WO2025000038A1

WO2025000038A1 - Blood pressure monitoring device, system and method

Info

Publication number: WO2025000038A1
Application number: PCT/AU2024/050689
Authority: WO
Inventors: Philip Mehrgardt; Anusha WITHANA; Simon Poon
Original assignee: The University Of Sydney
Priority date: 2023-06-28
Filing date: 2024-06-28
Publication date: 2025-01-02

Abstract

A wearable blood pressure measuring or monitoring device (100) comprises a housing (110) for receiving a limb (120), a first actuator (130) for applying a load force at a first location (150) of the limb to restrict blood flow (140) and a one or more first sensors (160) for detecting biometric parameter(s) at a second location (190) of the limb. The one or more first sensors (160) detects the biometric parameter(s) when the first actuator (130) applies the load force to determine the blood pressure of the patient. A one or more second sensors (270) detects biometric parameter(s) at the first location (150). An optional second actuator (520) may apply a load force at the second location, so that biometric parameter(s) can be measured at both locations when blood flow is restricted and unrestricted. An associated method and system use a neural network to calculate the blood pressure from the measured data.

Description

Blood Pressure Monitoring Device, System and Method

Technical Field

[001] The invention relates to a blood pressure monitoring device, system and method and in a particular to a portable device, system and method for blood pressure monitoring and/or measuring. The invention has been developed primarily for use as a continuous blood pressure monitoring device and will be described hereinafter by reference to this application.

Background

[0001] The following discussion of the prior art is intended to present the invention in an appropriate technical context and allow its advantages to be properly appreciated. Unless clearly indicated to the contrary, however, reference to any prior art in this specification should not be construed as an express or implied admission that such art is widely known or forms part of common general knowledge in the field.

[0002] Photoplethysmography (PPG) sensors emit light that is absorbed in different amounts by different tissues in the body, such as skin pigments, tissue, bones and blood. The emitted light is absorbed differently by oxygenated (usually arterial) and deoxygenated (usually venous) blood. Also, the amount of light absorbed depends on the amount of blood at the location of the PPG sensor, which varies with heart or pulse rate.

[0003] Due to these characteristics, devices incorporating PPG sensors are used in non-invasive applications to monitor and/or measure various biometric data (such as oxygen saturation levels, heart rate and blood pressure) of patients. These devices are primarily used in clinical or controlled environments, such as in hospitals and medical clinics. However, these devices are limited in application to these environments, due to the need for the patient to be relatively stationary or still during the monitoring/measuring process to provide sufficient accuracy. Hence, patients are unable to perform any significant physical activity when being monitored by these devices. Consequently, patients have to spend time visiting the hospital or medical clinic to have their blood pressure checked, leading to inconvenience. For patients with chronic conditions requiring continuous monitoring their blood pressure, this necessitates extended periods in hospital, increasing their discomfort, increasing the burden on hospitals or medical clinics. Furthermore, the measurement of a patient’s blood pressure in a clinical environment may not be a realistic measure as it may not reflect the patient’s blood pressure in day to day activities.

[0004] Outside clinical environments, wearable devices (such as smartwatches or fitness/activity trackers) also use PPG sensors to monitor and measure the same or similar biometric data, like heart rate. This enables the user to perform various physical activities while still enabling measurement of their biometric data. However, the accuracy of these wearable devices is less than the accuracy of the devices used in clinical environments. This is primarily due to a greater amount of noise and a wider variety of noise sources in non-clinical environments, resulting in lower signal to noise ratios in wearable devices. In particular, most wearable devices are attached to the arm of a user, frequently at the wrist. As such, frequent movement of the arm during activities like walking or running creates significant noise in the form of motion artefacts that need to be filtered out to obtain an accurate measurement.

[0005] In addition, noise can be created by improper or variable contact of the PPG sensors to the skin of the user. For example, slight shifts or movement of the wearable device on the wrist may generate noise from the ingress of ambient light between the PPG sensor and the skin (thus contaminating the signal received by the PPG sensor). It can also cause “signal crossover” where the PPG sensor is periodically moved resulting in the device monitoring a motion artefact or other noise signal instead of the actual signal. Since PPG sensors measure reflected or transmitted light from the skin, improper contact can cause large signal errors. For example, contact pressure on the skin can causes deformations of the dermis and its connective tissue components, resulting in significant PPG signal noise.

[0006] While wearable devices have attempted to address these errors to improve the signal to noise ratio, especially those arising from motion artefacts, further improvements are required to improve the accuracy of these devices. Moreover, there is not any ability in current prior art devices to rectify or self-correct any such errors in measurement, such as those caused by signal crossover.

[0007] It is an object of the present invention to overcome or substantially ameliorate one or more of the disadvantages of prior art, or at least to provide a useful alternative. It is an object of the invention in at least one preferred form to provide a portable device and method that enables continuous monitoring of a person’s blood pressure without restricting their movement while providing a similar accuracy to clinical devices.

Summary

[008] A first aspect of the invention provides a device for measuring the blood pressure in a patient, comprising: a housing for receiving a limb of the patient; a first actuator for applying a load force to the limb to restrict a blood flow in the limb; one or more first sensors located adjacent an inner wall of the housing for detecting one or more biometric parameters at a second location of the limb; and wherein the one or more first sensors detect the one or more biometric parameters when the first actuator applies the load force to determine the blood pressure of the patient.

[009] In some embodiments, the device comprises a one or more second sensors adjacent the inner wall for detecting the one or more biometric parameters at the first location of the limb. In some embodiments, the one or more second sensors detect the one or more biometric parameters at the first location when the first actuator does not apply the load force at the first location.

[0010] In some embodiments, the housing comprises a base portion and the inner wall is an inner wall of the base portion. In other embodiments, the housing comprises a first moveable portion, wherein the first actuator is configured to move the first moveable portion relative to the base portion. In further embodiments, the first moveable portion comprises a clamping portion for engaging the limb. [0011] In some embodiments, a first force sensor is operatively connected to the first actuator to measure for measuring the load force applied by the first actuator. In other embodiments, the first actuator comprises a servomechanism or a linear actuator. In further embodiments, the first force sensor comprises a load cell.

[0012] In some embodiments, the first actuator applies the load force at a first location of the limb and the one or more first sensors detect the one or more biometric parameters at the first location. In other embodiments, the first actuator applies the load force at a first location of the limb and the one or more first sensors detect the one or more biometric parameters at a second location of the limb.

[0013] In some embodiments, the device comprises: a second actuator for applying a load force to the limb at the second location; and wherein the one or more second sensors detect the one or more biometric parameters when the second actuator applies the load force at the first location.

[0014] In some embodiments, the device comprises a second moveable portion, wherein the second actuator is configured to move the second moveable portion relative to the base portion.

