Real-Time Biosignal Acquisition and Telemedicine

Platform for AAL based on Android OS

Hugo Silva

1,2,3

, Andr´e Lourenc¸o

1,2,4

and Nuno Paz

3,4

Instituto de Telecomunicac¸ ˜oes, Lisboa, Portugal

Instituto Superior T´ecnico, Lisboa, Portugal

PLUX Engenharia de Biosensores, Lisboa, Portugal

Instituto Superior de Engenharia de Lisboa, Lisboa, Portugal

Abstract. Among other strategic action points, the Ambient Assisted Living

Joint Programme framework has deﬁned telemedicine and remote monitoring as

major applications for improved quality of care. With the advent of novel, game-

changing mobile platforms, the technological basis is now in place to provide

more capable and usable systems for this practice. In this paper, we present a

real-time biosignal acquisition and telemedicine platform based on Android OS,

as both the system and the available handsets present highly desirable features

for the ﬁeld. The proposed platform was used in the context of a continuous real

time monitoring of ECG signal showing one simple example of its applicability.

1 Introduction

Telemedicine and mobile monitoring of health, and other related parameters, are re-

ferred as one of the major applications toward improved quality of care and independent

living in an Ambient Assisted Living (AAL) framework [1]. This is a rapidly growing

ﬁeld within clinical medicine, and modern Information and Communication Technolo-

gies (ICTs) are providing new opportunities, through the introduction of innovative

ways of monitoring wellbeing and conveying information both to healthcare profes-

sionals and users [2].

Although the history of telemedicine can be dated back to the Middle Ages, where

bonﬁres were used as a way of transmitting information about the bubonic plague

throughout Europe, only recently, with the advent of wireless and mobile communi-

cation technologies, novel tools have been made available that are able to provide more

effective remote monitoring and telemedicine methods and systems [3].

These systems are useful to assist the general population, however they are particu-

larly important in severe cases where the fact that the patient is unattended among visits

to the practitioner and/or moving the patient between his/her home and the hospital or

healthcare centers represents a great ordeal; Chronic Obstructive Pulmonary Disease

(COPD) and Amyotrophic Lateral Sclerosis (ALS) are just a few examples of such

cases [4,5].

Silva H., Lourenço A. and Paz N..

Real-time Biosignal Acquisition and Telemedicine Platform for AAL based on Android OS.

DOI: 10.5220/0003303801110121

In Proceedings of the 1st International Living Usability Lab Workshop on AAL Latest Solutions, Trends and Applications (AAL-2011), pages 111-121

ISBN: 978-989-8425-39-3

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

Smartphones in particular, are becoming a de facto tool for telemedicine appli-

cations as they evolve to become always-on, always-connected devices [6], with in-

creasingly higher computing capabilities and intuitive user interfaces [7]. Together with

biosignal acquisition and sensing devices, current ICTs have the potential of enabling

practitioners to remotely access a comprehensive set of follow-up and diagnosis param-

eters.

In this paper we present a telemedicine platform for real-time biosignal acquisition,

display, and transmission based on Android OS. Within the existing mobile platforms

landscape, Android OS is gaining particular attraction due to its hardware interoper-

ability, versatility, and usability characteristics. The proposed platform was devised as

a tool for ambulatory and/or home-care within an AAL framework, providing: a) in-

terface with measurement instrumentation through Bluetooth; b) local data acquisition

and visualization; c) local data compression and storage; and d) data transmission to a

remote server over TCP/IP.

The rest of the paper is organized as follows: Section 2 provides an insight on the

motivation behind the use of Android OS; Section 3 describes the related work on this

topic; Section 4 details the proposed platform; Section 5 describes an application of

this platform to the remote monitoring of hear rate data; and ﬁnally, 6 outline the main

results and discussion.

2 Mobile Platforms

2.1 Landscape

A wide variety of mobile development platforms is currently available, of which Sym-

bian, Blackberry, iOS, Android, J2ME, and Windows Mobile, currently represent the

mainstream. Among these, the ﬁrst four have the most representative market shares [8];

each has its own singularities, namely:

a) Symbian is currently the most widespread platform, with a share above 40% of

the smartphone market today. Despite the fact that it provides full hardware access

and that applications are programmed in a variant of the highly efﬁcient C++ lan-

guage, Symbian lacks a uniﬁed structure across editions. Among other aspects, this

forces applications to be speciﬁcally written for each user interface, and even binary

installers need to be revised, as different editions largely differ between them;

b) Blackberry devices are mostly know by their e-mail capabilities. Although the OS

and applications are mostly supported in Java, and data transmission is very well

supported, RIM’s proprietary APIs must be used in order to take advantage of the

device capabilities; this severely limits the developed applications to Blackberry

handsets. Furthermore, this platform was initially developed for keyboard-based

operation, which makes user interaction slightly below that provided by platforms

designed taking into account touchscreen devices;

c) iOS is featured in what can be considered the most desired consumer electronics

devices in the market such as the iPhone, iPad, and iPod. Recognized as the best

user experience in its class and with great computing capabilities, the iOS has a

severe handicap development-wise as application development is subject to Apple’s

