EP4626307A1

EP4626307A1 - Mobile device based oscillometric blood pressure measurement

Info

Publication number: EP4626307A1
Application number: EP23898760.6A
Authority: EP
Inventors: Colin Barry; Edward Jay Wang
Original assignee: University of California; University of California Berkeley; University of California San Diego UCSD
Current assignee: University of California; University of California Berkeley; University of California San Diego UCSD
Priority date: 2022-11-29
Filing date: 2023-11-28
Publication date: 2025-10-08
Also published as: KR20250114382A; WO2024118687A1; CN120897699A; JP2025537957A

Abstract

Disclosed are systems, devices and methods for mobile device based oscillometric blood pressure measurement. In some aspects, a mobile device includes a vibration motor, an inertial motion unit (IMU), a camera, a processor, and non-transitory computer-readable memory, in which the processor is configured to perform operations including: causing the vibration motor to vibrate; acquiring, using the inertial measurement unit of the mobile device, IMU data indicative of vibration damping that is caused by a user, using a body part, applying a force on the camera; determining the applied force based on a force model and the IMU data; capturing, using the camera of the mobile device, photoplethysmography (PPG) data of a set of blood vessels in the body part concurrently with the application of the force; and determining a blood pressure of the user using an oscillometric technique based on the PPG data and the applied force.

Description

MOBILE DEVICE BASED OSCILLOMETRIC BLOOD PRESSURE MEASUREMENT

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This patent document claims priority to and benefits of U.S. Provisional Patent Application No. 63/385,383 entitled “SMARTPHONE-BASED BLOOD PRESSURE MEASUREMENT” filed on November 29, 2022. The entire content of the aforementioned patent application is incorporated by reference as part of the disclosure of this patent document.

TECHNICAL FIELD

[0002] This patent document relates to mobile device based oscillometric blood pressure measurement.

BACKGROUND

[0003] The American Heart Association (AHA) and the American Medical Association (AMA) released a joint statement outlining the importance of integrating at-home blood pressure self-measurement and distinctly stating that the oscillometric method of calculating blood pressure is preferred. The oscillometric method is a standard calculation performed by most FDA approved blood pressure cuffs.

SUMMARY

[0004] The present document discloses methods, devices, and systems for blood pressure measurement based on the oscillometric technique using sensor data acquired by built-in sensors of mobile devices with no additional attachments.

[0005] An aspect of the present document relates to a mobile device configured to measure blood pressure based on the oscillometric technique using sensor data acquired by built-in sensors of the mobile device without attachments. The mobile device may include: a vibration motor, an inertial motion unit (IMU), a camera, a processor, and non- transitory computer-readable memory that stores instructions that, when executed by the processor, cause the processor to perform at least: cause the vibration motor to vibrate; acquire, by the inertial measurement unit of the mobile device, IMU data indicative of vibration damping that is caused by a user’s body part applying a force on the camera; determine the applied force based on a force model and the IMU data; capture, by the camera of the mobile device, photoplethysmography (PPG) data of a set of blood vessels in the body part concurrently with the application of the force; and determine a blood pressure of the user using an oscillometric technique based on the PPG data and the applied force.

[0006] An aspect of the present document relates to a method for measuring blood pressure based on the oscillometric technique using sensor data acquired by built-in sensors of the mobile device without attachments. The method may include: causing a vibration motor of the mobile device to vibrate; acquiring, using an inertial measurement unit (IMU) of the mobile device, IMU data indicative of vibration damping that is caused by a user’s body part applying a force on a camera of the mobile device; determining the applied force based on a force model and the IMU data; capturing, using the camera of the mobile device, photoplethysmography (PPG) data of a set of blood vessels in the body part concurrently with the application of the force; and determining the blood pressure of the user using an oscillometric technique based on the PPG data and the applied force.

[0007] A further aspect of the present document relates to one or more non-transitory computer-readable media storing instructions that, when executed by one or more processors of a mobile device, cause the mobile device to perform any one or more of the solutions described herein.

[0008] The above and other aspects of the present document and their implementations and applications are described in greater detail in the drawings, the description, and the claims.

BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 shows a diagram of an example system implementing the disclosed technology in accordance with the disclosed technology.

[0010] Figure 2 illustrates an exemplary block diagram of the various components of a mobile device in accordance with some embodiments of the present document.

[0011] Figure 3 illustrates an example application for a mobile device based blood pressure measurement in accordance with some embodiments of the present disclosure. [0012] Figure 4 illustrates a flowchart of a process for a mobile device based blood pressure measurement in accordance with some embodiments of the present document. [0013] Figure 5 illustrates regression and Bland Altman plots for force measurements of three different smartphones in accordance with some embodiments of the present document.

[0014] Figure 6(a) illustrates exemplary PPG and applied force data measured by a mobile device and a corresponding oscillogram generated in accordance with some embodiments of the present document.

[0015] Figure 6(b) illustrates average oscillograms corresponding to various blood pressure ranges in accordance with some embodiments of the present document.

[0016] Figure 7 illustrates a confusion matrix and ROC analysis of the results in Figures 6(a) and 6(b) in accordance with some embodiments of the present document.

DETAILED DESCRIPTION

[0017] The present document describes systems, devices, and methods for measuring blood pressure using the oscillometric method in substantially the same way a blood pressure cuff functions. This methodology is applicable to any mobile device with a camera, a vibration motor, and an inertial measurement unit (IMU). For determining blood pressure using the oscillometric method, a device needs to: 1 ) apply a known force on a blood vessel and simultaneously 2) measure the local blood volume. For a blood pressure cuff, the inflatable cuff applies pressure to a set of blood vessels, while an integrated sensor measures the change in the blood volume within the set of blood vessels. For the mobile phone application disclosed herein, a user applies force by pressing a body part (e.g., a finger of the user) on the camera of the mobile device, while the camera may utilize photoplethysmography (PPG) to measure the blood volume within the body part concurrently with the force application. The quantification of the applied force can be achieved by inducing vibrations in the mobile device through its internal vibration motor, and monitoring the damping of these vibrations, resulting from the applied force, using the mobile device's IMU. The mobile device may determine a blood pressure of the user based on the applied force and the PPG data acquired using the built-in sensors already in the mobile device, needs no hardware modifications, no physical attachments, nor any prior blood pressure measurements for calibration, and can be enabled solely with software by downloading an application on any mobile device, as discussed in further detail below. Accordingly, the technology may provide large scale screening opportunities for hypertension. As a software enabled measurement for any mobile device with a camera, IMU, and a vibration motor, this technology can be quickly proliferated into communities without sufficient blood pressure monitoring.

[0018] Widespread blood pressure screening, is a critical step towards improving healthcare worldwide. Hypertension, defined for this paper as a systolic and/or diastolic blood pressure of over 130/80 mmHg, long served as a major risk factor for heart disease, strokes, and chronic kidney disease. More recently, hypertension has been demonstrated to accelerate cognitive decline for middle-aged and older adults, with growing evidence that hyper- tension is associated with higher risk of all-cause mild cognitive impairment (MCI) and Non-Amnestic MCI. Given that the World Health Organization estimates nearly half (46%) of hypertensive adults are unaware of their condition, BP screening is imperative to reach the United Nations goal of reducing hypertension prevalence by 30% between 2010 and 2030. Furthermore, enabling large scale screening could vastly improve health literacy and empower at-risk communities to leverage non- traditional healthcare settings address hypertension. A mobile device based blood pressure measurement may provide a highly extensible and cost-efficient screening solution due to the widespread availability of mobile devices.

