Beat-by-Beat Monitoring of Systolic Blood Pressure based on an

ASIC and a Mobile Phone for Ambulatory Application

Wenxi Chen

, Ming Huang

, Xin Zhu

, Kei-ichiro Kitamura

and Tetsu Nemoto

Biomedical Information Technology Laboratory, The University of Aizu, Tsuruga, Aizu-wakamatsu City, Japan

Department of Laboratory Science, Faculty of Health Sciences, Institute of Medical, Pharmaceutical and Health Sciences,

Kanazawa University, Kodatsuno, Kanazawa City, Japan

Keywords: Blood Pressure, Electrocardiogram, Photoelectric Plethysmogram, Pulse Arrival Time, Asic, Mobile Phone,

Ambulatory Monitoring, Daily Healthcare.

Abstract: This paper describes an ambulatory monitor for beat-by-beat monitoring of systolic blood pressure (SBP)

based on an ASIC chip and a mobile phone. The ASIC is able to measure electrocardiogram (ECG),

photoelectric plethysmogram (PPG), and has a peripheral interface to control an air pump and valve for

inflating and deflating a sphygmomanometer cuff in conventional blood pressure measurement. Algorithms

for signal processing, characteristic point detection and SBP estimation are implemented on a mobile phone.

Pulse arrival time (PAT) is derived from the apex of QRS complex to the maximum slope of PPG, and is

used to estimate a rapid change component in SBP beat-by-beat. An oscillometric sphygmomanometer with

a cuff is used to determine SBP intermittently for calibration purpose. Data communication between a

mobile phone and the ambulatory monitor is conducted via a Bluetooth wireless connection. Performance of

the prototype is examined by data from five healthy college students. The results show that 65.9% of

estimated SBP fall into ±5% relative error, 96.6% in ±10% and 99.7% in ±15%. This prototype is a pilot

study aiming at integrating an innovative sphygmomanometry into a mobile phone for continuous blood

pressure monitoring. We expect to find potential applications in ambulatory monitoring and daily

healthcare.

1 INTRODUCTION

Ambulatory monitoring of multiple vital signs

attracts more and more attention in daily healthcare

domain. Monitoring devices with smaller size and

lighter weight are desirable. With dramatic

advancement in information technologies, a mobile

phone is nowadays not only a communication tool,

but also provides a universal platform with versatile

interfaces, large amount of computational power and

high capacity of data communication. Widespread

utilization of mobile phones in daily life makes it

practical and acceptable to extend their applications

in ambulatory healthcare by integrating advanced

biomedical sensors and functionalities.

Ambulatory applications based on mobile phones

have been widely explored in various aspects such

as diagnosis, monitoring and health management.

Without requiring any additional accessories, a

built-in camera can be used to monitor breath rate

and pulse rate simultaneously by analyzing chest

movement during breathing and chromatic tone

change due to blood flush during heartbeating in a

series of images (Philips, 2011).

More professionally, an external device with

specific function is connected to a mobile phone via

wired or wireless means to turn a mobile phone into

a significantly efficient medical checkup equipment.

For example, an electrode pad is attached to an

iPhone or Android smartphones to monitor ECG and

detect heart rate in real-time mode without directly

contacting the electrodes to the body surface

(AliveCor, 2011). Although it does not serve to

diagnose acute myocadial infarction, an ambulatory

ECG can mark cardiac events in emergency situation

conveniently.

A B-mode ultrasound imaging device based on a

mobile phone was developed to offer an imaging

tool for kidney, liver, eye and uterine screenings

(Richard and Zar, 2009). It is realized by simply

connecting a portable ultrasound probe to a mobile

phone via a USB cable. A mobile phone implements

Chen W., Huang M., Zhu X., Kitamura K. and Nemoto T..

Beat-by-Beat Monitoring of Systolic Blood Pressure based on an ASIC and a Mobile Phone for Ambulatory Application.

DOI: 10.5220/0004192600100013

In Proceedings of the International Conference on Biomedical Electronics and Devices (BIODEVICES-2013), pages 10-13

ISBN: 978-989-8565-34-1

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

algorithms for signal processing and image

reconstruction, and serves as a GUI display.

Peripherals applicable to mobile phones include

an electronic stethoscope for cardiac auscultation

(Bentley, 2011), a sphygmomanometer cuff for

blood pressure measurement (Withings, 2011), an

acousticon-type device for multiple vital signs

monitoring (CIM, 2008), and many others (Liu and

Liu, 2011).

Our study aims at developing an ASIC for

mobile phone application, and exploring the

possibility of ambulatorily monitoring systolic blood

pressure beat-by-beat based on the ASIC and a

mobile phone.

