Building IoT-Enabled Wearable Medical Devices: An Application to a

Wearable, Multiparametric, Cardiorespiratory Sensor

Arthur Gatouillat

, Bertrand Massot

, Youakim Badr

, Ervin Sejdi

and Claudine Gehin

Univ. Lyon, INSA-Lyon, LIRIS, UMR5205, F-69621, France

Univ. Lyon, INSA Lyon, INL, UMR5270, F-69621, France

Department of Electrical and Computer Engineering, Swanson School of Engineering,

University of Pittsburgh, Pittsburgh, PA, U.S.A.

Keywords:

Wearable Sensor, Medical IoT, Heart Rate Sensor.

Abstract:

Recent developments in personal and mobile healthcare have shown promising results in term of patients’ quality

of life and quality of care improvements. This can be achieved through continuous monitoring of patients’

physiological functions using wearable non-invasive biomedical sensors. The remote collection and processing

of such data can then be used to provide rapid medical response if a problem is detected or to offer preventive

measures. However, the integration of wearable sensors into wider-scale framework is still a major challenge,

as real-time data collection and remote conﬁguration capabilities must be integrated to strongly constrained

devices. Here, we show how such requirements can be integrated into a multiparameter, cardiorespiratory

wearable sensor and how this sensor can be integrated into wide-scale Internet-based frameworks. We thus

manufactured a biomedical-grade heart rate, instantaneous heart-rate variability and respiratory sensor. The

sensor was tested in real life ambulatory condition, and we showed an Internet-based proof of concept exhibiting

the integration of our sensor into wide-scale healthcare frameworks. Finally, we anticipate that wearable

healthcare will greatly improve patients’ quality-of-life by using IoT-based wearable devices similar to the

sensor developed in this paper.

1 INTRODUCTION

Personalized and mobile healthcare are growing ﬁelds

of interest in the biomedical community. The idea

behind such concepts is to provide patients with health

recommendations and diagnostic tailored to their indi-

vidual needs. This can be achieved using continuous

remote monitoring of the patients’ physiological func-

tions associated with data analysis to offer preventive

measures, or rapid medical response if a physiologi-

cal malfunction is detected. Consequently, personal-

ized healthcare is a promising approach to improve

the ﬁnancial and therapeutic efﬁciency of healthcare

(Van Hoof and Penders, 2013), by avoiding unneces-

sary hospitalization while preserving the safety of the

patients.

Personalized healthcare can be achieved using

wearable sensors networks (Massot et al., 2013), which

are used to accurately monitor vital signs. Cardiac

health estimation devices are of particular interest be-

cause of the crucial nature of the cardiac function and

because it has been well explored by the biomedi-

cal sensor community, for instance in (Massot et al.,

2015), (Magno et al., 2014), (Altini et al., 2011),

(Khayatzadeh et al., 2013) and (Tuominen et al., 2017).

Moreover, cardiac activity parameters can be used as

an indicator of several pathologies. In particular, heart

rate variability (HRV) provides insights about the au-

tonomic nervous system parameters which reﬂects the

patients’ emotional state (Task Force of the European

Society of Cardiology the North American Society

of Pacing Electrophysiology, 1996). However, one

of the major research problem at the moment is the

integration of developed wearable sensors into bigger

frameworks. Indeed, personalized healthcare assumes

health data to be collected in real-time and that such

data is analyzed by powerful algorithms on distant

servers. Personalized healthcare also makes remote

sensor reconﬁguration necessary in order to better suit

both patients and healthcare professional expectations.

All the identiﬁed wearable electrocardiogram-

based cardiac activity sensors, despite accurately mea-

suring cardiac activity parameters, do no feature

enough connectivity and functionality to be able to

Gatouillat, A., Massot, B., Badr, Y., Sejdi

c, E. and Gehin, C.

Building IoT-Enabled Wearable Medical Devices: An Application to a Wearable, Multiparametric, Cardiorespiratory Sensor.

DOI: 10.5220/0006729101090118

In Proceedings of the 11th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2018) - Volume 1: BIODEVICES, pages 109-118

ISBN: 978-989-758-277-6

109

use them in a personalized healthcare context. Only

wearable ECG or ECG-based cardiac parameters sen-

sors were studied because of their accuracy in terms

of cardiac parameters estimation. Although widely

used, the accuracy of other popular methods such as

plethysmography is still questioned by the community

(Sch

afer and Vagedes, 2013) and suffer from heavy

movement artifact. Authors of (Magno et al., 2014) de-

signed a wearable heart rate and respiratory rate sensor

with Bluetooth low energy and proprietary 802.15.4

connectivity by combining several development kits.