[0019] Most prior research into smartphone and smartwatch blood pressure measurement preform relative blood pressure measurements using optical methods that need only PPG measurements including, e.g., pulse transit time (PTT) or pulse wave analysis (PWA). However, these techniques are fundamentally different than the mobile device based oscillometric blood pressure measurement technology disclosed herein. PTT and PWA only provide a relative change in blood pressure but cannot provide the valuable systolic and diastolic blood pressure measures commonly used by physicians. Instead, a blood pressure cuff measurement need to be performed daily or weekly to calibrate the measurement. These measures of relative change assume an individual has access to a blood pressure cuff (or another absolute blood pressure measurement); otherwise, the relative measurements cannot be meaningfully interpreted. Existing smartphone based absolute blood pressure measurement often employs custom built attachments (e.g., an added force sensor) to measure force.

[0020] The mobile device based oscillometric blood pressure measurement as disclosed herein needs no attachment and functions on mobile devices of any models with a vibration motor, camera, and IMU. The oscillometric blood pressure measurement is not an “optical” blood pressure measurement relying on pulse transit time (PTT), pulse arrival time (PAT), pulse wave analysis (PWA), or any other technique estimating the relative change in blood pressure. The application may perform absolute blood pressure measurements based on the oscillometric method used in clinical blood pressure cuffs and recommended by the American Medical Association (AMA) and the American Heart Association (AHA). As discussed elsewhere in the present document, blood pressure measurement based on the oscillometric technique needs measuring two key metrics: applied force and local blood volume. Local blood volume of an artery can be accurately estimated via photoplethysmography (PPG) with the smartphone camera. Measuring an applied force with no attachments is a key insight to enable the blood pressure measurement as disclosed herein. By relying on components common to mobile devices of different types, the oscillometric blood pressure measurement as disclosed herein is functional across mobile devices of different types (e.g., a type may correspond to a model by a manufacturer).

[0021] Some embodiments of the present document disclose a force damping technique utilizing only the vibration motor and IMU of a mobile device to measure an applied force. Most mobile devices already contain an IMU with an accelerometer and gyroscope for other purposes including, e.g., to measure the linear acceleration and angular velocity that enables position inferences by such mobile devices widely used for rotating the screen, playing games, counting steps, gesture recognition, and much more. Similarly, most mobile devices already contain a vibration motor to provide haptic feedback for messages, phone calls, and screen interactions, etc. To measure force, the mobile device may be set to vibrate (e.g., maximally vibrate) while the IMU measures the motion as the device oscillates (from the vibration). When force is applied during vibration, the IMU signal changes as the oscillations are altered or dampened. According to embodiments of the present documents, the vibration dampening is measured and relied upon to model the applied force as a damping force on the vibration.

[0022] According to some embodiments of the present document, IMU data that includes multi-axis accelerometer data and multi-axis gyroscope data collectively representing a vibration damping is used to determine a corresponding force that causes the vibration damping. This method recognizes a coupling between different axes of gyroscope and accelerometer data that respond differently to an applied force. Due to the conservation of energy, an applied force does not solely cause vibration damping or decreasing in amplitude along all IMU axes. Instead, some axes may even increase as the applied force to damp the vibration motion in one axis may cause the vibration energy to be transmitted into another axis. The disclosed method takes multiple inputs from both the accelerometer and gyroscope as inputs, investigating frequency, amplitude, and relationships between different axes.

[0023] An applied force causing vibration damping of the mobile device may be quantified based on a force model trained to correlate (1 ) IMU data measured by the mobile device that is indicative of vibration damping of the mobile device with (2) forces that are applied to the mobile device and cause the corresponding vibration damping. The force model may include a machine learning algorithm, e.g., a multivariate linear regression algorithm. The correlation may depend on the configuration of the mobile device. Damping of vibration may depend on one or more factors that contribute to how the mobile device absorbs and dissipates vibrational energy. Examples of such factors include the strength of the vibration, the load distribution within the mobile device (including, e.g. , how the weight and components are distributed within the mobile device), materials used in the mobile device (including, e.g., density, elasticity, and internal friction), the overall design (including, e.g., shape and structural integrity, features such as ribs, gussets), interaction between components within the device, or the like, or a combination thereof. Moreover, the damping as measured by the IMU of the mobile device may depend on one or more factors including, e.g., the position of the IMU relative to the vibration motor(s), sensor parameters including, e.g., sensitivity, resolution, noise level, or the like, or a combination thereof. A cross device reliability study disclosed herein demonstrates that the technology disclosed herein provide similar force estimation performance across all smartphone models that were tested, indicating that the technology is versatile and translatable across mobile devices. A force model may be determined for a type of mobile devices (e.g., a model by a manufacturer) and used to calibrate mobile devices of the same type.

[0024] The calibration step may allow a standardized measurement across mobile devices of different types. The calibration may ensure that different mobile devices of a same type or different types are configured to measure force consistently. As such, different mobile devices of a same type or different types can use a same blood pressure estimation model, even though the force model (also referred to as a force estimation algorithm) may be different. This is beneficial because the developing the force model is an automated or at least partially automated process that can be completed within a short period of time (e.g., a day) and does not take large participant recruitment or clinical measurements. On the other hand, developing a blood pressure model may take many participants with different demographics to undergo clinical blood pressure measurements, which can take months or longer.

[0025] Figure 1 shows an exemplary system implementing the disclosed technology of mobile device based oscillometric blood pressure measurement. The system 100 includes a mobile device 102 that can include at least one of a camera 104, a vibration motor 105, an IMU 106, a processor 108, a wireless transmitter 110, and a display 112. The processor 108 can control operations of the mobile device 102 (e.g., causing the vibration motor 105 to vibrate, causing the camera 104 and/or the IMU 106 to acquire data), receive and process sensor data (e.g., image data acquired by the camera 104, IMU data acquired by the IMU 106), and run algorithms (e.g., a force model) on the sensor data and generate results (e.g., an applied force to bring about vibration damping, a force plot, a PPG plot, blood pressure, diagnosis based on the determined blood pressure, etc.). As shown in Figure 1 , the mobile device 102 may communicate with a user 114 or an external device or system (e.g., a cloud server 116). For example, the mobile device 102 can send a report of the state of the system to a cloud server 116 using, for example, a wireless transmitter 110. As another example, the mobile device 102 may communicate with the user 114 via the display 112. The display 112 may be a touch screen configured as a graphical user interface such that the mobile device 102 may present data or results to the user 114 via the display 112 and receive user input via the display 112. In some embodiments, the mobile device 102 may be a smartphone, a tablet, etc.

[0026] Figure 2 illustrates an exemplary block diagram of the various components of a mobile device in accordance with some embodiments of the present document. The mobile device 200 is an example of the mobile device 102 as illustrated in Figure 1. In some embodiments, the mobile device 200 may be a smartphone, a tablet, etc. In some embodiments, a mobile device 200 can vibrate driven by a built-in vibration motor 203, measure vibration damping caused by a user applying a force on the mobile device 200, capture PPG data recording a change of blood volume in a set of blood vessels caused by the applied force, and estimate a blood pressure using a force model based on the PPG data and the applied force. The mobile device 200 includes one or more sensors 202 that can gather data, a processing unit 204 connected to the one or more sensors 202 and capable of executing a force model on the gathered data, a wireless transceiver 206 connected to the processing unit 204, and a display 208 connected to the processing unit 204. The one or more sensors 202 may include one or more of a camara or an IMU. An IMU of the mobile device 200 may include an accelerometer and a gyroscope.