In this paper, we describe implementation of a

prototype using the ASIC for monitoring

electrocardiogram (ECG) and photoelectric

plethysmogram (PPG) continuously, and estimating

systolic blood pressure (SBP) beat-by-beat by using

pulse arrival time (PAT) and intermittent calibration

method proposed by (Chen et al., 2000). We

examined five subjects to assess the performance of

the prototype by comparing the estimated SBP with

reference from a tonometer beat-by-beat.

2 METHOD

This prototype consists of three parts: a conventional

sphygmomanometer with a cuff, an ambulatory unit

and a mobile phone, as showed in figure 1.

Figure 1: Three main parts of the monitor. A conventional

sphygmomanometer (ABS box) with a cuff (deep blue) in

the left; a mobile phone in the right; an ambulatory unit

making use of a mouse crust above the ABS box.

The sphygmomanometer is used to measure SBP

intermittently, for example once a day, as initial

value for calibration of estimated beat-by-beat SBP.

It is not necessary to be portable. The size is

110×196×37mm

The ambulatory unit embeds the ASIC and other

peripheral circuits into a mouse crust. This unit is

small to be portable and used to measure ECG and

PPG, and therefore to derive PAT beat-by-beat for

estimation of SBP.

A mobile phone is used to implement various

algorithms, such as signal processing, characteristic

point detection for ECG and PPG, and SBP

estimation. It also serves as a GUI terminal for data

visualization and transmission.

All communications for data and commands

among the sphygmomanometer, the ambulatory unit

and a mobile phone are available via a Bluetooth

connection. Daily results can be transmitted to a

specific database server via mobile network and be

accumulated for further data mining.

2.1 Sphygmomanometer

The sphygmomanometer with a built-in air pump

and valve can inflate and deflate the cuff by UART

commands from a mobile phone wirelessly. Pressure

change in the cuff can be monitored in real-time and

transmitted to a mobile phone.

Figure 2 shows pressure profiles in the cuff

during inflation and deflation. The cuff is firstly

inflated to approximately 150 mmHg and then

deflated gradually as showed by the pink trace (DC).

At the same time, the oscillation component in the

cuff is amplified as showed by the blue trace (AC).

The AC signal is used to determine SBP and mean

BP by the oscillometric method.

Figure 2: Cuff pressure profiles during inflation and

deflation procedures. Pink trace indicates a direct

component (DC) in the pressure signal. Blue trace is the

alternating component (AC).

2.2 Ambulatory Unit

The ambulatory unit utilizes a mouse crust to mount

the electronic circuit board, including the ASIC, for

measurements of Lead I ECG and two wavelengths

of PPG.

Time (sec)

Pressure (mmHg)

Beat-by-BeatMonitoringofSystolicBloodPressurebasedonanASICandaMobilePhoneforAmbulatoryApplication

The ASIC is customized to integrate several

functional blocks with analogue and digital mixed

circuits, such as front-end amplifiers, poor contact

detector, auto gain controller, filters and other logic

circuits. The ASIC is fabricated by UMC 0.18um

mixed-mode process with 1.8V and 3.3V dual power

supplies (United Microelectronics Corp., Taiwan).

Measured analogue signals are digitized by 12-

bit AD converters and transmitted to a mobile phone

as showed in figure 3 (right). Two Ag/AgCl film

electrodes are attached on two sides of the mouse

crust contacted by left hand. Another electrode

surrounding the light hole is touched by right hand.

Photo diode and detector are deployed in the hole of

the mouse. In this way, the ambulatory unit can

measure ECG and PPG simultaneously by two hands

as showed in figure 3 (left).

Figure 3: Left: Measurement of ECG and PPG using the

ambulatory unit. Two fingers of left hand touch one

electrode and right forefinger touches another to measure

Lead I ECG. At the same time, right forefinger is also

used to measure PPG. Right: Sample snapshot on a mobile

phone display captured by a camera. Upper trace indicates

ECG and heart rate (69). Lower trace indicates PPG and

the corresponding PAT (252) and estimated SBP (127).

The ASIC chip (middle lower) inside the mouse crust is

7×7mm

in size.

2.3 Estimation of SBP Beat-by-Beat

SBP is estimated beat-by-beat through PAT based

on the algorithm described in (Chen et al., 2000).

Definition of PAT varies on different studies. Beat-

by-beat PAT is usually defined as a time interval

from apex of an ECG’s QRS complex to the

following nadir or the maximum slope of PPG in the

corresponding heartbeat. Because the maximum

slope is more distinguishable than the nadir, we

choose the former to derive PAT in this study.

Estimated SBP is obtained by two steps:

intermittent calibration and beat-by-beat estimation.

Once an initial SBP calibration value is attained by

using a conventional sphygmomanometry method,

change in beat-by-beat SBP is tracked by summation

of the calibrated SBP value and the filtered PAT

value. This procedure is illustrated in figure 4.