[0015] In some embodiments, the first and second actuators alternate in applying their respective load forces at the first and second locations, respectively. In other embodiments, the first and second actuators simultaneously apply their respective load forces at the first and second locations, respectively.

[0016] In some embodiments, the detected biometric parameters from the one or more first sensors are used to calculate the blood pressure of the patient.

[0017] In some embodiments, a second force sensor is operatively connected to the second actuator to measure the load force applied by the second actuator. In other embodiments, the second actuator is a linear actuator. In further embodiments, the second force sensor comprises a load cell. [0018] In some embodiments, the one or more first sensors and/or one or more second sensors each comprise a photoplethysmography (PPG) sensor and/or an inertial measurement unit (IMU). In other embodiments, the PPG sensor comprises an infrared, red and/or green light transmitter and a light receiver. In some embodiments, the PPG sensor may comprise a light emitting diode (LED), photodiodes or a combination thereof. Alternatively, the PPG sensor in other embodiments may comprise a microlens, miniature optical element, miniature camera or a combination thereof. In further embodiments the IMU comprises a 6-axis or 9-axis IMU.

[0019] In some embodiments, there is a plurality of first sensors and/or second sensors. In some embodiments, the plurality of first sensors and/or second sensors are arranged to measure the spatial propagation of blood flow in the limb. In some embodiments, the first sensors and/or second sensors are spaced laterally apart relative to an axis of the limb. In some embodiments, the axis is a longitudinal axis of the limb. In some embodiments, the first sensors and/or second sensors are each aligned along a transverse axis to the longitudinal axis of the limb.

[0020] In some embodiments, the first sensors and/or second sensors are arranged in groups, wherein the groups are spatially separated from each other. In some embodiments, the first sensors and/or second sensors are spaced laterally apart relative to the limb in their respective groups. In some embodiments, the groups of first sensors and/or second sensors are each aligned along an axis. In some embodiments, the groups of first sensors and/or second sensors are each aligned along a transverse axis to a longitudinal axis of the limb.

[0021] In some embodiments, there is a plurality of second actuators, wherein the second actuators are configured to move the second moveable portion. In other embodiments, there is a plurality of first actuators, wherein the first actuators are configured to move the first moveable portion. In further embodiments, there is a plurality of actuators for respectively applying a load force to a plurality of locations of the limb.

[0022] In some embodiments, the limb comprises an appendage. In other embodiments, the appendage comprises a finger or thumb. In further embodiments, the first location is a proximal phalanx of a finger and the second location is a distal phalanx of the finger. In yet other embodiments, the first location is a distal phalanx of a finger and the second location is a proximal phalanx of the finger. In another embodiment, the appendage comprises a toe.

[0023] In some embodiments, the biometric parameters comprise to one or more characteristics of the blood flow in the limb.

[0024] A second aspect of the invention provides a method of measuring the blood pressure of a patient, comprising: applying a load force at a first location to restrict the blood flow of a limb of the patient; in response to the load force applying step, taking a measurement at a second location of the limb of one or more biometric parameters to create a first set of data; and calculating the blood pressure of the patient from the first set of data.

[0025] In some embodiments, the method comprises: taking a measurement at the first location of one or more biometric parameters to create a second set of data; and calculating the blood pressure of the patient from the first and second sets of data.

[0026] In some embodiments, the measurement at the first location is taken when the load force is not applied to the first location. In other embodiments, the measurement at the first location is taken prior to the load force applying step. In further embodiments, the measurement at the first location is taken after the load force applying step.

[0027] In some embodiments, the method comprises calibrating the second set of data with the first set of data.

[0028] In some embodiments, the method comprises: applying a load force at the second location to restrict the blood flow in the limb; in response to the load force applying step at the second location, taking a measurement at the first location of the limb of one or more characteristics of the blood flow to create a third set of data; and calculating the blood pressure of the patient from the first and third sets of data.

[0029] In some embodiments, the method comprises calibrating the second set of data with the first set of data or third set of data. In other embodiments, the method comprises calibrating the second set of data with the first and third sets of data.

[0030] In some embodiments, a neural network (NN) performs the calibrating step. In other embodiments, the NN performs the calculating step.

[0031] In some embodiments, the NN is trained with a preliminary set of data prior to performing the calculating step. In other embodiments, the NN is trained with a preliminary set of data prior to taking the measurement at the first location.

[0032] In some embodiments, the NN compares the calculated blood pressure against the second set of data in real time and determines a confidence value in the accuracy of the calculated blood pressure. In other embodiments, in response to the confidence level falling to or below a threshold value, the NN adjusts the calculated blood pressure based on the second set of data. In further embodiments, in response to the confidence level falling to or below a threshold value, the NN initiates the load force applying step and repeats the measurement at the second location to obtain a fourth set of data and repeats the calculating step using the fourth set of data.

[0033] In some embodiments, the measurement at the first location is performed continuously to create the first set of data.

[0034] In some embodiments, the load force is applied periodically.

[0035] In some embodiments, the measurement at the first location also comprises measuring the load force applied at the first location to create a first set of force data. In other embodiments, the measurement at the second location also comprises measuring the load force applied at the second location to create a second set of force data. In further embodiments, the first and second sets of force data are combined to produce a combined set of force data.

[0036] In some embodiments, the calculating step comprises applying a transformation of the set of force data (which can be the first set of force data, the second set of force data or the combined set of force data) and combining the transformed force data with the second set of data. In further embodiments, the transformation comprises selecting an AC component of the set of force data and segregating the AC component into discrete data clusters. In yet other embodiments, where a discrete data cluster lacks data, data values are interpolated and allocated to the empty discrete data cluster.

[0037] In some embodiments, the second aspect has the same features as embodiments of the first aspect of the invention stated above, where applicable.

[0038] A third aspect of the invention provides a system for determining the blood pressure of a patient, comprising: the blood pressure monitoring device of the first aspect; a neural network module for receiving the first set of data, the neutral network being configured to calculate the blood pressure of the patient from the first set of data.

[0039] A fourth aspect of the invention provides a system for continuously determining the blood pressure of a patient, comprising: the blood pressure monitoring device of the first aspect; and a neural network module for receiving the first set of data, the neutral network being configured to perform the method of the second aspect.

[0040] In some embodiments, the system comprises a display for displaying the calculated blood pressure. In other embodiments, the display may also display the one or more biometric data or characteristics of the blood flow, or a biometric parameter corresponding to or derived from the characteristics of the blood flow. [0041] In some embodiments, the third and fourth aspects have the same features as embodiments of the first and second aspects of the invention stated above, where applicable.

[0042] Unless the context clearly requires otherwise, throughout the description and the claims, the words “comprise”, “comprising”, and the like are to be construed in an inclusive sense as opposed to an exclusive or exhaustive sense; that is to say, in the sense of “including, but not limited to”.