[0027] Figure 3 illustrates an example application for a mobile device based blood pressure measurement in accordance with some embodiments of the present disclosure. For illustration purposes only and not intended to be limiting, a smartphone application is shown. It is understood that the mobile device based blood pressure measurement as disclosed herein may be implemented on other mobile devices including, e.g., a tablet.

[0028] Figure 3(a) illustrates a simplified diagram of a smartphone force damping. The smartphone includes a vibration motor, a camera, an IMU, and a display. The vibration motor drives oscillatory motion. A user applies a damping force by pressing a finger on the camera. The IMU of the smartphone records the motion (the vibration) and the smartphone camera records PPG concurrently. The base or opposite side of the fingernail presses on the camera of the smartphone such that the camera of the smartphone can measure real-time blood volume within the finger while the finger applies the force by pressing on the camera.

[0029] Figure 3(b) illustrates the vibration motor movement along the Z-axis (perpendicular to the display of the smartphone) in (I), the applied finger force along the Z-axis in (II), and the linear Z-axis acceleration of smartphone IMU are plotted vertically in (III). As illustrated, as the applied force increases, the acceleration of the oscillation or vibration amplitude of the Z-axis decreases. The raw PPG signal in (IV) is also plotted vertically to show how an increase in the applied force affects the blood volume in (V). [0030] The IMU of the smartphone may include an accelerometer and a gyroscope. Although only the linear acceleration along the Z-axis is illustrated (in (III)), the IMU of the smartphone measures acceleration in multiple directions (e.g., three perpendicular directions including the direction along the Z-axis and the two directions in a plane perpendicular to the Z-axis), and also gyroscope data in multiple directions (e.g., the same three directions as the accelerometer. The vibration, and damping thereof, of the smartphone may be monitored based on the IMU data including the multi-axis accelerometer data in combination with the multi-axis gyroscope data. Due to the way vibration energy dissipates, the linear accelerations of the IMU in various axes (indicative of the smartphone motion) may change in different patterns. For example, as illustrated in (II) and (III), while the applied force increases, the linear acceleration along the Z-axis decreases; however, the linear acceleration along a different direction may decrease at a different rate or even increase in some portion(s) of the period the damping force is applied. Similarly, the angular velocities of the IMU along different directions (indicative of the smartphone motion) may change in different patterns in response to the damping force the user applies. Accordingly, the IMU data including the multi-axis accelerometer data in combination with the multi-axis gyroscope data may be used in determining the applied force. See additional description in this regard elsewhere in the present document.

[0031] Figure 3(c) illustrates a graphical user interface (GUI) of the smartphone application with visual signifiers to guide a user to perform a blood pressure measurement using the application. A first visual signifier includes an image of a fingertip near the smartphone camera that may provide intuitive guidance for the user on how to position the finger. A second visual signifier includes a plot of the applied force in real time overlaid with a force guide line during the force damping measurement. The force guide may aid the user to apply a correct amount of force during the measurement.

[0032] The GUI may also include a plot of the acquired PPG signal during data collection. The acquired PPG signal may provide the user or someone else (e g., the research staff) with helpful information during data collection. The PPG signal may be plotted in real time before and during the measurement so the user and/or the research staff can ensure data quality. For example, if the PPG signal is not within the desired range, the user or the research staff can press a button to calibrate the PPG signal before the measurement to ensure the PPG signal is not over- or under- saturated. At the end of the measurement, the full force and PPG signals may be plotted within the application to provide an overview of the data. The GUI may also present predicted results including, e.g., systolic blood pressure (sysBP), diastolic blood pressure (diaBP), heartrate (HR), etc. The GUI may further present user identify (e.g., user ID), the number (or count) of attempts by the user. As illustrated, the smartphone is oriented upside down so that the camera of the smartphone is close to the user to make it convenient for the user to place a finger on the camera for force application and PPG signal acquisition.

[0033] Additionally or alternatively, the application may provide guidance to the user on how to use the application and the mobile device to perform a blood pressure measurement. Such guidance information may be in the form of text, audio, video, image, or the like, or a combination thereof. For example, the GUI may remind the user that a software update or calibration is needed, that the mobile device should be placed on a specific type of surface (e.g., a flat wooden surface, like a desk or table), that a surface calibration (as described elsewhere in the present document) is needed, that the user needs to follow a protocol. The protocol may be substantially the same as a standard blood pressure procedure including that the user needs to sit upright with the measurement device (the mobile device) at heart height, that the user’s feet need to be flat on the ground, that the user should remain calm and breathe normally during the measurement, that for the mobile device based measurement, the user needs to place the mobile device flat on the surface with the screen facing up and the front facing camera closest to the user (so that the user can press the camera conveniently and view the GUI of the application at the same time). The application may provide instructions and/or guidance on how to properly perform the measurement. For example, the instructions/guidance may include that the user places a hand flat on the surface with an index finger over the front facing camera of the mobile device. [0034] Figure 3(d) illustrates a graph showing the correlation between the amplitude of blood volume oscillations plotted against the applied finger pressure. The approximate points of Diastolic Blood Pressure (DBP), Mean Arterial Pressure (MAP), and Systolic Blood Pressure (SBP) are indicated for, e.g., interpretability.

[0035] Figure 4 illustrates a flowchart of a process for a mobile device based blood pressure measurement in accordance with some embodiments of the present document. The process 400 may be implemented on a mobile device (e.g., the mobile device 102, the mobile device 200).

[0036] At block 410, the process 400 includes causing a vibration motor of the mobile device to vibrate. The vibration motor may be one already built-in in the mobile device. For example, the vibration may be driven by the vibration motor that is used to provide haptic feedback for messages, phone calls, and screen interactions, etc. The vibration motor may be caused to vibrate at a specific level (e.g. , at a maximal level) of the vibration motor.

[0037] In some embodiments, the mobile device may be placed on a specific type of surface to perform the oscillometric blood pressure measurement, considering that the vibration damping relates to characteristic of the surface. For example, the process 400 may include providing information to guide or remind that the mobile device be placed on such a type of surface (e.g., a wooden flat surface such as a table, desk, etc.) as part of a preparation of a blood pressure measurement. The process 400 may include providing additional information including, e.g., how a user should sit, apply a force, or the like, or a combination thereof, as part of the preparation.

[0038] In some embodiments, the process 400 may include performing a surface calibration by measuring the characteristics of the surface where the mobile device is placed. Example characteristics of the surface include flatness, level, material, or the like, or a combination thereof. In some embodiments, the mobile device may measure whether the surface is level using the built-in sensors including the IMU (including the accelerometer and/or the gyroscope) by detecting the direction of gravity relative to the orientation of the mobile device, and/or the angle and direction of tilt of the surface. In some embodiments, the mobile device may identify the material of the surface. Merely by way of example, the process 400 may include causing the mobile phone placed on the surface to vibrate, driven by the built-in vibration motor, at a vibration mode with a known vibration profile, measuring the vibration of the mobile device, and determining the material of the surface by comparing the measured vibration profile with the known vibration profile. As another example, the process 400 may include causing the mobile phone placed on the surface to vibrate, driven by the built-in vibration motor, at different vibration modes with known vibration profiles, measuring the vibration profiles of the mobile device, and determining the material of the surface by comparing a change in the measured vibration profiles with the changes in the known vibration profiles. Additionally or alternatively, the process 400 may include identifying the characteristics of the surface based on one or more images of the surfaces. The process 400 may including performing a surface calibration based on the characteristics of the surface when the applied force is determined as described below.