Figure 4: SBP is estimated by summation of higher

frequency component (HFC) from filtered PAT and lower

frequency component (LFC) from measured SBP using a

sphygmomanometer intermittently. The PAT is rescaled to

a regular interval of 1 sec before passing through a narrow

band-pass filter, which had a low cut-off frequency F

0.00053 Hz, and high cut-off frequency F

= 0.004 Hz.

3 RESULTS

Five college students at their twenties were involved

in measurement and performance evaluation. Data

were collected from each subject for about ten

minutes. Blood pressure was elevated through cold

water immersion test. Estimated SBP was compared

with SBP measured simultaneously by a noninvasive

continuous blood pressure monitor (Jentow-7700,

Nihon Colin, Aichi, Japan), which measures radial

arterial pressure by the tonometry method.

The performance was evaluated by relative error

between the reference and the estimation. Table 1

summarizes the probability distribution of relative

error in five subjects. With a wide range of SBP

from 108 to 163 mmHg, nearly 100% of estimations

fall within the range of ±15% relative error.

Table 1: Summary of probability distribution of relative

error in five subjects with different SBP range during cold

water immersion test.

No.

Probability and

Relative Error (%)

Range of

SBP

(mmHg)

Data

Length

(min)

±15 ±10 ±5

1 100 97.3 49.8 120-163 7

2 100 96.1 71.4 115-147 10

3 100 96.7 62.5 108-139 10

4 100 95.8 68.4 124-156 10

5 98.6 97.0 77.2 116-152 11

Avg 99.7 96.6 65.9 9.6

Std 0.6 0.6 10.4 1.5

Figure 5 shows the continuous profiles of subject

BIODEVICES2013-InternationalConferenceonBiomedicalElectronicsandDevices

No. 1 in Table 1. Figure 6 shows the histogram of

relative error distribution of the same subject.

Figure 5: Comparison of the measured SBP and the

estimated SBP. Blue trace is the measured SBP by the

tonometer and green trace is the estimated SBP by our

prototype in the upper plot. Red trace in the lower plot

indicates the relative error between reference SBP and

estimated SBP.

Figure 6: Relative error distribution. About half of the

estimations have ±5% relative error. 97.3% of the

estimations fall into ±10% range of relative error. All of

the estimations are within ±15%.

4 DISCUSSION

Because mobile phones are becoming more

powerful than ever before, more algorithms are able

to be implemented within the mobile phone platform

in real-time processing mode; more information can

be bundled into the ambulatory monitor. Location-

based and personalized services are possible to be

provided in various scenarios.

However, a mobile phone with smaller size

brings about severe electromagnetic interference

which leads to difficulty in designing the ASIC.

Improving SNR in the ASIC is a big challenge,

especially in a compact volume of space embracing

a wide range of working spectral band.

Discrepancies between the reference and the

estimated SBP beat-by-beat are relatively low and

practical in daily use if we consider ±15% relative

error acceptable. Nevertheless, there is still a

vacancy to improve estimation accuracies if higher

SNR can be achieved.

5 CONCLUSIONS

We developed an ASIC-based ambulatory monitor

for measurement of ECG and PPG, and evaluated its

performance in estimation of beat-by-beat SBP on a

mobile phone platform. It is confirmed that the

ambulatory monitor based on the ASIC and a mobile

phone is promising in daily healthcare application.

ACKNOWLEDGEMENTS

This study was supported in part by JST innovative

business model project through cooperation with

Medical Trust Co. Ltd., Saitama, Japan.

REFERENCES

AliveCor, 2011, Mobile-powered devices. Available from:

http://alivecor.com/. [8 Nov. 2012].

Bentley, P., 2011, iStethoscope Pro. Available from:

http://www.peterjbentley.com/. [8 Nov. 2012].

Chen, W., Kobayashi, T., Ichikawa, S., Takeuchi, Y.,

Togawa, T., 2000. Continuous Estimation of Systolic

Blood Pressure Using the Pulse Arrival Time and

Intermittent Calibration. Med. & Bio. Eng. & Comp.,

38(5), pp. 569-574.

CIM, 2008, Care In Motion. Available from: http://cim-

t.net/. [8 Nov. 2012].

Liu, L. and Liu, J., 2011. Biomedical sensor technologies

on the platform of mobile phones, Frontiers of

Mechanical Engineering, 6(2), pp. 160-175.

Philips, 2011, Philips launches personal App to measure

heart rate and breathing rate. Available from:

http://www.newscenter.philips.com/. [8 Nov. 2012].

Richard, W. and Zar, D., 2009, Ultrasound imaging now

possible with a smartphone. Available from:

http://news.wustl.edu/. [8 Nov. 2012].

Withings, 2011, Smart Blood Pressure Monitor. Available

from: http://www.withings.com/. [8 Nov. 2012].

Beat-by-BeatMonitoringofSystolicBloodPressurebasedonanASICandaMobilePhoneforAmbulatoryApplication