[0043] Furthermore, as used herein and unless otherwise specified, the use of the ordinal adjectives "first", "second", "third", etc., to describe a common object, merely indicate that different instances of like objects are being referred to, and are not intended to imply that the objects so described must be in a given sequence, either temporally, spatially, in ranking, or in any other manner.

Brief Description of the Drawings

[0044] Preferred embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings in which:

[0045] Figure 1 is a schematic cross-sectional view of a blood pressure monitoring device according to an embodiment of the invention;

[0046] Figure 2 is a perspective view of the device of Figure 1 ;

[0047] Figure 3 is a partially exploded bottom perspective view of a blood pressure monitoring device according to another embodiment of the invention;

[0048] Figure 4 is a partially exploded top perspective view of the device of Figure 3;

[0049] Figure 5 is a partially exploded end view of the device of Figure 3;

[0050] Figure 6 is another partially exploded end view of the device of Figure 3;

[0051] Figure 7 is a partially exploded bottom view of the device of Figure 3; [0052] Figure 8 is a partially exploded top view of the device of Figure 3;

[0053] Figure 9A and 9B are partially exploded front views of the device of Figure 3;

[0054] Figure 10A and 10B are partially exploded rear views of the device of Figure 3;

[0055] Figures 11 and 12 are schematic cross-sectional views of the device of Figure 3 in operation;

[0056] Figure 13 is a schematic cross-sectional view of a blood pressure monitoring device according to a further embodiment of the invention;

[0057] Figure 14 is a schematic drawing illustrating the location of sensors used in the device of Figure 13 relative to a finger;

[0058] Figure 15 is a schematic diagram of a system for monitoring blood pressure according to a further embodiment of the invention;

[0059] Figures 16(a), 16(b), 16(c) and 17 are graphs illustrating measurements taken as part of a method for monitoring blood pressure according to yet another embodiment of the invention;

[0060] Figures 18(a) and 18(b) are graphs illustrating measurements of the applied load force art in the method; and

[0061] Figures 19(a) and 19(b) are graphs illustrating transformation of the measurements in Figure 15.

Description of Embodiments

[0062] The present invention will now be described with reference to the following examples which should be considered in all respects as illustrative and non-restrictive. In the Figures, corresponding features within the same embodiment or common to different embodiments have been given the same reference numerals. [0063] Referring to Figures 1 and 2, a device 100 for measuring the blood pressure in a patient according to one embodiment of the invention comprises a housing 110 for receiving a limb of the patient (in the form of a finger 120) and an actuator in the form of a servomechanism 130 for applying a load force to the finger to restrict a blood flow 140 in the finger. In this embodiment, the load force is applied at a first location or region 150, being the proximal phalanx of the finger 120.

[0064] A first sensor 160 is provided for detecting one or more biometric parameters of the finger 120. The first sensor 160 is located adjacent an inner wall 180 of the housing 110 to provide a more accurate measurement. In this embodiment, the first sensor detects the biometric parameters at a second location or region 190, being the distal phalanx of the finger 120. The first sensor 160 detects the one or more biometric parameters when the servomechanism 130 applies the load force at the proximal phalanx 150 to determine the blood pressure of the patient.

[0065] The housing 110 comprises a base portion 200 with the inner wall 180 being an inner wall of the base portion. In this embodiment, the housing 110 also has a moveable portion 210 and the servomechanism 130 is configured to move the first moveable portion relative to the base portion 200, either toward or away from the base portion, as indicated by arrow 215. In this way, the moveable portion 210 is able to apply the load force to the finger 120 at the proximal phalanx 150. The servomechanism 130 comprises a motor 220 to drive a gear assembly 230, an L-shaped arm 240 connected to the gear assembly and an output shaft 250 connected to L-shaped arm. In operation, the motor 220 drives the gear assembly 230, which moves the L-shaped 240 and causes the output shaft 250 to apply the load force. An adjustment mechanism in the form of a screw 260 enables manual adjustment of the output shaft 250.

[0066] In this embodiment, there is a second sensor 270 for detecting one or more biometric parameters at the proximal phalanx (first location) of the finger 120. The second sensor 270 is also located adjacent the inner wall 180 of the housing 110 to provide a more accurate measurement.

[0067] The device 100 enables the measurement of one or more biometric parameters at the distal and/or proximal phalanx 190, 140. This enables measurements to be made at both the distal and proximal phalanges when there is no load force to restrict the blood flow 140 in the finger 120 and when there is a load force at the proximal phalanx 150.

[0068] The biometric parameters are one or more characteristics of the blood flow 140 in the finger 120. In this embodiment, the first and second biometric sensors 160, 270 both take the form of a photoplethysmography (PPG) sensor, which has an infrared light transmitter 160a, 270a, a red light transmitter 160b, 270b, a green light transmitter 160c, 270c and a reflect light receiver 160d, 270d. The PPG sensors 160, 270 measure the absorbance of light at the infrared, red and green wavelengths by the blood when it is free flowing and when it is restricted by the applied load force at either the proximal phalanx 150 or distal phalanx 190. The PPG sensors 160, 270 may comprise light emitting diodes (LEDs) and/or photodiodes. Alternatively, the PPG sensors may take the form of microlenses, miniature optical elements or miniature cameras. In these embodiments, there is at least a light source (multiwavelength or monochromatic) for transmitting light for detection by the microlenses, miniature optical elements or miniature cameras. For example, a multi-wavelength PPG emitter and photodiode can be used as the light source. Such PPG emitters may have a footprint of <1mm². The first sensor 160 may also comprise an inertial measurement unit (IMU) in the form of a 9-axis IMU 290 that measures specific force (via a three-axis accelerometer), angular velocity (via a three-axis gyroscope) and magnetic field (via a three-axis magnetometer) of the finger 120 at the distal phalanx 190.

[0069] A first force sensor in the form of a load cell 300 is connected between the output shaft 250 and the moveable portion 210 to measure the load force applied by the servomechanism 130. A second force sensor, also in the form of a load cell 310, is connected to holding portion 320 generally opposite to the base portion 200 for passively measuring the load force at the distal phalanx to detect any changes in skin contact between the finger 120 and the housing 110 that could permit the ingress of ambient light and thus adversely affect the measurements taken by the PPG sensor 270.

[0070] In operation, the patient places the finger 120 into the housing 110 and rests against the inner wall 180 of the base portion 200 and the moveable portion 210. For comfort of the patient, the inner wall 180 and inner walls 330 of the moveable portion 210 and holding portion 320 are inwardly curved or ergonomically shaped to accommodate the general shape of the finger 120. One or both of the sensors 160, 270 measure the biometric parameters of the blood flow 140 in the finger 120 to obtain an initial set of biometric data. The device 100 is also able to take biometric data measurements continuously from one or both locations at the proximal and distal phalanges 150, 190 to provide real time data.