[0039] At block 420, the process 400 includes acquiring, using an inertial measurement unit (IMU) of the mobile device, IMU data indicative of vibration damping that is caused by a user’s body part applying a force on a camera of the mobile device. The body part may be a tip portion of a finger of the user, e.g. , an index finger of the user. In some embodiments, the process 400 may include providing a visual guide to be presented on a display of the mobile device. The visual guide (also referred to as a visual signifier) may include indicia to guide the user in placing a surface of the body part onto the camera.

[0040] The IMU of the mobile device may include a multi-axis accelerator and a multiaxis gyroscope. The vibration motor may be one already built-in in the mobile device for other purposes including, e.g., to measure the linear acceleration and angular velocity that enables position inferences by such mobile devices widely used for rotating the screen, playing games, counting steps, gesture recognition, etc.

[0041] The IMU data may include multi-axis accelerometer data and multi-axis gyroscope data. For example, the IMU data may include three-axis accelerometer data and three-axis gyroscope data. The three-axis accelerometer data may include linear accelerations of the IMU in various axes (indicative of the smartphone motion). Depending on how energy dissipates during vibration damping, linear accelerations of the IMU may change in different patterns. For example, the linear accelerations along different directions may decrease at different rates or even following opposite trends (e.g. , the linear acceleration in one direction increases, while the linear acceleration in one direction decreases) in some portion(s) of the period the damping force is applied. Similarly, the angular velocities of the IMU along different directions (indicative of the smartphone motion) may change in different patterns in response to the damping force the user applies. Accordingly, the IMU data including the multi-axis accelerometer data and the multi-axis gyroscope data may be used in combination to provide a comprehensive representation of the vibration damping and therefore improve the accuracy of the determined applied force. As used herein, “damping” or “vibration damping” refers to recorded alterations in the device’s motion that result from the applied force.

[0042] Each axis of the multi-axis accelerometer data and the multi-axis gyroscope data may include a time-series signal. A signal of an axis of the IMU data may contain high frequency components from the vibration motor and low frequency components corresponding to noise. In some embodiments, the raw IMU data may be sampled from the mobile device at a maximal rate with no on-device filtering or post processing. In some embodiments, the raw IMU data may be subjected to a series of processing. For example, the raw IMU data may be processed using one or more of the following techniques: low pass filtering, high pass filtering, bandpass filtering, savgol filtering, standard deviation, or empirical mode decomposition. The filtering may also involve multiple steps as in the case of a filter bank or multiple stages of different filter types. From the filtered or raw IMU data, further features can be obtained, such as a signal power for a combination of axes. One or more features may be evaluated using a variety of metrics including, e.g., a principal component analysis, recursive feature elimination, LASSO regression, and correlation.

[0043] At block 430, the process 400 includes determining the applied force based on a force model and the IMU data.

[0044] In some embodiments, the process 400 may include decomposing the IMU data based on an empirical mode decomposition (EMD) technique. According to the EMD technique, the process 400 may include for each axis of the multi-axis accelerometer data and the multi-axis gyroscope data, decomposing a signal of the axis into a plurality of intrinsic mode functions (IMFs) representing different frequency components of the signal, so that the applied force may be determined based on at least a portion of the IMFs that correspond to various axes of the multi-axis accelerometer data and the multi-axis gyroscope data.

[0045] Given that the high frequency components of the signals of various axes from the vibration motor are the strongest components of the corresponding signals, the first IMFs of various axes of the IMU data primarily correspond to the vibration motor signal. The upper envelope of the first IMF of each axis of the multi-axis accelerometer data and the multi-axis gyroscope data may be used as a feature for identifying the applied force. In some embodiments, the upper envelope of the first IMF on each IMU axis serves as a representative feature, and the first IMFs collectively may most significantly correlate to the force data and therefore used as input to a force model to determine the applied force, while other features may be excluded. Accordingly, merely by way of example, the process 400 may include for each axis of the multi-axis accelerometer data and the multiaxis gyroscope data, identifying, from the plurality of IMFs of the signal of the axis, a first IMF that corresponds to a highest frequency; and identifying an upper envelope of the first IMF of the axis; and inputting into the force model the upper envelopes of the first IMFs that respectively correspond to all axes of the multi-axis accelerometer data and the multi-axis gyroscope data.

[0046] The EMD technique for processing the IMU data is described here for illustration purposes and not intended to be limiting. As described elsewhere in the present document, before being input into the force model, the IMU data may be processed by at least one of filtering, a filter bank, averaging, standard deviation, wavelet decomposition, spectral analysis, or EMD, etc. In some embodiments, signals of at least two different axes of the IMU data (including the multi-axis accelerometer data and the multi-axis gyroscope data) may be processed using different techniques. In some embodiments, the raw IMU data may be directly input into the force model.

[0047] The force model may include a machine learning model trained to correlate (1 ) IMU data measured by the mobile device that is indicative of vibration damping of the mobile device with (2) forces that are applied to the mobile device and cause the corresponding vibration damping. The force model may be a data driven model including parametric modeling, a linear regression model, an ensemble learning model (random forest, adaboost, etc.), a support vector machines model, a neural network (e.g., a transformer model, a convolutional neural network (CNN), etc.), or the like, or a variation thereof, ora combination thereof. Merely by way of example, the force model may include a multivariate linear regression model that provides a value of the applied force on a continuous scale.

[0048] The force model may be trained using training data including applied forces measured using force sensors (e.g., force sensitive resistor (FSR)) and measured dampings caused by the applied forces. An example test case was performed involving five participants and three smartphones of different types including Google Pixel 4, Samsung Galaxy A53, and Motorola Moto G Power from a variety of smartphone manufacturers, physical shapes, costs, and embedded components. In the example test case, each of a group of five participants applied pressure with an index finger onto a 0.3 mm thick force sensor positioned on top of a front camera of a smartphone. A participant applied force with an index finger in the same vertical positioning as the blood pressure measurement to account for hydrostatic force. The force sensor, a calibrated, linear force sensor (SingleTact 15 mm 4.5N), was placed over the smartphone front camera. A participant pressed a finger on a smartphone force sensor (and smartphone camera) to measure the applied force of the finger. In the example case, a smartphone was placed on a flat wooden desktop surface during measurement. During the measurement, realtime readings from the force sensor were displayed on a screen in front of the user with a force guide. The user was instructed to use the real time feedback to follow the force guide in applying a range of pressures. Similar to a cuff-based blood pressure measurement, the force guide instructed users to continuously increase the applied force. The study included three repetitions of forty-second session of applying a range of ferees. These three repetitions are performed by each participant on each of three smartphones of different types for a resulting of ten minutes of raw applied force data on each phone (two minutes per participant). The IMU data acquired were processed based on EMD in substantially the same way as described above to provide vibration damping data. The forces measured using the force sensor and the corresponding vibration damping data were used to train and validate the force model. The mean absolute error, correlation coefficient, and bias of each smartphone is displayed in Table 1 and Figure 5. In Figure 5, solid dots represent data corresponding to Motorola Moto G Power, crosses represent data corresponding to Samsung Galaxy A53, and stars represent data corresponding to Google Pixel 4. The average correlation coefficient across all phones is 0.86. The Google Pixel 4 smartphone used in the blood pressure validation study has the lowest mean absolute error (MAE).

Table 1.

[0049] As discussed elsewhere in the present document, the vibration damping behavior of a mobile device may depend on its configuration. The results from the example test case suggest that relevant differences in force determination by mobile devices of different types (e.g., a type may correspond to a model by a manufacturer) may be calibrated using different force models. A force model may be trained when a mobile device of a new type (e.g., a new model by a manufacture) is released and the obtained force model may be used to calibrate mobile devices of the same type. The type of a mobile device my correspond to a model and/or a manufacturer of the mobile device and associated with factors including, e.g., a configuration of the vibration motor, a configuration of the IMU, a configuration of the camera, or the like, or a combination thereof, of the mobile device.

