not limited to, Read Only Memory (ROM), Random Access Memory (RAM), compact discs (CDs), digital versatile discs (DVD), etc. Therefore, the computer-readable media that is described in the present application comprises non-transitory storage media. Generally, program modules may include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Computer-executable instructions, associated data structures, and program modules represent examples of program code for executing steps of the methods disclosed herein. The particular sequence of such executable instructions or associated data structures represents examples of corresponding acts for implementing the functions described in such steps or processes.
[0065] The foregoing description of embodiments has been presented for purposes of illustration and description. The foregoing description is not intended to be exhaustive or to limit embodiments of the present invention to the precise form disclosed, and modifications and variations are possible in light of the above teachings or may be acquired from practice of various embodiments. The embodiments discussed herein were chosen and described in order to explain the principles and the nature of various embodiments and its practical application to enable one skilled in the art to utilize the present invention in various embodiments and with various modifications as are suited to the particular use contemplated. While operations are depicted in the drawings in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed, to achieve desirable results. The features of the embodiments described herein may be combined in all possible combinations of methods, apparatus, modules, and systems.Atorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT
[0066] Only a few implementations and examples are described and other implementations, enhancements and variations can be made based on what is described and illustrated in this patent document.
Claims
Atorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT CLAIMSWhat is claimed is:
1. A blood pressure measurement device, comprising:a pump configured to inflate and deflate a cuff;a pump control system coupled to the pump and configured to manage operations of the pump, the pump control system including a microcomputer configured for neutral network inference and control of the pump;a plurality of solenoid valves coupled to the pump control system and controlled by the microcomputer to regulate cuff pressure;an audio sensor configured to detect Korotkoff sounds; anda neural network implemented on the microcomputer, wherein the neural network is configured to process audio data from the audio sensor to classify detected sounds as Korotkoff sounds or non-Korotkoff sounds during both inflation and deflation of the cuff.
2. The blood pressure measurement device of claim 1 , comprising a touchscreen interface operatively connected to the pump control system for user interaction and system control, wherein the pump is a piezopump.
3. The blood pressure measurement device of claim 1, wherein the plurality of solenoid valves comprises:a first solenoid valve configured to control transition between the inflation and the deflation of the cuff, anda second solenoid valve configured to control a rate of the deflation.
4. The blood pressure measurement device of claim 1, wherein the microcomputer is configured to:estimate a rough diastolic blood pressure (DBP) and systolic blood pressure (SBP) values during the inflation of the cuff, anddetermine refined DBP and SBP values during a slow deflation phase focused on a pressure region identified during the inflation of the cuff.
5. The blood pressure measurement device of claim 4, wherein the microcomputer is configured to continuously perform operations, including:Atorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT recording audio data from the audio sensor and monitoring timestamped pressure data from the pump,processing the timestamped pressure data to extract cuff oscillations, and detecting peaks in the cuff oscillations.
6. The blood pressure measurement device of claim 5, wherein the microcomputer is configured to:for each detected peak, extract a corresponding audio window from the recorded audio data,process the audio window using the neural network to classify it as a Korotkoff sound or a non-Korotkoff sound, andannotate a corresponding peak of the detected peaks with classification result.
7. The blood pressure measurement device of claim 6, wherein the microcomputer is configured to analyze a sequence of annotated peaks to determine the rough DBP and SBP values during the inflation and to determine the refined DBP and SBP values during deflation, based on a pattern of Korotkoff sound classifications.
8. A system for automated blood pressure measurement, comprising:an arm cuff with an integrated digital stethoscope:a piezopump operatively connected to the arm cuff;a microcomputer configured to control the piezopump; anda user interface,wherein the microcomputer is configured to:control the piezopump to inflate and deflate the arm cuff,record audio data from the integrated digital stethoscope during inflation and deflation of the arm cuff,analyze the audio data using a neural network to detect Korotkoff sounds, and determine blood pressure values based on the detected Korotkoff sounds and corresponding cuff pressures.
9. The system of claim 8, comprising:a first solenoid valve configured to control transition between the inflation and the deflation of the arm cuff; andAtorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT a second solenoid valve configured to control a rate of the deflation of the arm cuff.
10. The system of claim 9, wherein the microcomputer is configured to:estimate rough diastolic blood pressure (DBP) and systolic blood pressure (SBP) values during the inflation of the arm cuff; anddetermine refined DBP and SBP values during a slow deflation phase focused on a pressure region identified during the inflation of the arm cuff.
11. A method of measuring blood pressure, comprising:inflating a cuff using a pump;recording audio data during inflation of the cuff;analyzing recorded audio data using a neural network to detect Korotkoff sounds during the inflation;determining estimated systolic blood pressure (SBP) and diastolic blood pressure (DBP) values based on the detected Korotkoff sounds during inflation;deflating the cuff at a controlled rate;recording audio data during the deflation of the cuff;analyzing the recorded audio data during the deflation using the neural network to detect Korotkoff sounds; anddetermining final SBP and DBP values based on the detected Korotkoff sounds during deflation and corresponding cuff pressures.
12. The method of claim 11, further comprising:continuously monitoring timestamped pressure data from the pump during inflation and deflation of the cuff;processing the timestamped pressure data to extract cuff oscillations; and detecting peaks in the cuff oscillations.
13. The method of claim 12, further comprising:for each detected peak, extracting a corresponding audio window from the recorded audio data;processing the audio window using the neural network to classify it as a Korotkoff sound or a non-Korotkoff sound; andassociating the corresponding peak with classification result.Atorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT 14. The method of claim 13, wherein determining the estimated SBP and DBP values and determining the final SBP and DBP values comprises analyzing a sequence of peaks having an associated classification result based on a pattern of Korotkoff sound classifications.
15. The method of claim 11, wherein deflating the cuff at a controlled rate comprises: controlling a first solenoid valve to transition from the inflation to the deflation; and controlling a second solenoid valve to regulate the rate of the deflation.
16. The method of claim 15, wherein determining the final SBP and DBP values comprises:focusing on a pressure region identified during the inflation; andperforming a slow deflation phase within the identified pressure region to refine determining of the final SBP and DBP values.
17. The method of claim 11, further comprising a terminating process before a cuff pressure reaches an end pressure in response to detecting an interrupt condition.
18. The method of claim 17, wherein the interrupt condition comprises at least one of: receiving a user abort input through a user interface,determining the SBP value during a sweep measurement mode,determining an final DBP value during a ramp measurement mode, andfailing to detect significant oscillogram peaks for a predetermined duration during a measurement phase.
19. The method of claim 17, further comprising:after completion or the terminating process, transmitting a stop signal to a plurality of background processing threads,stopping the pump,equalizing pressure in the cuff to atmospheric pressure, andwaiting for the plurality of background processing threads to complete cleanup operations and exit.
20. The method of claim 19, further comprising:Atorney Docket No.: 009062.8598. WOOO Client Ref. No.: SD2025- 278-2PCT saving collected measurement data to one or more timestamped files, the collected measurement data comprising audio data, pressure data, neural network output data, oscillogram peak information, final blood pressure values, and measurement settings, generating and saving a visualization summarizing measuring of blood pressure, closing a debug log file, anddisplaying final blood pressure results in a user interface and re-enabling one or more user interface controls.