112

approval and the standard set of APIs does not allow low level access to the device

resources (e.g. Bluetooth);

d) Android was developed by Google and it is available under an open source license.

Speciﬁcally designed for touchscreen devices, this platform is based on a modi-

ﬁed version of the Linux kernel. While the Android platform is built upon Java

applications, it does not use the established Java standards as J2SE or J2ME, this

being the main drawback of Android as it limits the compatibility between existing

applications written for these platforms and the Android OS;

e) J2ME debuted as a subset of the Java language, with a reduced set of APIs targeted

at mobile phones and other devices that implemented the Virtual Machine. This

platform is available on nearly every handset and application development follows

the philosophy of write once, run anywhere making portability a keystone of this

platform. Nonetheless, the J2ME APIs were initially designed for connection lim-

ited devices, with small displays and reduced computing power, therefore limiting

the developed applications capabilities. Furthermore, applications are bound to the

VM, which greatly conditions the access to core hardware features;

f) Windows Mobile belongs to the ﬁrst wave of mobile platforms for smartphones

ever to appear; designed to mimic in some extent the environment of the desktop

versions of Windows. Although highly robust and versatile, since applications are

programmed in C++ and the APIs provide full hardware access, Windows Mobile

has not been able to adequately keep up with the competition regarding usability,

and changes among versions of the OS generally do not take backward compatibil-

ity into account.

2.2 Outline

In all platforms, except Android and Windows Mobile, applications must either be dig-

itally signed or undergo a thorough approval process by the manufacturer prior to their

deployment or for being made available to end-user. Still, after these processes, the ac-

cess to core resources is typically restrained or sand-boxed, limiting the functionality

of the applications and/or requiring a great deal of attention from the end-user to grant

permission in the operations considered most sensitive in terms of resource access. In

either case, the user experience, functionality, and attainable automation levels are, in

general, highly restrained.

Android however, is an open source platform, currently supported by a wide range

of devices, it is hardware independent, and greatly focused on interaction via touch

reducing the learning curves. As of 2010, over 70 devices from different makers can be

found on the market with a wide price range and most of them with large touchscreen

displays; these allow users with different cognitive and motor skills to use the device as

screens can be easily adapted to the end-users needs.

Given this scenario, our research found the Android platform to be the most versatile

in an AAL framework. Among other advantages regarding its competition, Android

provides: a) Advanced UI with touch-driven interaction; b) Portability; c) Exposed

APIs for Bluetooth radio access; d) Unrestrained hardware access through uniﬁed APIs

to access the different services; e) Multihoming overWiFi or cellular network (2G/3G);

and f) Unprecedented security and application permissions management policy.

113

3 Related Work

Other authors have proposed approaches for telemonitoring and real-time signal ac-

quisition from biosignal measurement devices through mobile platforms. [9] proposes

telemedicine system for home care and patient monitoring based on a mobile phone.

In their approach, the mobile phone collects the data from existing monitors through a

cabled RS232 interface and then sends them at distance through TCP/IP for a central

computer. Furthermore, the client application does not provide local feedback to the

user and is based on Java MIDP, initially designed for devices with reduced interface

capabilities and with severe security restrictions for hardware resources access, which

in the current landscape may not be the most adequate platform in a AAL framework.

In a work by [10], a GSM-based, mobile telemedicine platform for blood pressure

monitoring is presented. This system is based on a wireless platform, that integrates

a Bluetooth receiver and a GSM module; the Bluetooth connectivity allows the wire-

less data acquisition from a Bluetooth-enabled sphygmomanometer, while the GSM

module provides the cellular network access layer for data transmission to a remote

server. While dedicated hardware devices can be highly optimized for power and efﬁ-

ciency, they’re scalability and versatility or AAL applications is generally highly lim-

ited. They perform well the functions for which they were designed, but new features

require ﬁrmware or hardware changes which are difﬁcult to implement on new devices

and to propagate among devices already in use.