[0050] A model library maya be built by collecting such force models corresponding to mobile devices of different types. In some embodiments, the process 400 may include obtaining, from such a model library, the force model based on a type of the mobile device as a preparation of a blood pressure measurement procedure.

[0051] The process 400 may determine the applied force substantially real time. In some embodiments, during the measurement, the process 400 may include providing a visual guide to be presented on a display of the mobile device. The visual guide may include a force plot that illustrates the applied force in substantially real time. In some embodiments, the visual guide may further include a force guide line overlaid with the applied force. The user may be instructed to use the real time feedback to follow the force guide in applying a range of forces. Merely by way of example, the force guide may instruct a user to continuously increase the applied force within the range, similar to how a cuff-based blood pressure measurement is performed.

[0052] In some embodiments, the process 400 may include measuring an area of a surface of the body part, in which the force is applied by pressing the surface of the body part against the camera. Merely by way of example, the process 400 may cause the camera to take a picture of the surface of the body part when the body part is pressed against the camera to apply the damping force, and determine the area based on the image. As another example, the mobile device includes a display including a touch screen configured to sense contact by the user, and the process 400 may determine the area by identifying a contact area of the body part with the display. The process 400 may include determining an applied pressure based on the applied force and the area.

[0053] At block 440, the process 400 includes capturing, using the camera of the mobile device, photoplethysmography (PPG) data of a set of blood vessels in the body part concurrently with the application of the force. The body part of the user may include a tip portion of a finger, and the set of blood vessels may include at least one blood vessel in a transverse palmar arch branch of a digital artery in the tip portion of the finger. The PPG data may relate to a volume of blood flow within the set of blood vessels that varies with an amplitude of the applied force. The PPG data may record intensity in a red channel of the camera. As the volume of blood flowing through the artery near the fingernail bed changes with the amplitude of the applied force, the reflective properties of the tissue in the region changes. As such, the camera of the mobile device can record changes in blood volume by measuring the changes in the reflective properties.

[0054] Merely by way of example, the display (or screen) of the mobile device may be set to a pure white background with a maximal brightness. The user may be instructed to place a body part (e.g., an index finger) over the camera (e.g., as aided by a visual signifier). The bright white screen illuminates the finger and the camera records pixel intensity changes in the red channel as a proxy for blood volume changes. To account for skin tone and lighting variations, the mobile device may perform a calibration with the user’s finger over the camera to adjust the ISO (the sensitivity of the camera's sensor to light). For example, the red channel intensity is between 70 and 200. This calibration may protect against under- or over- saturated PPG measurements.

[0055] At block 450, the process 400 includes determining the blood pressure of the user using an oscillometric technique based on the PPG data and the applied force. The PPG measurements may be aligned with the applied force measurements such that each PPG peak can be assigned an applied force value. The prominence of each PPG peak may be determined by finding the peak to trough difference. During the blood pressure measurement, the camera of the mobile device records the PPG signal while the user exerts force with a body part (e.g., an index finger) on the camera. Applying force causes changes in the PPG signal to create an oscillogram shape. When the applied force is greater than the diastolic blood pressure, the blood volume during the systolic pulse phase increases and the blood volume in the diastolic phase decreases. The total volume of blood flowing through the artery is approximately the same, but the pressure on the artery causes more blood to flow through at the systolic pulse phase and less at the diastolic pulse phase. This increases the prominence of the PPG signal. When the applied force surpasses the mean arterial pressure, the blood volume begins to be limited even during systolic phase of the pulse. As such, the systolic phase PPG prominence begins to decrease and the blood volume during the diastolic phase approaches zero to remains approximately constant. The PPG prominence continues to decrease even after the applied force surpasses the systolic blood pressure. As the pressure increases, the signal will decrease to zero (or noise) as the blood flow through the artery during both the systolic and diastolic phase approach zero. A plot of the pulse prominence through the range of applied forces (from less than diastolic blood pressure to greater than systolic blood pressure) for each PPG peak value creates a Gaussian or Skew-Gaussian like shape that may be referred to as the oscillogram. The blood pressure may be determined based on the oscillogram and a blood pressure model.

[0056] As described elsewhere in the present disclosure, the blood pressure model may be substantially independent of the type of the mobile device. Accordingly, it is unnecessary to perform a calibration with respect to the blood pressure model based on the type of the mobile device.

[0057] The blood pressure model may be a data driven model including parametric modeling, linear regression models, ensemble learning models (random forest, adaboost, etc.), a support vector machines model, a neural network (e.g., a transformer model, a CNN), or the like, or a variation thereof, or a combination thereof.

[0058] The oscillometric method utilized by most FDA-approved blood pressure cuffs utilize the oscillogram shape to estimate blood pressure. Accordingly, the blood pressure model may be trained to estimate blood pressure based on oscillographs. For training, an oscillograph may be processed to extract characteristic features including, e.g., a maximum value of the oscillogram, the applied force at the maximum value (an argument of the maximum value of the oscillogram), an extremum gradient (e.g., a maximum gradient, a minimum gradient) of the oscillogram, an argument of the extremum gradient (an argument of the maximum gradient, an argument of the minimum gradient), prominence at an extremum gradient (e.g., prominence at the maximum gradient, prominence at the minimum gradient), or the like, or a combination thereof. These exemplary characteristic features are visualized in Figure 6(a) and Figure 6(b).

[0059] Figure 6(a) illustrates exemplary PPG and applied force data measured by a mobile device and a corresponding oscillogram generated in accordance with some embodiments of the present document. Specifically, Figure 6(a) illustrates filtered PPG data and applied force data captured using a smartphone, in which the amplitude of the filtered PPG signal is computed and the force is converted to pressure via an fixed estimation of fingertip surface area. The filtered amplitude of the PPG signal plotted over the applied pressure is referred to as the oscillogram. The shape of the oscillogram resembles a skewed gaussian, where the peak indicates the mean arterial pressure.

[0060] Figure 6(b) illustrates average oscillograms corresponding to various blood pressure ranges in accordance with some embodiments of the present document. The plot of the average oscillograms by blood pressure ranges demonstrates that the oscillogram shape is indicative of blood pressure. The average oscillogram shape is plotted as a solid line. The average skew gaussian fit is plotted as a dotted line. For higher blood pressure, the oscillogram peak shifts to the right because, e.g., the mean arterial pressure is greater.

[0061] In an example case of training a blood pressure model, characteristic features of the oscillometric method including a maximum value of the oscillogram, the applied force at the maximum value (an argument of the maximum value of the oscillogram), an extremum gradient (e.g., a maximum gradient, a minimum gradient) of the oscillogram, an argument of the extremum gradient (an argument of the maximum gradient, an argument of the minimum gradient), prominence at an extremum gradient (e.g., prominence at the maximum gradient, prominence at the minimum gradient) of each of training oscillograms are the only inputs into a blood pressure model. In the example case, attributes related to exercise including heart rate and pulse shape were not included in the features provided to the model; indirect measurements of heart rate or pulse were avoided by interpolating all oscillograms to have the same number of points with the same force values. A least absolute shrinkage and selection operator (LASSO) regression model was trained to predict systolic and diastolic measures of blood pressure. The model was trained and tested using a hold one participant out validation. In this training scheme, the model was trained on all data excluding the one or more blood pressure measurements from a single participant. The model was then used predict the blood pressure of the holdout participant. This process was repeated for each participant.