This solution lacks remote conﬁguration capabilities

and proper integration, which both are critical aspects

of personalized healthcare. In (Tuominen et al., 2017),

a remotely conﬁgurable ECG sensor was designed,

but data is stored locally on a SD card, which makes

this solution unsuitable for remote healthcare applica-

tions. Another wearable ECG sensor was developed

by (Izumi et al., 2015), is remotely conﬁgurable, and

data are sent using wireless connectivity. However,

this device use near ﬁeld communication (NFC) for all

of its connectivity. The short range of NFC application

makes it unpractical for real life data collection, as data

can only be collected when a NFC reader is brought in

proximity of the sensor. A heart rate (HR) and HRV

sensor is described in (Massot et al., 2015). This de-

vice offers on-board cardiac parameters calculation,

and computed parameters are sent to a smartphone

using wireless connectivity. However, this sensor is

not remotely conﬁgurable, and data is only stored lo-

cally on the smartphone, which makes some adaptation

necessary in order for this device to be used in a per-

sonalized healthcare concept.

Recently, the study of the interconnection of mul-

tiple devices featuring real-time data collection and

remote conﬁguration capabilities was studied under

the scope of the Internet-of-Things (IoT) (Gubbi et al.,

2013). In the IoT, a variety of heterogeneous com-

municating devices are interconnected, but also com-

municate with external services implemented on re-

mote servers. Consequently, the IoT seems like an

approach of particular interest when trying to solve the

challenges of wide-scale personalized healthcare (Fer-

nandez and Pallis, 2014), and this paper will describe

the integration of IoT characteristics on a biomedical

wearable cardiac and respiratory sensor to enable its

integration into a wide scale personalized healthcare

framework.

Our solution focuses on the integration of remote

reconﬁguration capabilities to our sensor, the imple-

mentation of different functional modes to enable

adaptive capabilities, and the real-time streaming of

biomedical data. All these functionalities must pre-

serve measurements accuracy. The impact of integrat-

ing IoT characteristics to the sensor on both hardware

and software designs, but also on the overall consump-

tion of the device was carefully studied in order to

ensure our sensor validity from a real-life perspective.

The following paper is organized as follows. Sec-

tion 2 introduces the materials and methods enabling

the integration on IoT properties to a biomedical grade

sensor, but also describes the design of analog/digital

electronics to achieve the desired functional and non-

functional goals. In section 3, we present a compre-

hensive caracterization of the sensor from a power

consumption and data quality evaluation perspective,

and ﬁnally section 4 discusses the potential application

of such sensor for patients and medical practitioners.

2 MATERIALS AND METHODS

The manufactured IoT-enabled wearable cardiac and

respiratory activity sensor is displayed in Figure 1.a.

This circuit board is small (

40 mm × 20 mm × 6 mm

and it is packaged along with a

300 mAh

battery in a

plastic enclosure. The complete sensor (i.e., PCB, bat-

tery, plastic enclosure and cables) weighs only 26.7 g,

making it light enough for wearable applications.The

sensor is designed to be used with three electrodes:

left arm, right arm and a common mode rejection elec-

trode, as illustrated in Figure 1.b. The left arm and

right arm electrodes are connected to the sensor us-

ing wires, and the sensor is directly attached to the

common mode rejection electrode. Consequently, the

device can be comfortably worn by the patients for

extended periods of time. The battery of the sensor is

charged using the micro-USB port, and the sensor can

be reset to its default state using the single push-button

of the device. This sensor is able to measure both

the heart rate, heart rate variability parameters and the

respiration waveform (RWF).

2.1 Sensor Hardware Description

The global sensor architecture is given in Figure 2.

Because of the potential overhead due to the addition

of IoT-related characteristic to our device, components

were selected in order to maximize their computational

and power efﬁciency.

In order to minimize CPU load and CPU wake-up

time caused by real-time signal processing, it must

be performed using dedicated hardware. This led to

the choice of a PSoC 5LP (Cypress Semiconductor,

San Jose, CA) for the device microcontroller and ded-

icated signal processor. Indeed, this integrated cir-

cuit (IC) offers both an ARM Cortex Core M3 CPU

BIODEVICES 2018 - 11th International Conference on Biomedical Electronics and Devices

110

40 mm

20 mm

(a) (b)

(c)

Figure 1: Manufactured sensor (a), body placement (b) and

sensor worn in ambulatory conditions (c).

core and a programmable digital ﬁlter block in a sin-

gle package, leading to smaller PCB, thus improving

the comfort of the patients. Because Bluetooth Low

Energy (BLE) is the main protocol used in resource

constrained devices for the IoT and because of its

widespread adoption (Harris III et al., 2016), it was

selected to provide wireless connectivity to our sen-

sor. The BLE113 integrated chip (Silicon Labs, Austin,

TX) was selected as the BLE dedicated microprocessor

because it integrates both a full hardware and software

BLE stack with integrated antenna design in a small

15.75 mm × 9.15 mm × 1.9 mm

package. This inte-

grated circuit is fully programmable, and all commu-

nication dedicated ﬁrmware can be deployed on this

IC, thus reducing the microprocessor computing load.