A remote patient monitoring system using a Java-enabled 3G mobile phone is pre-

sented in [11]. Designed for bedridden patients, this system allows doctors to remotely

monitor their patients through a mobile phone; patients are connected to standard bed-

side monitors, which in turn relay the data to a remote information server through a

central monitoring station. Through a 3G mobile phone, doctors can use a Java applica-

tion to access the remote informationserver and monitor at distance multiple parameters

from their patients. For continuous monitoring in an AAL framework, devices should

perform their measurement functions without conditioning the individuals throughout

their regular activities and daily life, in which case, the fact that it relies on bedside

monitors limits the practical applications of this system.

[12] have devised a multipurpose mobile platform for ubiquitous personal health

information assessment, that can be used for a wide range of applications. It is an open

source framework that includes a set of interfaces for communication with devices,

storage, network connectivity and user interface. The mobile application is able to col-

lect data from Bluetooth end-devices, provide local feedback with summary data, and

stream it through 3G/GPRS network to a remote server. Although extremely complete

and modular, this platform is based in the Java MIDP framework.

Recent work by [13] has described a smartphone based telemedicine system for

recording limited lead body surface potential maps by means of a Bluetooth-enabled,

12-lead ECG end-device. A software application developed for Windows Mobile and

the .NET compact framework collects the data in real-time from the end-device and

stores it on a removable solid state memory card. Among all smartphone platforms,

Windows Mobile is currently among the group with lowest market shares; moreover,

most handsets designed for this OS require a stylus for proper operation as the UI was

114

not designed to take advantage of the modern devices features. For AAL applications

more user-friendly and intuitive platforms are required.

Previous approaches to this problem have explored different solutions and plat-

forms. We address this problem in a novel approach, through the Android OS mobile

platform and the integrated capabilities provided by modern smartphones, whose ad-

vantages were previously commented. In its current version, our system interfaces with

a general purpose biosignal acquisition hardware platform, nonetheless, we devised

a software-based architecture with plugin end-device and processing blocks. Further-

more, combined with high-level facilities provided by the OS such as the Android Mar-

ket, the proposed platform presents increased versatility for AAL applications since it

can be remotely updated with new features.

4 Proposed Approach

4.1 System Overview

Figure 1 presents an overview of the global architecture for the proposed approach.

Sensors (A) are applied to the subject with the purpose of simultaneously monitoring

one or more parameters. An end-device (B), the acquisition unit, performs the analog-

to-digital conversion and data encoding of the biosignals coming from the different

sensors. Furthermore, the end-device (B) provides Bluetooth wireless connectivity to

the smartphone (C) based on the Serial Port Proﬁle (SPP).

The data streamed by the end-device (B) is continuously collected and decoded in

real-time on the smartphone (C) which performs storage, instant display, and trans-

mission to a remote database (D) where the data becomes available for observation by

healthcare professional and other designated caregivers. In the smartphone, algorithmic

transformations are also performed over the signals, which include data compression

(based on deﬂate algorithm), and signal processing operations consisting of transfer

functions computation and automated biosignal modeling.

Local real-time monitoring of incoming data is shown in the smartphone (C) screen

to provide local feedback to the user, based on our UI screen architecture; also, the

data is stored in standard ASCII format on the local memory. When a connection to

the remote database (D) is available, the stored data is then compressed for network

payload optimization, and streamed in real-time to the remote database (D) using an

atomic transaction model.

For data transmission at distance, our platform uses the multihoming capabilities

of Android OS. A connection to the remote database (D) for data transmission is per-

formed through standard TCP/IP protocol over an established WiFi, 2G/3G cellular

network connection. Since current devices and associated mobile providers data plans

are designed in an always-on, always-connected approach, our platform is able to pro-

vide near-real time continuously updated data at the remote server. The remote server

(D) is able to handle incoming concurrent connection from multiple smartphones.

115

Fig.1. Global system architecture overview.

4.2 Framework

Our framework provides a comprehensive set building blocks for the development and

deployment of ﬂexible and versatile telemonitoring tools. We provide core components

and interfaces for controlling the mobile application at different levels:

a) End-device Communications, through a communications component that is re-

sponsible for establishing and managing the communication with the acquisition

unit over bluetooth;

b) Signal Acquisition, through an end-device dependent component, that decodes the

collected signals; an interface enables the abstraction of the acquisition unit and

addition of complementary or alternative devices;

c) Signal Processing, through a component that implements the conversion algo-

rithms, taking the signals coming from the device as input and providing the results

from the processing to the presentation layer;

d) Data Presentation, which is the core component of the framework; it receives

data from the acquisition and processing components, stores them persistently in

the solid state (SD) memory card, and enables the visualization of the the selected

channel in the user display;

e) Data Compression, which is performed using the components with the same name;

for compression we the selected the deﬂate algorithm (used by gzip), that uses a

combination of the LZ77 algorithm and Huffman coding [14].

f) Transmission Over TCP/IP, through a component that takes the compressed data

and transmits it to a remote server.