[0062] N=30 validation study was performed based on the trained blood pressure model. Table 2 shows participant demographics.

Table 2 [0063] In Table 2 of the participant demographics, “Skin” refers to Fitzpatrick Skin Type, “Phone” refers to the average smartphone estimation (rounded to nearest integer), and “Cuff” refers to the cuff measurement used for ground truth. The averaged phone and cuff measurements displayed are from natural (non-exercise induced) measurements only. Among the thirty participants, each participated in multiple measurements.

[0064] In the blood pressure validation study, the smartphone device was compared against an FDA-approved cuff-based device (Omron BP7350). The participant first recorded a trial measurement using the smartphone to learn how to use the device and practice applying force. The trial measurement was not used for analysis. The cuff-based device was used to measure the participants blood pressure following standard protocols. Immediately afterwards, the participant performed three blood pressure measurements using the smartphone. After completing three blood pressure measurements on the smartphone, the cuff-based device was used to measure the blood pressure again. The full measurement sequence was as follows: cuff, smartphone, smartphone, smartphone, cuff. If the two cuff systolic BP readings had a difference less than three mmHg, the first cuff measurement was used as the label for all three smartphone measurements in the analysis. If the two blood pressure readings differed by more than three mmHg, the two readings were averaged.

[0065] The participants were also asked to optionally complete a second phase of the data collection involving exercise. The purpose of this exercise phase was to obtain high blood pressure data. Consenting participants were asked to perform a wall sit for approximately one minute. During the wall sit, the participants simultaneously measured their blood pressure with the blood pressure cuff device and the smartphone device. Both blood pressure measurements were initiated approximately ten seconds after the start of the wall sit.

[0066] Among the thirty participants, N=six participants were entirely excluded. Of these six excluded participants, one participant was unable to perform any valid force measurements; one participant had low prominence in all measurements; and one participant had poor gaussian skew fit in all measurements. The other three excluded participants had a mixture of criteria excluding each measurement. Participant measurements are excluded based on the following exclusion criteria: • Cuff blood pressure exceeds 160mmHg: If the reference measurement exceeded 160mmHg, the measurement is excluded based on prior re- search demonstrating that the fingertip and arm blood pressure values differ at extremely high BP values induced from exercise.

• Inconsistent/Undetected Pulse or Excessive Noise: The Fast Fourier Transform (FFT) of the PPG signal between zero and two Hz (frequencies relevant for human pulse) must have a narrow, identifiable peak indicating the pulse rate.

• Insufficient Applied Force: The applied force measurement must have a correlation coefficient greater than 0.88. The first 0.5 seconds of the force signal must have a correlation coefficient less than 0.21 N. Also, the minimum applied force must be less than six N and the maximum applied force value must be greater than seven N.

• Poor Skew Gaussian Fit: The mean absolute error for the normalized Skew Gaussian of best fit must be less than 0.18.

• Saturation in PPG signal: If the PPG signal rises above 254 pixel intensity units, the measurement is discarded. This is indicative of the finger being lifted off of the smartphone camera during a measurement.

• Unchanging in PPG Prominence: The standard deviation of the normalized PPG prominence must be greater than 0.17.

[0067] From the validity study, the mobile device based oscillometric blood pressure measurement achieved an MAE of 9.9 mmHg and 8.9 mmHg of systolic and diastolic BP, respectively, compared to an FDA approved blood pressure cuff. The Pearson correlation coefficients are 0.67 and 0.23 for systolic and diastolic measurements, respectively. The blood pressure cuff measurements ranged 62 to 107 mmHg for diastolic and 89 to 147 mmHg systolic. See Figure 6(a) and Figure 6(b) showing regression and Bland Altman plots for systolic blood pressure (Systolic BP) and diastolic blood pressure (Diastolic BP) for all measurements of N=24 included participants, respectively. There were 73 total measurements because each participant had multiple measurements. The Systolic and Diastolic BP values were computed with the same model.

[0068] Around a use of screening based on the technology disclosed herein, blood pressure (BP) values may be categorized based on the AMA/ACC standards. Figure 7 illustrates a confusion matrix based on results shown in Figures 6(a) and 6(b) and adjusted towards ensuring correctly labelling elevated and high blood pressure individuals. Specifically, for the categorical prediction, the sensitivity and specificity were adjusted from the original predictions. The categorical prediction thresholds were determined by systolic BP values only based on standards developed by AMA/ACC. The receiver operating characteristic (ROC) analysis reveals an area under the curve (AUC) of 0.88 for detecting high blood pressure individuals, characterized by systolic blood pressure greater than 130 mmHg. The ROC curve informs on the sensitivity and specificity of the prediction. The confusion matrix as illustrated in Figure 7 demonstrates effective blood pressure screening using downloadable applications with no required attachments or hardware according to embodiments of the present document. With a false positive rate of three percent and a true positive rate of 88% for identifying hypertensive or elevated blood pressures from normal blood pressures, the present technology may be used as a hypertension screening tool that can be easily downloaded at home and used on one’s own mobile device.

[0069] These results suggest the potential for blood pressure screening using only a downloadable smartphone software application. The technology may be used as a low- barrier hypertension screening measurement fit for widespread adoption.

[0070] In addition, the mobile device based oscillometric blood pressure technology as disclosed herein may provide a calibration mechanism. For example, methods to monitor relative changes in blood pressure using wearable devices like smartwatches or smart glasses are becoming increasingly common; these devices generally rely on pulse transit time (PTT), Pulse Wave Analysis (PWA), or Pulse Arrival Time (PAT) methodologies. A key limitation for this type of blood pressure measurement is the required calibration using an absolute blood pressure measurement, often performed daily or weekly via a blood pressure cuff device. The absolute blood pressure measurement on a mobile device, as disclosed here may provide an effective, low-cost method for calibrating wearable or other continuous BP relative measurements.

[0071] Moreover, the mobile device based oscillometric blood pressure technology as disclosed herein may be combined with other cardiovascular measurements for an analysis of heart health. For example, the system disclosed herein may be combined with another optical system using a camera to perform PWA, heart rate variability (HRV) measurements, or pulse measurements to provide additional information to blood pressure.

[0072] Technical improvements offered in the embodiments disclosed herein include the methods for performing absolute blood pressure measurements that can be applied across different types (e.g., models) of mobile devices without hardware modifications or attachments and instead can be achieved by downloadable applications. Generalizing sensor-based measurements across the large variation of mobile devices is an established issue that significantly limits the impact of work in the mobile health space. The cross-device compatibility study result is evidence that the force damping technique disclosed herein allows the blood pressure measurement to be functional on a range of mobile devices. The data processing and modeling demonstrate that many mobile devices of different types are capable to perform an oscillometric blood pressure measurement after a mobile device type based calibration with respect to force measurements. With tens of millions of users for mobile devices of each one of popular types, a single factory calibration may enable millions of mobile devices to perform oscillometric blood pressure measurements.

EXAMPLES

[0073] The following examples are illustrative of several embodiments in accordance with the present technology. Other exemplary embodiments of the present technology may be presented prior to the following listed examples, or after the following listed examples.

[0074] 1 . A mobile device, including: a vibration motor, an inertial motion unit (IMU), a camera, a processor, and non-transitory computer-readable memory that stores instructions that, when executed by the processor, cause the processor to perform at least: cause the vibration motor to vibrate; acquire, by the inertial measurement unit of the mobile device, IMU data indicative of vibration damping that is caused by a user's body part applying a force on the camera; determine the applied force based on a force model and the IMU data; capture, by the camera of the mobile device, photoplethysmography (PPG) data of a set of blood vessels in the body part concurrently with the application of the force; and determine a blood pressure of the user using an oscillometric technique based on the PPG data and the applied force.