In order to be able to simultaneously measure both the

ECG and the respiration waveform, the ADS1292R

(Texas Instruments, Dallas, TX) was selected. This

low-power analog front-end (AFE) integrates two dif-

ferential ampliﬁers and two 24 bits analog to digital

converters. It also features a right-leg drive (RLD) am-

pliﬁer to implement common mode rejection (Winter

and Webster, 1983) along with lead-off and respiration

signal modulation and demodulation circuitry. The

measurements are sent using a SPI bus at a conﬁg-

urable sampling rate (from 125 to 8000 samples per

seconds).

In the default state, the analog front end is con-

ﬁgured to measure both the respiration waveform on

channel 1 and the ECG signal on channel 2 at a sam-

Analog front-end

Signal amplification

Respiration measurements

Signal filtering

Analog-to-digital conversion

Bluetooth module

Advertising

Connection/Disconnection handling

Data communication

ADS1292R

BLE113

Disposable ECG

electrodes

Digital bus

Programmable system-on-chip

Hardware ECG processing

Soware parameters computation

IoT state machine

Battery level monitoring

PSoC5LP

Digital bus

Figure 2: Hardware architecture of the sensor.

pling rate of a 1000 SPS using the internal clock of the

IC. This sampling rate enables a

1 ms

resolution for RR

interval measurements. The RLD ampliﬁer and lead-

off detection circuitry are both enabled, along with the

respiration modulation and demodulation modules.

The RR interval computation is implemented us-

ing hardware ﬁltering and peak detection. The digital

ﬁlter block of the PSoC5LP is used to compute the

smoothed moving average derivative of the ECG sig-

nal, which is then used in combination with adaptive

thresholding in order to compute the RR interval. De-

tails about the implementation are given in Figure 3

and in (Massot et al., 2016).

count

capture

reset

comp

value

Up Counter

Sampling CLK

1 kHz

DMA1

Trig.

Int.

Counter value

tranfered to RAM

with DMA1

DFB

Delayed

DFB

Output

Adaptative

Threshold

ECG signal from AFE

transfered with

external DMA

Figure 3: Hardware detection of RR intervals.

Using this hardware-centric approach to perform

computationally intensive real-time signal processing

reduces the CPU load and increases the CPU sleep

time, thus reducing global power consumption. All the

other sensor functional tasks are implemented using a

software approach, and will thus be further described

in the following section.

Building IoT-Enabled Wearable Medical Devices: An Application to a Wearable, Multiparametric, Cardiorespiratory Sensor

111

2.2 Sensor Software Description

The ﬁrmware of the sensor was developed to achieve

several functional and non-functional goals. The ﬁrst

goal is focused on the measurements and their trans-

mission: the real-time communication of the instanta-

neous heart rate, the computation and communication

of HRV parameters every 5 minutes, the real-time com-

munication of 6 respiration samples every second. The

second goal is to transmit non-functional information

about the sensor, typically the battery level or if the

sensor is unattached. Finally, the last goal is the im-

plementation of IoT dedicated software, which is used

to integrate the sensor into a wider IoT-based frame-

work. Such software is used to transmit information

about the sensor state in order to trigger remote sensor

conﬁguration.

2.2.1 Measurements Dedicated Software

This ﬁrmware is implemented on the PSoC5LP, and

it can be divided into three sections: the computation

of the instantaneous heart rate when a heart beat is

detected (i.e., when an interruption is triggered by the

peak detection hardware). The heart rate can be easily

computed using the following equation:

BPM

60 × SR

int.

60 × 1000

int.

(1)

where

BPM

is the instantaneous heart rate in beats

per minutes (BPM),

is the sampling rate in samples

per seconds and

int.

is the RR interval in samples.

This computation is implemented directly on the mi-

crocontroller and the RR interval values are buffered

to be used in the HRV parameters computation.

Every 5 minutes, the HRV parameters are com-

puted from the buffered RR interval values in accor-

dance with the recommendations of the joint Task

Force of the European Society of Cardiology (Task

Force of the European Society of Cardiology the North

American Society of Pacing Electrophysiology, 1996),

and these parameters are summarized in Table 1.

The computation of the frequency domain pa-

rameters are based on the power spectral density

(PSD) estimation. Because the RR intervals signal

is unevenly sampled, the Lomb-Scargle periodogram

(Lomb, 1976) was determined using the fast computa-

tion algorithm developed by (Press and Rybicki, 1989).

We used the same optimal algorithmic parameters than

(Massot et al., 2016). Indeed, authors of (Massot et al.,

2016) conducted a systematic evaluation of the impact

of the parameters on the trade-off between computa-

tion time and periodogram precision, resulting in the

proposition of optimal parameters with respect to both

the periodogram precision and the computation time.