In ﬁgure 2 is illustrated the class diagram, where: IDevice is the interface that de-

ﬁnes the functionality for communication with the end-devices and signal acquisition;

116

Fig.2. Simpliﬁed class diagram.

IFrames derived objects store each collected data sample; IConvertionAlgorithm is the

interface that deﬁnes the structure for signal processing blocks; DataServer is the object

responsible for the data presentation; IStorage is an interface that deﬁnes the SD card

data storage procedures; ICompressData is the interface that deﬁnes the data compres-

sion mechanisms; and ﬁnally IConnection and IRemoteConnection are the interfaces

that deﬁne the data communication and associated mechanisms.

4.3 Mobile Application

The mobile application requires Android OS version 2.0 or above, a handset with SD

card storage capabilities, and WiFi or 2G/3G cellular network connectivity. It has been

deployed and tested in the HTC Hero, Google Nexus One, and Samsung SPICA hand-

set; the application runs in all devices, and has been tested successfully for all features

in the ﬁrst two devices.In the case of Samsung SPICA, an unresolved ﬁrmware problem

with the Bluetooth interface prevents the application from executing with all features,

namely the connection to Bluetooth wireless end-devices.

The developed application (see next section) has a footprint of 59KB.The generated

uncompressed ASCII ﬁles with raw data that are stored in the SD card have a 199 byte

header overhead, and each sample takes up 50 bytes of physical memory; Figure 3

117

illustrates an example of the ﬁle structure with header and data. Through the employed

compression algorithm we are able to achieve an average compression ratio (size of

original/size of compressed ﬁle) of 6 in the ﬁle size.

Fig.3. ASCII ﬁle example.

For the mobile client, the user interface takes advantage of the advanced features

provided by the Android platform. Figure 4 depicts several conﬁguration screens and

an oscilloscope-like screen for real-time signal visualization. The ﬁrst is used to con-

ﬁgure the biosignal acquisition parameters by selecting what is the set of channels,

sampling rate, and resolution that should be used for signal acquisition; the second, for

the remote server conﬁguration; and the later, allows the monitored channel selection

and besides the raw signal is able to provide instant on-screen measurements of the

computed parameters for local feedback to the user, in this case the Heart Rate of the

subject in beats per minute(BPM).

(a) Channel selection

screen

(b) Remote Server screen (c) Real-time visualization

screens

Fig.4. User interface examples.

118

5 Case Study

5.1 Scenario

For testing purposes, a case study was devised for remote ECG and Heart Rate mon-

itoring based on the presented framework. As end-device we have used a bioPLUX

research generic biosignal acquisition unit, together with an ecgPLUX ECG triode sen-

sor. Table 1 describes the main speciﬁcations of this system. Attached to this unit is an

ecgPLUX active ECG triode, and its speciﬁcations are listed in Table 2. The interface

with the skin is done through pre-gelled, self-adhesive round AgCl triode electrodes.

Table 1. bioPLUX research speciﬁcations.

Connectivity Bluetooth Class II

Sampling Rate 1000Hz

Channels 8 An. + 1 Dig.

Size 84x53x18mm

Weight 86g

Table 2. ecgPLUX sensor speciﬁcations.

Gain 1000

Filtering 0.05-30Hz

CMRR 110dB

Input Impedance >1MOhm

5.2 ECG Signal Processing

A signal processing block was implemented to analyze the raw signal in real-time and

automatically compute the corresponding Heart Rate (HR) value.

The algorithm used for this propose is based on the determination of the R peaks

of the ECG signal. We followed one of the proposed approaches of [15], based on

amplitude and ﬁrst derivative of the signal.

The ECG signal, x[n], is rectiﬁed, and passed through a low level clipper:

[n] = |x[n]|, (1)

[n] = y

[n] if y

[n] ≥ A

[n] = A if y

[n] < A (2)

The amplitude threshold, A, is determined as a fraction of the maximum amplitude

of the signal in the analyzed window:

A = 0.4max[x[n]], (3)

119

Finally, the ﬁrst derivative is calculated:

[n] = y

[n + 1] − y

[n − 1], (4)

A R peak candidate occurs when a point in y

[n] exceeds a ﬁxed constant threshold.

The application was deployed in an HTC Hero handset enabling the continuous

real-time monitoring of the ECG signal.