[0071] The servomechanism 130 is then activated, causing the L-shaped arm 240 (and hence output shaft 250) to move towards the base portion 200, which in turn causes the moveable portion 210 to exert the load force to the finger 120, restricting the blood flow 140. The load cell 300 measures the amount of the load force while the load cell 310 is checked to determine if the finger 120 has shifted to vary the initial skin contact with the sensor 160. The blood pressure in the blood vessels pushes against this external load force, causing a relation between the load force and measured blood flow at the distal phalanx 190 (and partially the proximal phalanx 150). Another set of biometric measurements is also taken by one or both sensors 160, 270, although it is preferred that the measurement of biometric data is taken at the distal phalanx 190 and the measurement of the load force is taken at the proximal phalanx 150 by the load cell 300. This ensures that measurement of the biometric data is not adversely affected by artefacts that could be induced by the applied force. That is, applying a force to the finger 120 in the vicinity of a PPG sensor 160, 270 may cause minor movements between the PPG sensor and the skin surface, resulting in an incorrect measurement of biometric data due to the high sensitivity of PPG sensors to motion. When the blood flow 140 is restricted by the applied load force, the blood pressure in the vessels pushes against this external force, causing a relation between load force and measured blood flow at the distal phalanx (and partially proximal phalanx). Hence, the device 100 takes “snapshots” or discrete points of biometric data when the load force is applied to the proximal phalanx 150.

[0072] The two sets of data are transmitted through wires 340 (as shown in Figure 2) or wirelessly (as shown by dotted line 350 in Figure 1) to a control module (not shown) to perform a more accurate calculation of the systolic and diastolic blood pressures of the patient based on the acquired sets of biometric data and the force measurements, resulting in a more accurate calculation of the overall blood pressure. In some embodiments, the device 100 comprises the control module, which may be a central processing unit (CPU), programmable logic controller (PLC) or other computer. Alternatively, the device 100 transmits the biometric and the force measurement data to an external CPU, PLC or the computer to perform the calculation.

[0073] Also, since measurements are taken when force is being actively applied by the servomechanism 130, the relationship between the load force and measured blood flow can be used to compare against the measurements taken previously to correct for any noise or errors caused by motion artefacts, improper or variable contact with the skin and/or signal crossover. Hence, this results in a more accurate calculation of the patient’s blood pressure, provides a way for the device 100 to self-correct or rectify any errors in the calculations and enables calibration of the device 100 based on real time data. Moreover, the measurements can be taken when desired by selectively activating the servomechanism 130. Hence, the load force may be selectively applied when a measurement is required and need not be applied indiscriminately.

[0074] Figures 3 to 12 illustrates a device 500 for measuring the blood pressure in a patient according to another embodiment of the invention. In this embodiment, the housing has two moveable portions in the form of clamping portions 510, 520 arranged opposite the sensors 160, 270 at the proximal and distal phalanges 150, 190. The clamping portions 510, 520 are moveably connected to cantilevered arms 530, 540 that are removably mounted to curved arm portions 550, 560 extending substantially perpendicular from the base portion 200 from one side 570.

[0075] In addition, first and second linear actuators 580, 590 are respectively operatively connected to the clamping portion 510, 520 for selectively moving the clamping portions toward or away from the base portion 200, as indicated by arrows 215. Load cells 585, 595 are incorporated into the structure of the linear actuators 580, 590 for convenience and a more compact design. As in the earlier embodiment, the linear actuators 580, 590 are able to move the clamping portions 510, 520 so that they apply the load force to one or both of the proximal or distal phalanges 150, 190 so as to restrict the blood flow 140 in the finger 120. The amount of load force applied by the clamping portions 510, 520 are easily controlled by the linear actuators 580, 590. [0076] Consequently, biometric measurements can be taken from the proximal phalanx 150 (the first location on the finger 120) and/or the distal phalanx 190 (the second location on the finger), either when the load force is applied to the same location where the measurement is being taken or at the other location where no local load force is present. That is, the sensor 270 can detect the biometric parameters at the proximal phalanx 150 when the second linear actuator 590 applies the load force at the distal phalanx 190. Similarly, the sensor 160 can detect the biometric parameters at the distal phalanx 190 when the first linear actuator 580 applies the load force at the proximal phalanx 150. This means that more biometric data can be obtained and improve the calculation of the patient’s blood pressure.

[0077] This dual actuator arrangement of this device 500 enables the measurement of biometric parameters at the proximal and distal phalanges 150, 190 under a load force at either phalanx. It is contemplated that the first and second actuators 580, 590 alternate in applying their respective load forces at the first (proximal) and second (distal) locations, respectively. This enables measurements to be made where the load force is not limited to a single location, minimising any measurement errors that are inherent at that site or could be made at that location. However, in some embodiments, the first and second actuators 580, 590 simultaneously apply their respective load forces at the first (proximal) and second (distal) locations, respectively, for measurements to be taken at both locations at the same time. Hence, a greater variety of measurement data that can be obtained in different situations using the device 500 that significantly improve the accuracy in the calculation of blood pressure and enhance the ability of the device 500 to self-correct any errors and perform a self-calibration.

[0078] The device 500 also incorporates a removable data module in the form of a USB 600 that receives the measurement data from the sensors 160, 270, I MU 290 and load cells 300, 310 of the linear actuators 580, 590. The USB 600 communicates (via wires or wirelessly as shown by dotted line 350) to an external control module (not shown) to transmit the measurement data for calculation of systolic, diastolic and/or overall blood pressures. In one embodiment, the device 500 and/or control module can also communicate with a display (not shown) for displaying the raw biometric data and force data measurements, as well as the calculated systolic, diastolic and overall blood pressures. [0079] The housing 110 of the device 500 in this embodiment is also made in two halves, each half portion 592, 593 pivotally connected at respective hinge points or axes 595, preferably by respective arm portions 597, 598. This configuration enables the base portion 200 to be adjusted to accommodate different sized fingers 120 and well as enabling the device 500 to be folded at the hinge points 595 to have a more compact profile and improved ergonomics. As the finger 120 in a “relaxed” state tends to be partly curled or curved, rather than straight, the housing 110 can be partly folded at the hinge points 595 to accommodate this non-linear position of the finger. This also shortens the device 500, which allows for greater freedom of movement of the finger 120. Ideally, the body of the device would be designed to be as short as possible to allow for the greatest freedom of movement.

[0080] In a variation to this embodiment shown in Figures 11 and 12, the housing 110 further comprises a display 650 for displaying the same information noted above. The display 650 is mounted on a printed circuit board (PCB) 660 that has a microcontroller (not shown) coupled to a speaker 670 to provide acoustic instructions. The device 500 also has a rechargeable battery 680, control buttons 690 and a Bluetooth interface 692. Charging and communication with wired devices can be achieved by a USB type C port 695 (the cable 698 being shown in dotted lines).