Table 1: Computed HRV parameters

Variable Unit Domain Description

SSDN ms Time Standard deviation of buffered RR

intervals

RMSSD ms Time Quadratic mean of differences be-

tween consecutive RR intervals

LF/HF n.u. Freq. Low-frequency (0.04 to 0.15 Hz)

to high-frequency (0.15 to 0.4 Hz)

components ratio of the PSD of the

buffered intervals

Norm. LF % Freq. Normalized low-frequency compo-

nents to sum of low- and high-

frequency components of the PSD

ration, i.e., LF/(LF+HF)

The respiration waveform is buffered in the RAM

of the PSoC5LP, and is transmitted wirelessly every

second. The ﬁrst step of the respiration signal pro-

cessing is the downsampling of the respiration sig-

nal from 1000 SPS to 6 SPS. A sampling frequency

of 6 Hz, resulting in a Nyquist rate of 3 Hz, is well

within the bandwidth of respiratory signals, which is

at most 1.5 Hz (Zhao et al., 1994), and decreasing the

sampling rate reduces energy consumption because of

the smaller amount of transmitted information. Af-

ter signal downsampling, the respiration waveform is

smoothed using an exponential ﬁlter:

y[n] = (1 − α)y[n − 1] − αx[n] (2)

with

a conﬁgurable ﬁlter parameter. For the prelim-

inary results, this parameter was set to

0.5

. Once the

signal is ﬁltered, the value are communicated to the

Bluetooth IC using a signed 24 bits integer format.

2.2.2 Communication Dedicated Software

In this section, the ﬁrmware of the BLE113 IC will be

detailed. The scripting language provided by Silicon

Labs, BGScript, was used to program this ﬁrmware. A

custom GATT server was implemented with respect

to the heart rate proﬁle detailed in the Bluetooth 4

speciﬁcation: the mandatory generic access proﬁle, de-

vice information, heart rate and battery services were

implemented. The heart-rate service was augmented

with a customized heart rate variability characteris-

tic. This characteristic is of length 5, and the SSDN,

RMSSD, LF/HF ratio and normalized LF components

ratio are encoded as single-byte unsigned integers. An

additional service was implemented for the respiration

along with another service dedicated to IoT-related

information (i.e., information about the sensor non-

functional states).

The BLE113 also implements device advertise-

ment when the sensor is not connected and if a discon-

nection occurs according to the heart rate proﬁle: the

advertising interval is between 20 and 30 ms for the

BIODEVICES 2018 - 11th International Conference on Biomedical Electronics and Devices

112

ﬁrst 30 seconds, and it is increased to be between 1

and 2.5 s for the next 30 seconds. If no device connect

to the sensor after this advertising period, the sensor is

put in a very low-energy deep sleep mode.

2.2.3 IoT Dedicated Software

To enable our respiratory and cardiac activity sensor

with IoT-based characteristics in order to include the

object in wider IoT-based frameworks, the require-

ments are three-fold:

•

The data must be streamed in real time, meaning as

soon as a data is acquired, it must be transmitted.

•

The sensor must be able to provide information

about its functional and non-functional state in

order to be able to build smart-scenarios based on

the IoT-enabled devices.

•

The sensor must be remotely conﬁgurable in or-

der to be integrated to wider-scale auto-adaptation

scenarios.

In order to fulﬁll this set of requirements, the sen-

sor was modeled using a labeled transition system

(LTS). LTSs are widely used models of computation

(MoC) because of their relative simplicity and high

expressiveness.

The LTS of our sensor is given in Figure

. The

transitions between the states are either controllable

(i.e., a remote tier can force the transition) or non-

controllable (i.e., the sensor automatically determines

its state from environmental measurements). As a

convention, a transition written as “

a \ b

” means the

transition is triggered on detection of internal event

or is remotely triggered using event

. When a transi-

tion is only labeled with a single event, it is assumed

that this event is internal.

As displayed in Figure

, the sensor has 5 non-

functional states:

•

When turned on or after a reset operation, the sen-

sor is in the

initialization

state. In this state, the

AFE is conﬁgured with the default parameters and

the BLE113 is software reset. This state automat-

ically calls the normal operation state when the

initialization process is over.

•

The default state is the

normal

state. In this state,

both ECG and respiration signal are acquired. The

instantaneous heart-rate is sent in real time (i.e.,

the new value is transmitted as soon as it is avail-

able), the HRV parameters are computed every 5

minutes, and the respiration signal is packaged and

transmitted every second.

•

The

failsoft

state is an energy saving state. It is

triggered either externally or when the battery level

of the sensor reaches a low level (i.e., 20% in our

case).

•

When the sensor detects a disconnected lead, it

is placed in the

unattached

state. In this state, a

timer is launched, and the sensor is turned off if

the timer overﬂows. If the sensor is reattached, it

goes back to the normal operation state.