[0075] 2. The mobile device of any one or more of the solutions described herein, in which the force model includes a machine learning algorithm trained to correlate (1 ) the IMU data measured by the mobile device that is indicative of the vibration damping of the mobile device with (2) the force that is applied to the mobile device and cause the vibration damping.

[0076] 3. The mobile device of any one or more of the solutions described herein, in which the IMU includes an accelerometer and a gyroscope, and the IMU data includes multi-axis accelerometer data and multi-axis gyroscope data.

[0077] 4. A method for measuring blood pressure using a mobile device, the method including: causing a vibration motor of the mobile device to vibrate; acquiring, using an inertial measurement unit (IMU) of the mobile device, IMU data indicative of vibration damping that is caused by a user’s body part applying a force on a camera of the mobile device; determining the applied force based on a force model and the IMU data; capturing, using the camera of the mobile device, photoplethysmography (PPG) data of a set of blood vessels in the body part concurrently with the application of the force; and determining the blood pressure of the user using an oscillometric technique based on the PPG data and the applied force.

[0078] 5. The method of any one or more of the solutions described herein, in which the IMU data includes multi-axis accelerometer data and multi-axis gyroscope data.

[0079] 6. The method of any one or more of the solutions described herein, in which determining the applied force from the IMU data includes processing the IMU data based on at least one of filtering, a filter bank, averaging, standard deviation, wavelet decomposition, spectral analysis, or empirical mode decomposition (EMD).

[0080] 7. The method of any one or more of the solutions described herein, in which: each axis of the multi-axis accelerometer data and the multi-axis gyroscope data includes a signal, determining the applied force from the IMU data includes for each axis of the multi-axis accelerometer data and the multi-axis gyroscope data, decomposing a signal of the axis into a plurality of intrinsic mode functions (IMFs) representing different frequency components of the signal, and the applied force is determined based on at least a portion of the IMFs that correspond to various axes of the multi-axis accelerometer data and the multi-axis gyroscope data.

[0081] 8. The method of any one or more of the solutions described herein, in which determining the applied force from the IMU data further includes: for each axis of the multi-axis accelerometer data and the multi-axis gyroscope data, identifying, from the plurality of IMFs of the signal of the axis, a first IMF that corresponds to a highest frequency; and identifying an upper envelope of the first IMF of the axis; and inputting into the force model the upper envelopes of the first IMFs that respectively correspond to all axes of the multi-axis accelerometer data and the multi-axis gyroscope data.

[0082] 9. The method of any one or more of the solutions described herein, in which the PPG data relates to a volume of blood flow within the set of blood vessels that varies with an amplitude of the applied force.

[0083] 10. The method of any one or more of the solutions described herein, in which the PPG data records intensity in a red channel of the camera.

[0084] 11 . The method of any one or more of the solutions described herein, in which the mobile device is placed on a supporting surface when the user applies the force, the method further includes: performing a surface calibration by characterizing a material of the supporting surface; and determining the applied force based further on the surface calibration.

[0085] 12. The method of any one or more of the solutions described herein, further including: providing a visual guide to be presented on a display of the mobile device, in which the visual guide includes a force plot that illustrates the applied force in substantially real time.

[0086] 13. The method of any one or more of the solutions described herein, in which the visual guide further includes a force guide line overlaid with the applied force.

[0087] 14. The method of any one or more of the solutions described herein, further including: providing a visual guide to be presented on a display of the mobile device, in which the visual guide includes indicia to guide the user in placing a surface of the body part onto the camera.

[0088] 15. The method of any one or more of the solutions described herein, further including: measuring an area of a surface of the body part, in which the force is applied by pressing the surface of the body part against the camera; and determining an applied pressure based on the applied force and the area, and determining the blood pressure includes: generating an oscillogram based on the PPG data and the applied pressure; and deriving the blood pressure from the oscillogram.

[0089] 16. The method of any one or more of the solutions described herein, further including: causing the oscillogram to be presented on a display of the mobile device.

[0090] 17. The method of any one or more of the solutions described herein, in which: the body part of the user is a tip portion of a finger, and the set of blood vessels includes at least one blood vessel in a transverse palmar arch branch of a digital artery in the tip portion of the finger.

[0091] 18. The method of any one or more of the solutions described herein, in which: determining a blood pressure of the user using the oscillometric technique including: generating an oscillogram based on the PPG data and the applied force; extracting, from the oscillogram, characteristic features including at least one of a peak of the oscillogram, an applied force corresponding to the peak, and an extremum value of a gradient of the oscillogram, prominence at an extremum gradient of the oscillogram, or an argument of an extremum gradient, and generating the blood pressure based on the characteristic features and a blood pressure model.

[0092] 19. The method of any one or more of the solutions described herein, further including: obtaining, from a model library, the force model based on a type of the mobile device, in which the type of the mobile device is associated with at least one of a configuration of the IMU, a configuration of the vibration motor, or a configuration of the camera.

[0093] 20. One or more non-transitory computer-readable media storing instructions that, when executed by one or more processors of a mobile device, cause the mobile device to perform any one or more of the solutions described herein.

[0094] Implementations of the subject matter and the functional operations described in this patent document can be implemented in various systems, digital electronic circuitry, or in computer software, firmware, or hardware, including the structures disclosed in this specification and their structural equivalents, or in combinations of one or more of them. Implementations of the subject matter described in this specification can be implemented as one or more computer program products, i.e. , one or more modules of computer program instructions encoded on a tangible and non-transitory computer readable medium for execution by, or to control the operation of, data processing apparatus. The computer readable medium can be a machine-readable storage device, a machine-readable storage substrate, a memory device, a composition of matter effecting a machine-readable propagated signal, or a combination of one or more of them. The term “data processing unit” or “data processing apparatus” encompasses all apparatus, devices, and machines for processing data, including by way of example a programmable processor, a computer, or multiple processors or computers. The apparatus can include, in addition to hardware, code that creates an execution environment for the computer program in question, e g., code that constitutes processor firmware, a protocol stack, a database management system, an operating system, or a combination of one or more of them.

[0095] A computer program (also known as a program, software, software application, script, or code) can be written in any form of programming language, including compiled or interpreted languages, and it can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment. A computer program does not necessarily correspond to a file in a file system. A program can be stored in a portion of a file that holds other programs or data (e.g., one or more scripts stored in a markup language document), in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files that store one or more modules, sub programs, or portions of code). A computer program can be deployed to be executed on one computer or on multiple computers that are located at one site or distributed across multiple sites and interconnected by a communication network.

[0096] The processes and logic flows described in this specification can be performed by one or more programmable processors executing one or more computer programs to perform functions by operating on input data and generating output. The processes and logic flows can also be performed by, and apparatus can also be implemented as, special purpose logic circuitry, e.g., an FPGA (field programmable gate array) or an ASIC (application specific integrated circuit).

[0097] Processors suitable for the execution of a computer program include, by way of example, both general and special purpose microprocessors, and any one or more processors of any kind of digital computer. Generally, a processor will receive instructions and data from a read only memory or a random access memory or both. The essential elements of a computer are a processor for performing instructions and one or more memory devices for storing instructions and data. Generally, a computer will also include, or be operatively coupled to receive data from or transfer data to, or both, one or more mass storage devices for storing data, e.g., magnetic, magneto optical disks, or optical disks. However, a computer need not have such devices. Computer readable media suitable for storing computer program instructions and data include all forms of nonvolatile memory, media and memory devices, including by way of example semiconductor memory devices, e.g., EPROM, EEPROM, and flash memory devices. The processor and the memory can be supplemented by, or incorporated in, special purpose logic circuitry.