[0081] In operation, the device 500 operates in substantially the same way as the device 100, but with the ability to apply the load force selectively at the proximal phalanx 150, distal phalanx 190 or at both phalanges to collect the biometric data. Figure 11 shows the device 500 with the linear actuators 580, 590 inactive, where the PPG sensors 160, 270 and I MU 290 take continuous or discreate biometric measurements. The load cells 585, 595 may also take passive measurements to check that appropriate skin contact with the sensors 160, 270 has been maintained. Figure 12 shows the device 500 with the linear actuator 590 activated and the linear actuator 580 inactive. The linear actuator 590 causes the clamping portion 520 to apply a load force (as shown by arrows 215) to the proximal phalanx 150. This time, the PPG sensors 160, 270 and IMU 290 take continuous or discrete biometric measurements as the load force is applied at the proximal phalanx 150. The load cell 595 also measures the applied load force exerted by the linear actuator 590. The measurements may be taken as the load force increases to a predetermined value and as the load force decreases when the actuator 590 is deactivated. In one embodiment, only the PPG sensor 160 and I MU 290 records biometric measurements at the distal phalanx 190 when the load force is applied at the proximal phalanx 150. In another embodiment, only the PPG sensor 270 records biometric measurements at the proximal phalanx 150 when the load force is applied to the proximal phalanx.

[0082] Alternatively, the linear actuator 580 is used to apply a load force at the distal phalanx 190 and the PPG sensors 160, 270 and IMU 290 take continuous or discrete biometric measurements as the load force is applied at the distal phalanx 190. The load cell 585 also measures the applied load force exerted by the linear actuator 580. The measurements may be taken as the load force increases to a predetermined value and as the load force decreases when the actuator 580 is deactivated. Again, in some embodiments, only the PPG sensor 160 and IMU 290 records biometric measurements at the distal phalanx 190 when the load force is applied at the distal phalanx 190. In a further embodiment, only the PPG sensor 270 records biometric measurements at the proximal phalanx 150 when the load force is applied to the distal phalanx 190.

[0083] In a further embodiment, both linear actuators 580, 590 are used the apply a load force at both the proximal and distal phalanges 150, 190 and any combination of the PPG sensors 160, 270 and IMU 290 are used take continuous or discrete biometric measurements as the load force is applied at both phalanges. Also, the load cells 585, 595 also measure the exerted force applied by the linear actuators 580, 590. It should also be noted that different load forces may be applied by the linear actuators 580, 590.

[0084] Testing has shown that the achieved signal-to-noise ratio (SNR) of the photoplethysmographic (PPG) signals generated by the system 100, 500 is advantageous in obtaining accurate and reliable BP measurements. A high SNR ensures that the physiological data obtained is clear and less susceptible to interference from noise, which is essential for medical and health monitoring applications.

[0085] Consequently, Figures 13 and 14 illustrate a system 700 for measuring the blood pressure in a patient according to a further embodiment of the invention designed to provide a high SNR. This embodiment is substantially the same as the embodiment of Figures 3 to 12, except that instead of a single PPG sensor at each (first and second) location, there are multiple PPG sensors 705 having multiple wavelengths. The PPG sensors 705 are arranged into groups 707, which in this embodiment take the form of microarrays, as best shown in Figure 14. The microarrays 707 measure the spatial propagation of the blood flow in the finger, thus generating a large number of PPG input signals.

[0086] The spatial propagation of blood flow is an important factor because it provides detailed information about the hemodynamics within the vascular system of the finger 120. These microarrays 707 use a combination or matrix of multiple wavelengths emitted by the LEDs and/or photodiodes. Each wavelength is carefully selected to penetrate the skin at different depths, allowing the microarrays 707 to capture comprehensive data about the blood flow at various layers within the tissue.

[0087] This configuration enables the simultaneous capture of PPG signals from multiple points on the finger 120, as best shown in Figure 14, providing a detailed spatial map of blood flow propagation. This spatial mapping enhances understanding of the dynamics of blood flow and to calculate the blood pressure.

[0088] The fusion or combination of the large number of PPG input signals can significantly enhance the robustness and accuracy of the measurements, especially where deep learning is used to calculate the blood pressure. By leveraging the high SNR, deep learning algorithms can more effectively analyse the PPG signals to extract a more accurate blood pressure measurement.

[0089] Where the photodiodes used for detection are replaced by microlenses or miniature optical elements, they allow for the focusing of light with high precision, further enhancing the quality and resolution of the captured signals. Where the photodiodes used for detection are replaced by cameras, they can capture visual information that can be processed using image analysis techniques to complement the PPG data, providing an even richer set of data for analysis.

[0090] Figure 14 illustrates a bottom schematic view of the locations of the microarrays 707, showing their arrangement relative to the finger 120. The microarrays 707 are arranged in parallel to each other, preferably laterally aligned to each other relative to the longitudinal axis of the finger 120. Thus, the microarrays 707 are organised into groups 709a, 709b around each respective location 150, 190 of the finger 120. Thus, they have a spatial configuration spread along the finger 120 to capture the spatial propagation of blood flow.

[0091] As an alternative to using reflective PPG signals, the system 700 can be configured to utilise transmissive PPG signals. In this configuration, light emitted from the LEDs passes through the tissue, and the transmitted light is detected by photodiodes or cameras positioned on the opposite side of the tissue.

[0092] Referring to Figures 15 to 19, a system and method of measuring the blood pressure of the patient using the devices 100, 500 according to an embodiment of the invention will now be described. To provide some context for the method, a brief discussion of calculating blood pressure from blood flow in a patient’s limb is set out below.

[0093] With each cardiac cycle, blood is pumped to the body’s periphery. Although the flow is damped by the flexibility of human tissue, blood flow is not constant and can be observed as a pressure wave. This pressure wave can be monitored non-invasively through PPG sensors. Photoplethysmography illuminates the skin and measures changes in light absorption induced by varying blood flow, therefore measuring blood volume changes.

[0094] The shape and timing of each heartbeat’s pressure wave contains information about the respective blood pressure. However, this is physiologically not uniform for the whole population but overlaid with an individual component depending on the subject/person. This effect is exacerbated where the patient is mobile and not still, as tends to occur in non-clinical environments, inducing motion artefacts that can confuse the PPG sensor. Hence, it is therefore required to calibrate the method for each patient.