•

The

stop

state denotes a very low-energy deep

sleep mode. The sensor can be placed in this state

at any time.

The BLE113 can generate “no connection” (abbre-

viated as “no conn” in the LTS) events if the sensor

does not receive any connection request during the

advertising interval. This causes the sensor to go into

the very low power stop state.

Init.

Normal FailsoUnatt.

Stop

init_done

batt < 20% \ setMode(failso)

batt ≥ 20% \ setMode(normal)

(batt < 5% ∨ no_conn) \ setMode(stop)

no_conn \ setMode(stop)

(timeout ∨ no_conn) \ setMode(stop)

hardware_reset

unattached

Figure 4.

Practically, this LTS is implemented on the ARM

core of the PSoC5LP. Events are either generated in-

ternally or transmitted by the BLE113, and the sensor

state is exposed using a custom service.

2.3 Gateway Software and IoT-Based

Framework

Android phones were chosen as the main gateways

to our sensor. This choice was motivated principally

because phones can be carried by the subjects, result-

ing in a mobile gateway with perpetual connection to

the Internet. Consequently, an Android application

was developed to connect to the sensor, but also to

plot the data, store it locally on the phone or trans-

mit them over the Internet. The application is based

on the nRF Toolbox open-source application

devel-

oped by Nordic Semiconductor (Oslo, Norway), and

screenshots of the application are given in Figure 5.

Because data is collected in real time and sensor

conﬁguration can be triggered at any moment, a pub-

lish/subscribe type protocol was chosen. Practically,

https://github.com/NordicSemiconductor/Android-

nRF-Toolbox

Building IoT-Enabled Wearable Medical Devices: An Application to a Wearable, Multiparametric, Cardiorespiratory Sensor

113

Figure 5: Companion Android application.

the Eclipse Paho MQTT client was integrated in the

application. MQTT is a lightweight publish/subscribe

messaging Internet-based protocol. It was designed to

be used in memory and bandwidth constrained devices.

This protocol is built around the notion of topics. A

topic is an URI-like string used to describe transiting

data. In the MQTT terminology, a publisher sends data

in a topic. If a subscriber subscribes to the same topic,

it will receive the data instantaneously. Formally, all

data is sent through a central broker that can be seen as

a hub distributing data between the relevant publishers

and subscribers. This publish/subscribe mechanism

thus implements real-time data streaming over the In-

ternet, and enables connection to various external Web

services or other connected devices.

3 RESULTS

3.1 Power Consumption

Characterization

In this section, a comprehensive evaluation of the sen-

sor power consumption with respect to connection

parameters and non-functional IoT-state deﬁned in Fig-

ure 4 is performed.

Current consumption measurements were per-

formed using a Keithley 2400 sourcemeter (Beaverton,

OR, USA). The ﬁrst characterization to be performed

was a static power consumption evaluation of the nor-

mal, failsoft, and stop states. The initial state was not

measured because it is only a transitional state with a

short duration, and the unattached state was not mea-

sured because the sensor is in the same hardware con-

ﬁguration than in the normal state, resulting in equal

power consumption. This static characterization was

performed using default Android 7.0 BLE connection

parameters (i.e., connection interval of

48.75 ms

, time-

out of

20 s

and latency of 0). The consumption results

are given in Table 2. With a power consumption of

10.53 mW in the normal state, the battery life of the

sensor if of about 75 hours using a 300 mAh battery.

This battery life is extended to about 85 hours if the

failsoft mode is used. In the stop mode, the sensor can

last more than two months on a fully charged battery.

Table 2: Static power characterization of IoT states.

IoT state Power consumption (mW)

Normal 10.53

Failsoft 9.18

Stop 0.505

In order to further optimize power consumption

of the sensor, the inﬂuence of the BLE connection

parameters were also studied. There are three BLE

connection parameters: interval, latency and timeout.

The connection interval designates the period between

two master’s requests to the slave, while the latency

deﬁnes the number of connection intervals that can

be ignored safely by the slave. Finally, the connec-

tion timeout designates the period after which the BLE

master will consider that the connection with the slave

is lost. It is only after the timeout period that the BLE

master can attempt with reconnection. For the sensor,

we ﬁxed both the connection interval and connection

timeout, and the impact of the latency was studied on

the overall power consumption. The results of this

characterization are given in Figure 6, and the con-

sumption of both the normal IoT state and the failsoft

IoT state are plotted against the latency for various

connection interval and connection timeout. For all

states and connection interval and timeout parame-

ters, and increase in the latency results in decreasing

overall power consumption, with a stronger decrease

for smaller latency values. It is worth noting that for

a latency of 8, connection instability was observed.