[0098] It is intended that the specification, together with the drawings, be considered exemplary only, where exemplary means an example. As used herein, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. Additionally, the use of “or” is intended to include “and/or,” unless the context clearly indicates otherwise.

[0099] While this patent document contains many specifics, these should not be construed as limitations on the scope of any invention or of what may be claimed, but rather as descriptions of features that may be specific to particular embodiments of particular inventions. Certain features that are described in this patent document in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination. Moreover, although features may be described above as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination can in some cases be excised from the combination, and the claimed combination may be directed to a subcombination or variation of a subcombination. [0100] Similarly, 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. Moreover, the separation of various system components in the embodiments described in this patent document should not be understood as requiring such separation in all embodiments.

[0101] Various embodiments described herein are described in the general context of methods or processes, which may be implemented in one embodiment by a computer program product, embodied in a computer-readable medium, including computerexecutable 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), Blu-ray Discs, etc. Therefore, the computer-readable media described in the present application include 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.

[0102] For example, one aspect of the disclosed embodiments relates to a computer program product that is embodied on a non-transitory computer readable medium. The computer program product includes program code for carrying out any one or and/or all of the operations of the disclosed embodiments.

[0103] 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

CLAIMS What is claimed is:

1 . A method for measuring blood pressure using a mobile device, the method comprising: causing a vibration motor of the mobile device to vibrate; acquiring, using an inertial measurement unit (IMU) of the mobile device, IMU data indicative of vibration damping that is caused by a user’s body part applying a force on a camera of the mobile device; determining the applied force based on a force model and the IMU data; capturing, using the camera of the mobile device, photoplethysmography (PPG) data of a set of blood vessels in the body part concurrently with the application of the force; and determining a blood pressure of the user using an oscillometric technique based on the PPG data and the applied force.

2. The method of claim 1 , wherein the IMU data comprises multi-axis accelerometer data and multi-axis gyroscope data.

3. The method of claim 2, wherein determining the applied force from the IMU data comprises processing the IMU data based on at least one of filtering, a filter bank, averaging, standard deviation, wavelet decomposition, spectral analysis, or empirical mode decomposition (EMD).

4. The method of claim 3, wherein: each axis of the multi-axis accelerometer data and the multi-axis gyroscope data comprises a signal, determining the applied force from the IMU data comprises for each axis of the multi-axis accelerometer data and the multi-axis gyroscope data, decomposing a signal of the axis into a plurality of intrinsic mode functions (IMFs) representing different frequency components of the signal, and the applied force is determined based on at least a portion of the IMFs that correspond to various axes of the multi-axis accelerometer data and the multi-axis gyroscope data.

5. The method of claim 4, wherein determining the applied force from the IMU data further comprises: for each axis of the multi-axis accelerometer data and the multi-axis gyroscope data, identifying, from the plurality of IMFs of the signal of the axis, a first IMF that corresponds to a highest frequency; and identifying an upper envelope of the first IMF of the axis; and inputting into the force model the upper envelopes of the first IMFs that respectively correspond to all axes of the multi-axis accelerometer data and the multi-axis gyroscope data.

6. The method of claim 1 , wherein the PPG data relates to a volume of blood flow within the set of blood vessels that varies with an amplitude of the applied force.

7. The method of claim 1 , wherein the PPG data records intensity in a red channel of the camera.

8. The method of claim 1 , wherein the mobile device is placed on a supporting surface when the user applies the force, the method further comprising: performing a surface calibration by characterizing a material of the supporting surface; and determining the applied force based further on the surface calibration.

9. The method of claim 1 , further comprising: providing a visual guide to be presented on a display of the mobile device, wherein the visual guide comprises a force plot that illustrates the applied force in substantially real time.

10. The method of claim 9, wherein the visual guide further comprises a force guide line overlaid with the applied force.

11 . The method of claim 1 , further comprising: providing a visual guide to be presented on a display of the mobile device, wherein the visual guide comprises indicia to guide the user in placing a surface of the body part onto the camera.

12. The method of claim 1 , further comprising: measuring an area of a surface of the body part, wherein the force is applied by pressing the surface of the body part against the camera; and determining an applied pressure based on the applied force and the area, wherein determining the blood pressure comprises: generating an oscillogram based on the PPG data and the applied pressure; and deriving the blood pressure from the oscillogram.

13. The method of claim 12, further comprising: causing the oscillogram to be presented on a display of the mobile device.

14. The method of claim 1 , wherein determining a blood pressure of the user using the oscillometric technique comprising: generating an oscillogram based on the PPG data and the applied force; extracting, from the oscillogram, characteristic features including at least one of a peak of the oscillogram, an applied force corresponding to the peak, and an extremum value of a gradient of the oscillogram, prominence at an extremum gradient of the oscillogram, or an argument of an extremum gradient, and generating the blood pressure based on the characteristic features and a blood pressure model.

15. The method of claim 1 , wherein: the body part of the user is a tip portion of a finger, and the set of blood vessels includes at least one blood vessel in a transverse palmar arch branch of a digital artery in the tip portion of the finger.

16. The method of claim 1 , further comprising: obtaining, from a model library, the force model based on a type of the mobile device, wherein the type of the mobile device is associated with at least one of a configuration of the IMU, a configuration of the vibration motor, or a configuration of the camera.

17. A mobile device, comprising: a vibration motor, an inertial motion unit (IMU), a camera, a processor, and non-transitory computer-readable memory that stores instructions that, when executed by the processor, cause the processor to at least: cause the vibration motor to vibrate; acquire, by the IMU of the mobile device, IMU data indicative of vibration damping that is caused by a user’s body part applying a force on the camera; determine the applied force based on a force model and the IMU data; capture, by the camera of the mobile device, photoplethysmography (PPG) data of a set of blood vessels in the body part concurrently with the application of the force; and determine a blood pressure of the user by an oscillometric technique based on the PPG data and the applied force.

18. The mobile device of claim 17, wherein the force model comprises a machine learning algorithm trained to correlate (1 ) the IMU data measured by the mobile device that is indicative of the vibration damping of the mobile device with (2) the force that is applied to the mobile device and cause the vibration damping.

19. The mobile device of claim 17, wherein the IMU comprises an accelerometer and a gyroscope, and the IMU data comprises multi-axis accelerometer data and multi-axis gyroscope data.

20. One or more non-transitory computer-readable media storing instructions that, when executed by one or more processors of a mobile device, cause the mobile device to perform operations including: causing a vibration motor of the mobile device to vibrate; acquiring, using an inertial measurement unit (IMU) of the mobile device, IMU data indicative of vibration damping that is caused by a user’s body part applying a force on a camera of the mobile device; determining the applied force based on a force model and the IMU data; capturing, using the camera of the mobile device, photoplethysmography (PPG) data of a set of blood vessels in the body part concurrently with the application of the force; and determining a blood pressure of the user using an oscillometric technique based on the PPG data and the applied force.

21 . The one or more non-transitory computer-readable media of claim 20, wherein the operations further include: obtaining, from a model library, the force model based on a type of the mobile device, wherein the type of the mobile device is associated with at least one of a configuration of the IMU, a configuration of the vibration motor, or a configuration of the camera.