[0095] As described above, the devices 100, 500 enable various sets of biometric data to be obtained from the PPG sensors 160, 270, IMU 290 and load cells 300, 310, 585, 595 by taking measurements at various times. The relation between the measurements taken by the PPG sensors 160, 270 and the force measurements from the load cells 300, 310, 585, 595 can be used to calibrate the PPG signals (in the form of waveforms) that are recorded when there is not a load force applied to the finger 120. The detected signals are composed of AC and DC components. Typically, the DC component 860 is the 0 Hz component that acts as an offset in the time domain, as best shown in Figure 17. The AC component typically comprises all other frequencies of the signal 830, also shown in Figure 17. The measurements taken by the IMU 290 reduce the effects of any noise created by motion artefacts induced by motion of the finger 120. Consequently, subsequent measurements by the PPG sensors 160, 270 (without the presence of a load force) can be calibrated for the particular patient. This improves the accuracy of the measurement and hence the calculation of the blood pressure of the patient.

[0096] In this embodiment of the invention, a neural network (NN) is used to perform the calibration and calculation of the blood pressure, either as part of the control module or as an external component. Other types of artificial intelligence (Al) or machine learning may be employed to perform the calibration and calculation, such as deep learning systems and the like, as well as machine learning regression algorithms like Decision Tree, SVM, Lasso, Random Forest and Linear regression algorithms.

[0097] Figure 15 illustrates a schematic diagram of a system 710 according to one embodiment of the invention, employing an NN in the form of a deep learning module 715 and one of the devices 100, 500. The system 710 comprises the device 100, 500, a control module 720 and a display unit 730. As discussed above, the control module may be a CPU, PLC or other computer, as well as a smart device, such as a smartphone. The control module 720 includes both the deep learning module 715 and a supervisor module 740.

[0098] The system may optionally include a user interface (Ul) 750 and/or links to telemedicine displays or devices 760. The Ul 750 may be an external monitor, computer or smart device (smartphone or tablet). The telemedicine display 760 may display the continuous blood pressure, the raw measurement data and other biometric information that can be derived from the measurement data, including heart rate, heart rate variability, saturation level of oxygen (SpO2), perfusion index and ECG. The telemedicine display 760 is able to provide long term monitoring of this biometric information.

[0099] The system 710 utilises an initial set of measurement data to help train the NN. For example, the PPG sensors 160, 270, IMU 290 and load cells 300, 310, 585, 595 take an initial set of measurements and then the set of measurements taken when the load force is applied is used to calibrate the initial set of measurements. Alternatively, the initial set of measurement data can be sourced from a prior set of known biometric data obtained from a set of known individuals, as discussed below. In a further embodiment, the initial set of measurement data can be a combination of both sources; i.e. taken by the PPG sensors 160, 270, IMU 290 and load cells 300, 310, 585, 595 and from the known biometric data. This initial set of measurement data is called “pretraining data” for its purpose in training the NN before it analyses a set of measurement data from the device 100, 500 in respect of a particular patient.

[0100] In this embodiment, the device 100, 500, 700 continuously records the biometric data, and this set of data is called “continuous waveform” data 770, referring to the continuous waveforms created by the continuous biometric measurement taken by the PPG sensors 160, 270, IMU 290 and load cells 300, 310, 585, 595. When the load force is applied by the device 100, 500, the recorded biometric data is kept as a separate set of data, called “triggered actuator” data 780. There is also corresponding pretraining data for both data sets, one for the continuous waveform data (pretraining data 775) and another for the triggered actuator data (pretraining data 785). The pretraining data 775, 785 and the measurement data 770, 780 is then sent by the device 100, 500 to the control module 720 as two data streams; one being the continuous waveform data stream 800 and the other being the triggered actuator data stream 810.

[0101] The deep learning module 715 then proceeds to calibrate the measurement data and the resultant calibrated measurement is then sent to the display unit 730 to display the continuous blood pressure and related biometric information mentioned above, including heart rate, heart rate variability, saturation level of oxygen (SpO2), perfusion index and ECG. The same information can also be transmitted from the control module 720 to the Ul 750 and telemedicine display 760. [0102] The pretraining data 775 for the continuous waveform data 770 involves taking measurements from participants ideally from diverse backgrounds and with a high blood pressure (BP) variation using the device 100, 500. Mild exercise was used to increase the BP of participants. While recording data with the device 100, 500, 700 the BP was also recorded from a commercial BP monitor to correlate the results. The data from all participants is then aggregated and “cleaned”, including removal of as much noise as possible using the measurements taken by the IMU 290, to obtain the pretraining data 775.

[0103] The pretraining data 785 for the triggered actuator data 780 is obtained in the same way, but the BP is directly measured when the load force is applied (i.e. servomechanism 130 or actuator 580, 590 is triggered) and not continuously.

[0104] The deep learning module 715 receives the continuous waveform data stream 800 and the triggered actuator data stream 810 from the device 100, 500. The deep learning module 710 uses the pretraining data 775 from the population of participants and the measurement data 770 of the particular patient from the continuous waveform data stream 800 to calculate a preliminary BP value of the patient.

[0105] The deep learning module 715 also uses the pretraining data 785 from the population of participants and the measurement data 780 of the particular patient from the triggered actuator data stream 810 to calculate an adjusted BP value for the patient. This adjusted BP value is then used to calibrate the preliminary BP value to a final measured value of the BP that is displayed on the display unit 730, III 750 and telemedicine display 760. Subsequent changes in BP measured by the device 100, 500 are then displayed based on the continuous waveform data 770 being generated.

Hence, once pretraining is completed, the directly measured BP by the device 100, 500 is used to calibrate the continuous waveform data 770 and this continuous BP is displayed.

[0106] The supervisor module 740 continuously monitors the output of the deep learning module 715 and a confidence score 820 generated by the deep learning module to indicate reliability of the prediction. It will be understood by those skilled in the art that the confidence score is a by-product of many deep learning models and can be typically accessed through the deep learning framework API https://rdrr.io/cran/keras/man/predict_proba.html. Deep learning methods often predict probabilities for different labels internally and combine these with an operation such as argmax or arg max, which is a mathematical operation that finds the argument that gives the maximum value of a designated function. That is, argmax finds the input values for a particular function that will return the highest or maximum output values. In the context of machine learning, argmax would find the values that will give the highest predicted probability, and in the case of this embodiment of the invention, the most accurate value for BP, If the confidence score 820 drops below a threshold or if a certain time period has passed, the control module 720 sends a signal to the device 100, 500, 700 to active the servomechanism 130 or linear actuator(s) 580, 590 to apply a load force and repeat the above algorithm to retrain the deep learning module 715.