In order to enable smaller power consumption, while

keeping acceptable connectivity quality, the sensor

requests a latency of 2 when a new connection is es-

tablished. However, masters (such as Android phones)

can reject connection parameters update and force a

new set of less constrained connection parameters (i.e.,

smaller connection intervals, zero latency and smaller

connection timeout), causing an increase in the overall

power consumption. This problem was solved with a

slight oversizing the battery capacity in order to always

achieve at least 48 hours of battery life.

BIODEVICES 2018 - 11th International Conference on Biomedical Electronics and Devices

114

0 1 2 3 4 5 6 7 8

Latency

6.5

7.0

7.5

8.0

8.5

Sensor power consumption (mW)

Normal mode, interval = 67.5 ms, timeout = 1.0 s

Failsoft mode, interval = 67.5 ms, timeout = 1.0 s

Failsoft mode, interval = 742.5 ms, timeout = 10.0 s

Figure 6: Characterization of the BLE connection latency

parameter effect on the overall power consumption.

3.2 Data Acquisition

In this section, various data acquisition are presented.

This ﬁrst experiment to be performed was the valida-

tion of RR-interval acquisition using synthetic ECG

signals generated using an Agilent 33220A (Santa

Clara, CA, USA) arbitrary waveform generator. The

objective of this test is twofold: it validates the com-

putation of RR-interval and the reliability of the BLE

connection. In order to simulate heart rate variability,

triangular frequency modulation was applied to the

synthetic ECG signal:

base frequency = 1.25 Hz

frequency deviation = 200 mHz

(3)

which theoretically results in RR intervals ranging

between:

int

1000

1.25 − 0.2

≈ 952.4 ms

and, RR

int

1000

1.25 + 0.2

≈ 689.7 ms

(4)

The sensor was then connected to a Raspberry Pi

used as a BLE data logger to record RR-intervals and

HRV parameters, and an excerpt of the collected RR

interval values is given in Figure 7. The experiment

was performed during 1 day, 17 hours and 9 minutes.

A total of 185153 were logged. Data was analyzed

for records violating the theoretical limits (with a tol-

erance of 1% over the limits), and no violations were

found.

The next experiment was data acquisition during

short respiratory exercises. During this exercise, a

young healthy patient was asked to perform a series

of forced inspiration and expiration during 1 minute

in order to trigger sinus arrhythmia. Results from

this experiment are given in Figure 8, where both the

instantaneous heart rate and the normalized respiratory

waveform are plotted. This ﬁgure clearly illustrates

sinus arrhythmia, with increasing heart rate during

228900 228950 229000 229050 229100 229150 229200

time (s)

700

750

800

850

900

950

RR Interval (ms)

20.0 s ⇔ 50 mHz

Figure 7: RR interval determined from synthetic ECG signal.

inspiration because of a decreasing vagal tone and

decreasing during expiration.

600

700

800

RR Interval (ms)

60 70 80 90 100 110 120

time (s)

0.0

0.2

0.4

0.6

0.8

1.0

Normalized RWF

Figure 8: RR-Interval and respiratory waveform during con-

trolled respiratory exercise.

The ﬁnal experiment consisted on a short-term am-

bulatory testing of the sensor. For this experiment, a

patient was asked to wear the sensor during one hour,

while performing normal daily activities. Cardiorespi-

ratory data was recorded using the companion Android

application, and results from this experiment are given

in Figures 9 and 10. Figure 9 displays both the instan-

taneous RR interval and normalized respiration signals.

No measurement artifact were observed for both the

signals. ECG artifacts, resulting in RR interval arti-

facts, are typically caused by electrical muscular activ-

ity interfering with ECG signals. Such adverse effects

were minimized by placing on the lower part of the rib

cage. Indeed, this location features minimal muscle

thickness, thus reducing interactions between ECG sig-

nal and electrical muscular activity. It is worth noting

that the respiration signal features a slowly evolving

component. This is because the respiratory activity is

in fact an impedance measurement, and because the

impedance of the electrode-skin interface can slowly

evolve over time because of various factors (such as

humidity, the presence of sweat, etc.). This can be

corrected by applying a high-pass ﬁlter during signal

post-processing.

Building IoT-Enabled Wearable Medical Devices: An Application to a Wearable, Multiparametric, Cardiorespiratory Sensor

115

400

600

800

1000

RR Interval (ms)

0 500 1000 1500 2000 2500 3000 3500

time (s)

0.0

0.2

0.4

0.6

0.8

1.0

Normalized RWF

Figure 9: Real-life RR-Interval and respiratory waveform.

Figure 10 displays the 4 heart rate variability pa-

rameters plotted as a function of time. These HRV

parameters correspond to the HR measurements ex-

hibited in Figure 9, and fall within the typical ranges

deﬁned by the literature (Nunan et al., 2010).