6 Conclusions

In an aging society, where life expectancy is increasing and birth rate is decreasing,

there is a clear effort of addressing this problem with tools and technologies capable of

helping in the self-management and delivery of improved healthcare services to peo-

ple. The Ambient Assisted Living Joint Programme points telemedicine as a strategic

application to address the needs of this growing market.

Other platforms have been proposed in the past that either use dedicated hardware

or conventional mobile platforms that, by design, are limitations and high permissions

restrictions that are preventing the faster evolution of telemedicine systems in terms

of functionality and usability. In this paper we have presented a real-time biosignal

acquisition and telemonitoring platform for AAL based on Android OS.

This paper presents an approach which takes advantage of the novel capabilities of

that mobile platform and of the compliant handsets. We’ve provided a brief discussion

on the current mobile platform landscape, and why Android OS is ahead of the compe-

tition in this ﬁeld, we’ve described our proposed approach, and ﬁnally a case study is

described where this framework was applied to the continuous real-time monitoring of

ECG signals.

Preliminary results have shown the validity of this proposed approach for the in-

tended application due to the always-on, always-connected philosophy of modern de-

vices and to the capabilities introduced by Android OS as uniﬁed APIs, unrestricted

hardware access, and unprecedented security management policy.

References

1. Steg, H., Strese, H., Loroff, C., Hull, J., Schimdt, S.: Europe is facing a demographic chal-

lenge ambient assisted living offers solutions. Technical report, VDI/VDE-IT (2006)

2. Yamauchi, K., Chen, W., Wei, D.: 3g mobile phone applications in telemedicine - a sur-

vey. In: Computer and Information Technology, 2005. CIT 2005. The Fifth International

Conference on. (2005) 956 – 960

3. Wootton, R., Craig, J., Patterson, V.: Introduction to Telemedicine. Oxford University Press

(2010)

4. Vodafone: Pulido valente hospital and vodafone foundation present telemonitoring of long

term oxygen therapy (2009)

5. Vitacca, M., Comini, L., Tentorio, M., Assoni, G., Trainini, D., Fiorenza, D., Morini, R.,

Bruletti, G., Scalvini, S.: A pilot trial of telemedicine-assisted, integrated care for patients

with advanced amyotrophic lateral sclerosis and their caregivers. Journal of Telemedicine

and Telecare 16 (2010) 83–88

120

6. Wan, D., Mentzer, K., Nash, R.: Always on, always connected how technology will transform

the future of chronic care. Technical report, Accenture (2009)

7. Zheng, P., Ni, L.: Smart Phone and Next Generation Mobile Computing. Morgan Kaufmann

(2005)

8. Vergne, H.D.L., Milanesi, C., Cozza, R., Gupta, A., Lu, C., Nguyen, T., Zimmermann, A.:

Competitive landscape: Mobile devices, worldwide, 2q10. Technical report, Gartner (2010)

9. Figueredo, M., Dias, J.: Mobile telemedicine system for home care and patient monitor-

ing. In: Proceedings of the 26th Annual International Conference of the IEEE EMBS, IEEE

EMBS (2006)

10. Wang, C.S., Lee, J.H., Chu, Y.T.: Mobile telemedicine application and technologies on gsm.

In: Bioinformatics and Biomedical Engineering, 2007. ICBBE 2007. The 1st International

Conference on. (2007) 1125 –1128

11. Zhang, P., Kogure, Y., Matsuoka, H., Akutagawa, M., Kinouchi, Y., Zhang, Q.: A remote pa-

tient monitoring system using a java-enabled 3g mobile phone. In: Engineering in Medicine

and Biology Society, 2007. EMBS 2007. 29th Annual International Conference of the IEEE.

(2007) 3713 –3716

12. Laakko, T., Leppanen, J., Lahteenmaki, J., Nummiaho, A.: Multipurpose mobile platform

for telemedicine applications. In: Pervasive Computing Technologies for Healthcare, 2008.

PervasiveHealth 2008. Second International Conference on. (2008) 245 –248

13. Finlay, D., Nugent, C., Breen, C., Bond, R., Moore, G.: A smartphone based telemedicine

system for recording limited lead body surface potential maps. In: Computers in Cardiology,

2009. (2009) 237 –240

14. Salomon, D.: Data Compression: The Complete Reference. Springer-Verlag (1998)

15. Friesen, G., Jannett, T., Yates, M.J.S., Quint, S., Nagle, H.: A comparison of the noise

sensitivity of nine QRS detection algorithms. IEEE Transactions on Biomedical Engineering

37 (1990) 85 –98

121