[0107] As discussed above, the method involves a machine learning method combining two measurement methodologies; continuous short term PPG and pressure waveforms (<5 seconds interval) with conditionally triggered actuations of restricted blood flow. This involved the fusion of the resulting continuous waveform data 770 in the series domain (time series) with transformed data in the pressure (applied load force) domain. The PPG waveform at a designated position “0” (for example, the distal phalanx 190) is recorded continuously. When the actuator at a designated position “1” (for example, the proximal phalanx 150) activates, it temporarily restricts the blood flow at position 0. To combine these two data sets for deep learning, a data transformation was developed to create 2D data for finger blood flow depending on exerted force. As a measure of blood flow, a metric was developed using the following steps illustrated in Figures 16(a), 16(b) and 16(c). These figures illustrate the actuator position (Figure 16(a)); continuous time series from 1 PPG channel (PPG location 0 red) (Figure 16(b)) and extracted AC component and peak-peak amplitude (Figure 16(c)). The blood flow at location 0 is shown restricted, depending on the actuator position at location 1 (e.g. samples 10,000-20,000).

[0108] Figure 17 shows an actual measurement for one PPG channel signal 830, where the DC component 860 acts as an offset in the time domain of the signal 830. There is a region or “envelope” 870 defined by boundary lines 880 around the peak-peak amplitude of the AC component of the signal 830. [0109] The AC component of the waveform is continuously calculated for each PPG and loadcell waveform channel (only one channel PPG0_R shown) by finding the rolling upper and lower envelope of the signal (PPG0_R_DC_min and PPG0_R_DC_max). The rolling mean PPG0_R_DC is consecutively removed from the PPG signal to gain only the AC envelope (PPG0_R_AC_min and PPG0_R_AC_max). The absolute signal amplitude PPGO_R_AC_p-p_amplitude can be calculated from PPG0_R_AC_max- PPG0_R_AC_min.

[0110] To fuse this data with the continuous PPG data, it is transformed by creating recorded force bins for the load cell at location 1 , as shown by the x-axis of the graphs in Figures 18(a) and 18(b). These force bins are discrete ranges of lower to upper force limits through the range from no applied force to the maximum comfortable attachment pressure. The ranges were selected in the order of 1k bins. The y-axis of each graph shows the PPGO_R_AC_p-p_amplitude. It was observed that some hysteresis effects were present between increasing and decreasing force data. Accordingly, to take into account these hysteresis effects and to provide additional data for the deep learning module 715, increasing and decreasing force data are transformed separately, as shown in Figures 18(a) and 18(b).

[0111] However, not all force bins will always occur with a measurement data point. Therefore, to improve the accuracy of the machine learning method, these values were interpolated as shown in Figures 19(a) and 19(b). Best results were achieved with a momentum-based spline interpolation but replaced by a linear interpolation due to its computational expense.

[0112] For the measured continuous waveform data 770, a rolling window with timeseries of all PPG channels (2x3), two load cell channels and 9-axis IMU is fused and used as the group of independent variables X for the dependent variable Y, being blood pressure. It will be readily understood by those skilled in the art that X and Y are commonly used in machine learning to designate the independent and dependent variables, respectively, and that X and Y are typically split into training and test data sometimes additionally into validation data. Example

[0113] In a comparative test between an embodiment of the invention (device 100), the correlation of the blood pressure measurements was assessed against clinical standard blood pressure monitors in common use.

[0114] The blood pressure of seven volunteer participants aged 25-55 was tested using a commercial cuff-based BP monitor and compared with the device 100. The BP of each participant was artificially raised through exercising on an exercise bike.

[0115] Commercial cuff-based monitors take in the order of a minute to measure and it is not exactly defined at which point during this timeframe BP is taken. As commercial BP monitors are not 100% accurate, BP naturally fluctuates and likely to continuously drop after the exercise (not measuring at the exact time as the device 100), some error in the commercial BP monitor is to be expected. The achieved overall absolute accuracy of the seven participants was 6.9 mm Hg, as shown in Table 1 below.

Table 1 : Accuracy for seven test participants

Sample Systolic BP (commercial) Systolic BP (Invention) RMSE [mm Hg] [mm Hg] [mm Hg]

1 91 93.1 2.1

2 87 89.8 2.8

3 93 92.8 0.2

4 111 107.5 3.5

5 104 111.1 7.1

6 106 89.5 16.5

7 112 102.1 9.9

8 136 140.7 4.7

9 133 117.4 15.6

10 118 117.4 0.6

11 116 118.0 2.0

12 149 136.0 13.0

13 156 163.2 7.2

14 122 136.5 14.5

15 122 122.1 0.1

16 129 142.7 13.7

17 133 136.4 3.4

18 122 120.1 1.9

19 118 129.1 11.1

20 130 135.5 5.5

21 140 142.1 2.1

22 128 114.7 13.3

Average 120.7 120.8 6.9 [0116] It can be observed that the root mean square error (RMSE) is relatively low at 6.9, indicating that the measured (and calibrated) blood pressure from the device 100. is relatively close in accuracy compared to commercial BP monitors.

[0117] The embodiments on the invention have been described as employing biometric sensors 160, 270 at the proximal and distal phalanges 150, 190 of a finger 120. However, it will be appreciated that the invention is able to achieve the substantially the same or similar effects by using a single sensor or set of sensors at only one location.

For example, there may only be the sensor 160 (and optionally the IMU 290) at the distal phalanx 190 and only the actuator 130, 580 is at the proximal phalanx 150 and/or the actuator 590 at the distal phalanx. Alternatively, there may only be a sensor 270 at the proximal phalanx 150 and the actuator 130, 580 at the proximal phalanx and/or the actuator 590 at the distal phalanx 190.

[0118] In yet other embodiments of the invention, more than one actuator may be provided at distal and/or proximal phalanges 190, 150 to move the respective moveable portion 210 or clamping portions 510, 520. Similarly, there may be a plurality of actuators at a plurality of locations, with or without associated biometric sensors.

[0119] In further embodiments of the invention, the control module may use a predetermined set of criteria or data to determine the confidence score, instead of using the measured data or pretraining data. In some embodiments, the control module may use a fixed set of data to determine an initial confidence score that is subsequently adjusted in response to the pretraining data or measurement data from the devices 100, 500.

[0120] While the embodiments of the invention have been described as using servomechanisms and linear actuators, in some embodiments other types of actuators may be used, such as electromagnetic and preloaded spring supported actuators. Similarly, the sensors are not limited to PPG sensors, inertial measurement units or load cells, but can include other biometric sensors and sensors in general. Examples of suitable biosensors include blood glucose, sweat glucose, blood lactate and sweat lactate biosensors. Examples of sensors include sensors measuring pH, oxygen saturation, humidity and galvanic skin response.

[0121] It will also be appreciated that although the finger 120 is used to take the biometric measurements, in other embodiments, different sites of the patient can be chosen. For example, another appendage of one of the patient’s limbs, such as a thumb or even a toe.