SDNN (ms)

RMSSD (ms)

0 1000 2000 3000 4000 5000

time (s)

LF n.u. (%)

0 1000 2000 3000 4000 5000

time (s)

LF/HF

Figure 10: Real-life HRV parameters.

The preliminary results presented herein-above

proves that our wearable sensor provides high quality

data. The next experiment will deal with the integra-

tion of our wearable sensor in wider-scale IoT-based

healthcare framework.

3.3 Integration to an IoT-Based

Framework

As speciﬁed in Section 2, the integration of our wear-

able sensor to wider-scale healthcare frameworks oc-

curs through the Android companion application. In-

deed, upon sensor connection, the application attempts

to connect to a local MQTT broker. If the connection

is successful, the application will publish heart rate

and HRV values as soon as they are available on two

topics:

• interface/hr

• interface/hrv

As a result, cardiac parameters are streamed in real

time using these two topics, and this is the basis to

build wider-scale healthcare framework. As a proof

of concept, a monitoring Web-based graphical user

interface was implemented. It simulates the kind of

GUI that can be made available to physicians in order

to have real time health information on their patients.

A sample screenshot of the Web-based GUI is given

in Figure 11.

Figure 11: Web application visual.

4 DISCUSSION

In this paper, we introduced a wearable cardiorespira-

tory sensor that can be easily integrated to wider-scale

healthcare framework. This sensor satisﬁes crucial

healthcare-related requirements such as data realia-

bility, remote conﬁguration capabilities or scalability.

Wearable healthcare and the Internet-of-Things have

similar objectives: the use of a mass of connected ob-

jects (which also are worn by patients in the context

of wearable healthcare) that monitor physical parame-

ters in order to trigger medical advice or intervention

based on data analysis. For instance, our wearable sen-

sor can be integrated in a living-lab, where the living

environment of the patient is continually monitored.

This integration relies on the use of widely adopted

technologies such as BLE, but also on Internet-based

technologies such as MQTT in order to enable wide-

scale connectivity. Internet-based connectivity implies

that our wearable sensor is able to connected to Inter-

net healthcare services, which can perform advanced

data analysis and detect potential health crisis of the

patient and trigger urgent medical response.

Wearable healthcare have a strong potential on the

improvement of the patients’ quality of life and quality

of care. Indeed, the wearable nature of devices en-

able patients to be continuously remotely monitored.

Combined with recent advances in home-automation,

BIODEVICES 2018 - 11th International Conference on Biomedical Electronics and Devices

116

wearable healthcare will drastically improve in-home

care for a variety of patients. Collected data processed

using big-data techniques or advanced signal process-

ing could also have great predictive values for the

evolution of chronic diseases and could be used to

provide better and earlier care for patients.

However, the personal nature of collected health

data mandates strong security mechanism, which were

not explored in our contribution. Indeed, at the mo-

ment our system only features basic data encryption

in compliance with the BLE standard, and the MQTT-

based solution does not use any security mechanism.

Security of our overall system needs to be improved,

more particularly in terms of access control (i.e., the

patients of physicians must be able to know and con-

trol who accesses their medical data) and identity

management, and recently developed decentralized

blockchain-based solutions (Zhu et al., 2017) can be

explored in order to provide comprehensive werable

healthcare system security.

5 CONCLUSION

In this paper, we presented a multiparametric, car-

diorespiratory wearable sensor. In order to answer

healthcare requirements, more particularly in terms of

the ability to integrate wearable sensors into wider-

scale frameworks, we considered an IoT-based ap-

proach. Indeed, we equipped our sensor with remote

conﬁguration capabilities while preserving quality-of-

data and real-time streaming capabilities, which are

key requirements of wearable healthcare systems. This

sensor was implemented using carefully selected hard-

ware, and it was comprehensively characterized in

terms of energy consumption, which is another major

concern of wearable healthcare devices. Indeed, be-

cause battery charging usually implies the sensor is

not collecting physiological data, potentially relevant

data can be lost, and the charging time to battery life

ration must thus be as big as possible. The data col-

lection capabilities of our sensor were also extensively

tested on both synthetic ECG signals and in real-life

ambulatory conditions. Successful testing and inte-

gration to Internet-based framework proved that our

sensor can be used in a wide-scale wearable healthcare

framework.

ACKNOWLEDGMENT

The authors would like to thank the COOPERA fund-

ing program of R

egion Auvergne Rh

one-Alpes for

their generous ﬁnancial support.

REFERENCES

Altini, M., Polito, S., Penders, J., Kim, H., Van Helleputte,

N., Kim, S., and Yazicioglu, F. (2011). An ECG patch

combining a customized ultra-low-power ECG SoC

with bluetooth low energy for long term ambulatory

monitoring. In Proceedings of the 2nd Conference

on Wireless Health, pages 15:1–15:2, New York, NY,

USA. ACM.

Fernandez, F. and Pallis, G. C. (2014). Opportunities and

challenges of the internet of things for healthcare: Sys-

tems engineering perspective. In Proceedings of the

EAI International Conference on Wireless Mobile Com-

munication and Healthcare, pages 263–266.

Gubbi, J., Buyya, R., Marusic, S., and Palaniswami, M.

(2013). Internet of things (IoT): A vision, architectural

elements, and future directions. Future Generation

Computer Systems, 29(7):1645 – 1660.

Harris III, A. F., Khanna, V., Tuncay, G., Want, R., and

Kravets, R. (2016). Bluetooth low energy in dense

IoT environments. IEEE Communications Magazine,

54(12):30–36.

Izumi, S., Yamashita, K., Nakano, M., Kawaguchi, H.,

Kimura, H., Marumoto, K., Fuchikami, T., Fujimori,

Y., Nakajima, H., Shiga, T., and Yoshimoto, M. (2015).

A wearable healthcare system with a 13.7

A noise tol-

erant ecg processor. IEEE Transactions on Biomedical

Circuits and Systems, 9(5):733–742.

Khayatzadeh, M., Zhang, X., Tan, J., Liew, W. S., and Lian,

Y. (2013). A 0.7-V 17.4-

W 3-lead wireless ECG

SoC. IEEE Transactions on Biomedical Circuits and

Systems, 7(5):583–592.

Lomb, N. R. (1976). Least-squares frequency analysis of

unequally spaced data. Astrophysics and Space Science,

39(2):447–462.

Magno, M., Spagnol, C., Benini, L., and Popovici, E.

(2014). A low power wireless node for contact and

contactless heart monitoring. Microelectronics Journal,

45(12):1656–1664.

Massot, B., Noury, N., Gehin, C., and McAdams, E. (2013).

On designing an ubiquitous sensor network for health

monitoring. In Proceedings of the International Con-

ference on e-Health Networking, Applications and Ser-

vices, pages 310–314.

Massot, B., Risset, T., Michelet, G., and McAdams, E.

(2015). A wireless, low-power, smart sensor of cardiac

activity for clinical remote monitoring. In Proceedings

of the International Conference on E-health Network-

ing, Application Services, pages 488–494.

Massot, B., Risset, T., Michelet, G., and McAdams, E.

(2016). Mixed hardware and software embedded sig-

nal processing methods for in-situ analysis of cardiac

activity. In Proceedings of the 9th International Joint

Conference on Biomedical Engineering Systems and

Technologies, pages 303–310.

Nunan, D., Sandercock, G. R., and Brodie, D. A. (2010).

A quantitative systematic review of normal values

for short-term heart rate variability in healthy adults.

Pacing and Clinical Electrophysiology, 33(11):1407–

1417.

Building IoT-Enabled Wearable Medical Devices: An Application to a Wearable, Multiparametric, Cardiorespiratory Sensor

117

Press, W. H. and Rybicki, G. B. (1989). Fast algorithm

for spectral analysis of unevenly sampled data. The

Astrophysical Journal, 338:277–280.

Sch

afer, A. and Vagedes, J. (2013). How accurate is pulse

rate variability as an estimate of heart rate variability?:

A review on studies comparing photoplethysmographic

technology with an electrocardiogram. International

Journal of Cardiology, 166(1):15 – 29.

Task Force of the European Society of Cardiology the North

American Society of Pacing Electrophysiology (1996).

Heart rate variability: Standards of measurement, phys-

iological interpretation, and clinical use. Circulation,

93(5):1043–1065.

Tuominen, J., Lehtonen, E., Tadi, M. J., Koskinen, J., Pnkl,

M., and Koivisto, T. (2017). A miniaturized low power

biomedical sensor node for clinical research and long

term monitoring of cardiovascular signals. In Proceed-

ings of the International Symposium on Circuits and

Systems, pages 1–4.

Van Hoof, C. and Penders, J. (2013). Addressing the health-

care cost dilemma by managing health instead of man-

aging illness: An opportunity for wearable wireless

sensors. In Proceedings of the Conference on De-

sign, Automation and Test in Europe, pages 1537–1539.

EDA Consortium.

Winter, B. B. and Webster, J. G. (1983). Driven-right-leg

circuit design. IEEE Transactions on Biomedical Engi-

neering, 30(1):62–66.

Zhao, L., Reisman, S., and Findley, T. (1994). Respiration

derived from the electrocardiogram during heart rate

variability studies. In Proceedings of 16th International

Conference of the IEEE Engineering in Medicine and

Biology Society, pages 123–124.

Zhu, X., Badr, Y., Pacheco, J., and Hariri, S. (2017). Auto-

nomic identity framework for the internet of things. In

Proceedings of the International Conference on Cloud

and Autonomic Computing, pages 69–79.

BIODEVICES 2018 - 11th International Conference on Biomedical Electronics and Devices

118