[0122] It will further be appreciated that any of the features in the preferred embodiments of the invention can be combined together and are not necessarily applied in isolation from each other. For example, the clamping portions 510, 520 can be implemented in the device 100 or a servomechanism 130 could be used to replace one of the linear actuators 580, 590 . Similar combinations of two or more features from the above described embodiments or preferred forms of the invention can be readily made by one skilled in the art.

[0123] By providing an actuator to apply a load force and a sensor for measuring one or more biometric parameters when the actuator applies the load force in a monitoring or measuring device, the invention confers the advantage of enabling a more accurate measurement of biometric parameters like blood pressure, self-correction or adjustment of previous measurements made by the device and calibration of the device. This in turn enables the device to be used as a portable device capable of providing continuous blood monitoring to permit a patient to be mobile or active while still being able to accurately measure the patient’s blood pressure and other related biometric information. This means that a patient’s blood pressure can be measured in non-clinical environments with greater accuracy than wearable devices of a similar nature. This also permits continuous blood pressure monitoring over a longer period without extended visits to hospitals or medical clinics or significantly disrupting the patient’s day to day activities. It also enables more realistic measurements to be made in the patient’s usual environment and not in an artificial clinical environment. Moreover, the ability to selective apply the load force and take measurements in response to that load force enables the invention to self-calibrate its measurements continuously, leading to greater accuracy and a more realistic representation of a patient’s blood pressure throughout the day. All these advantages of the invention result in greater patient comfort, less burden on hospitals and medical clinics and more personalised care. In all these respects, the invention represents a practical and commercially significant improvement over the prior art.

[0124] Although the invention has been described with reference to specific examples, it will be appreciated by those skilled in the art that the invention may be embodied in many other forms.

Claims

1. A device for measuring the blood pressure in a patient, comprising: a housing for receiving a limb of the patient; a first actuator for applying a load force at a first location of the limb to restrict a blood flow in the limb; one or more first sensors located adjacent an inner wall of the housing for detecting one or more biometric parameters at a second location of the limb; and one or more second sensors adjacent the inner wall for detecting the one or more biometric parameters at the first location of the limb; wherein the one or more first sensors detect the one or more biometric parameters at the second location when the first actuator applies the load force at the first location to determine the blood pressure of the patient; and wherein the one or more second sensors detect the one or more biometric parameters at the first location when the first actuator does not apply the load force at the first location.

2. The device of claim 1, wherein the housing comprises a first moveable portion, wherein the first actuator is configured to move the first moveable portion relative to a base portion of the housing, and the first moveable portion comprises a clamping portion for engaging the limb.

3. The device of claim 1 or 2, wherein a first force sensor is operatively connected to the first actuator to measure the load force applied by the first actuator.

4. The device of any one of the preceding claims, wherein the first actuator applies the load force at the first location and the one or more first sensors detect the one or more biometric parameters at the first location.

5. The device of any one of the preceding claims, comprising a second actuator for applying a load force to the limb at the second location; wherein the one or more second sensors detect the one or more biometric parameters when the second actuator applies the load force at the first location.

6. The device of claim 5, comprising a second moveable portion, wherein the second actuator is configured to move the second moveable portion relative to the base portion.

7. The device of claim 5 or 6, wherein the first and second actuators alternate in applying their respective load forces at the first and second locations, respectively.

8. The device of claim 5 or 6, wherein the first and second actuators simultaneously apply their respective load forces at the first and second locations, respectively.

9. The device of any one of claims 5 to 8, wherein the detected biometric parameters from the one or more first and second sensors are used to calculate the blood pressure of the patient.

10. The device of any one of claims 5 to 9, wherein a second force sensor is operatively connected to the second actuator to measure the load force applied by the second actuator.

11. The device of any one of the preceding claims, wherein the one or more first sensor and/or the one or more second sensors each comprises a photoplethysmography (PPG) sensor and/or an inertial measurement unit (IMU).

12. The device of claim 10, wherein there is a plurality of first sensors and/or second sensors, wherein the plurality of first sensors and/or second sensors are arranged to measure the spatial propagation of blood flow in the limb.

13. The device of claim 12, wherein the plurality of first sensors and/or second sensors are arranged in groups, wherein the groups are spatially separated from each other and are each aligned along a transverse axis to a longitudinal axis of the limb.

14. A method of measuring the blood pressure of a patient, comprising: applying a load force at a first location of a limb of the patient to restrict the blood flow in the limb; in response to the load force applying step, taking a measurement at a second location of the limb of one or more biometric parameters to create a first set of data; taking a measurement at the first location of one or more biometric parameters when the load force is not applied to the first location to create a second set of data; and calculating the blood pressure of the patient from the first set of data and second sets of data.

15. The method of claim 14, wherein the measurement at the first location is taken prior to or after the load force applying step.

16. The method of claim 14 or 15, comprising calibrating the second set of data with the first set of data.

17. The method of any one of claims 14 to 16, comprising: applying a load force at the second location to restrict the blood flow in the limb; in response to the load force applying step at the second location, taking a measurement at the first location of the limb of one or more characteristics of the blood flow to create a third set of data; and calculating the blood pressure of the patient from the first and third sets of data.

18. The method of claim 16, comprising calibrating the second set of data with the first set of data and/or third set of data.

19. The method of any one of claims 14 to 18, wherein a neural network (NN) performs the calculating step, wherein the NN is trained with a preliminary set of data prior to performing the calculating step and/or prior to taking the measurement at the first location.

20. The method of claim 18, wherein a neural network (NN) performs the calibrating step, wherein the NN compares the calculated blood pressure against the second set of data in real time and determines a confidence value in the accuracy of the calculated blood pressure.

21. The method of claim 20, wherein, in response to the confidence level falling to or below a threshold value, the NN adjusts the calculated blood pressure based on the second set of data or the NN initiates the load force applying step and repeats the measurement at the second location to obtain a fourth set of data and repeats the calculating step using the fourth set of data.

22. The method of any one of claims 13 to 21, wherein the measurement at the first location is performed continuously to create the second set of data

23. The method of any one of claims 14 to 22, wherein the load force is applied periodically.

24. The method of any one of claims 14 to 23, wherein the measurement at the first location also comprises measuring the load force applied at the first location to create a set of force data.

25. The method of claim 24, wherein the calculating step comprises applying a transformation of the set of force data and combining the transformed force data with the second set of data.

26. The method of claim 25, wherein the transformation comprises selecting an AC component of the set of force data and segregating the AC component into discrete data clusters.

27. The method of claim 26, wherein, where a discrete data cluster lacks data, data values are interpolated and allocated to the empty discrete data cluster.

28. A system for determining the blood pressure of a patient, comprising: the blood pressure monitoring device of any one of claims 1 to 13; and a neural network module for receiving the first set of data, the neutral network being configured to perform the method of any one of claims 14 to 27.

29. The system of claim Error! Reference source not found., comprising a display for displaying the calculated blood pressure.