ABCDEFG

UNIVERS ITY OF OULU P.O.B . 7500 F I -90014 UNIVERS ITY OF OULU F INLAND

A C T A U N I V E R S I T A T I S O U L U E N S I S

S E R I E S E D I T O R S

SCIENTIAE RERUM NATURALIUM

HUMANIORA

TECHNICA

MEDICA

SCIENTIAE RERUM SOCIALIUM

SCRIPTA ACADEMICA

OECONOMICA

EDITOR IN CHIEF

PUBLICATIONS EDITOR

Professor Mikko Siponen

University Lecturer Elise Kärkkäinen

Professor Hannu Heusala

Professor Olli Vuolteenaho

Senior Researcher Eila Estola

Information officer Tiina Pistokoski

University Lecturer Seppo Eriksson

Professor Olli Vuolteenaho

Publications Editor Kirsti Nurkkala

ISBN 978-951-42-6387-3 (Paperback)ISBN 978-951-42-6388-0 (PDF)ISSN 0355-3213 (Print)ISSN 1796-2226 (Online)

U N I V E R S I TAT I S O U L U E N S I SACTAC

TECHNICA

U N I V E R S I TAT I S O U L U E N S I SACTAC

TECHNICA

OULU 2010

C 373

Young-Dong Lee

WIRELESS VITAL SIGNS MONITORING SYSTEM FOR UBIQUITOUS HEALTHCARE WITH PRACTICAL TESTS AND RELIABILITY ANALYSIS

UNIVERSITY OF OULU,FACULTY OF TECHNOLOGY,DEPARTMENT OF ELECTRICAL AND INFORMATION ENGINEERING

C 373

ACTA

Young-Dong Lee

C373etukansi.kesken.fm Page 1 Tuesday, November 16, 2010 4:02 PM

A C T A U N I V E R S I T A T I S O U L U E N S I SC Te c h n i c a 3 7 3

YOUNG-DONG LEE

WIRELESS VITAL SIGNS MONITORING SYSTEM FOR UBIQUITOUS HEALTHCAREWITH PRACTICAL TESTS AND RELIABILITY ANALYSIS

Academic dissertation to be presented with the assent ofthe Faculty of Technology of the University of Oulu forpublic defence in Auditorium TS101, Linnanmaa, on 10December 2010, at 12 noon

UNIVERSITY OF OULU, OULU 2010

Copyright © 2010Acta Univ. Oul. C 373, 2010

Supervised byDocent Esko AlasaarelaProfessor Risto Myllylä

Reviewed byProfessor Mohammed ElmusratiProfessor Jukka Vanhala

ISBN 978-951-42-6387-3 (Paperback)ISBN 978-951-42-6388-0 (PDF)http://herkules.oulu.fi/isbn9789514263880/ISSN 0355-3213 (Printed)ISSN 1796-2226 (Online)http://herkules.oulu.fi/issn03553213/

Cover DesignRaimo Ahonen

JUVENES PRINTTAMPERE 2010

Lee, Young-Dong, Wireless vital signs monitoring system for ubiquitous healthcarewith practical tests and reliability analysisUniversity of Oulu, Faculty of Technology, Department of Electrical and InformationEngineering, P.O.Box 4500, FI-90014 University of Oulu, FinlandActa Univ. Oul. C 373, 2010Oulu, Finland

AbstractThe main objective of this thesis project is to implement a wireless vital signs monitoring systemfor measuring the ECG of a patient in the home environment. The research focuses on two specificresearch objectives: 1) the development of a distributed healthcare system for vital signsmonitoring using wireless sensor network devices and 2) a practical test and performanceevaluation for the reliability for such low-rate wireless technology in ubiquitous health monitoringapplications.

The first section of the thesis describes the design and implementation of a ubiquitoushealthcare system constructed from tiny components for the home healthcare of elderly persons.The system comprises a smart shirt with ECG electrodes and acceleration sensors, a wirelesssensor network node, a base station and a server computer for the continuous monitoring of ECGsignals. ECG data is a commonly used vital sign in clinical and trauma care. The ECG data isdisplayed on a graphical user interface (GUI) by transferring it to a PDA or a terminal PC. Thesmart shirt is a wearable T-shirt designed to collect ECG and acceleration signals from the humanbody in the course of daily life.

In the second section, a performance evaluation of the reliability of IEEE 802.15.4 low-ratewireless ubiquitous health monitoring is presented. Three scenarios of performance studies areapplied through practical tests: 1) the effects of the distance between sensor nodes and base-station, 2) the deployment of the number of sensor nodes in a network and 3) data transmissionusing different time intervals. These factors were measured to analyse the reliability of thedeveloped technology in low-rate wireless ubiquitous health monitoring applications.

The results showed how the relationship between the bit-error-rate (BER) and signal-to-noiseratio (SNR) was affected when varying the distance between sensor node and base-station, throughthe deployment of the number of sensor nodes in a network and through data transmission usingdifferent time intervals.

Keywords: BER, ECG, IEEE 802.15.4, acceleration signal, personal area networks,reliability analysis, vital signs monitoring

5

Acknowledgements

I wish to express my warm and sincere thanks to the University of Oulu and

Dongseo University for giving me the opportunity and resources to write this

thesis through their excellent dual-degree program. This work would not have

been possible without support from the dual-degree program between the

University of Oulu and Dongseo University.

I would also like to express my deep and sincere gratitude to my supervisor,

Professor Esko Alasaarela, for his understanding, encouragement and guidance

throughout the work of this thesis. His wide knowledge and technical thinking

have been of great value to me.

I am, as well, deeply grateful to my second supervisor, Professor Risto

Myllylä, Head of the Optoelectronics and Measurement Techniques Laboratory,

for his important support throughout this work.

Furthermore, I wish to express my sincere thanks to Professor Hoon-Jae Lee

who gave me important guidance and untiring help during my difficult moments.

Additionally, during this work I have collaborated with many colleagues for

whom I have great regard, and I wish to extend my warmest thanks to all those

who have helped me with my work in the Optoelectronics and Measurement

Techniques laboratory at the University of Oulu, and the USN laboratory at

Dongseo University, Korea.

Finally, I would like to convey my loving thanks to my wife Mi-Young, and

my son Hyeong-Gyu, for their encouragement and understanding.

Oulu, March 2010 Young-Dong Lee

6

7

List of terms, symbols and abbreviations

A/D Analog/Digital

ACK Acknowledgement

ADC Analog-to-Digital Converter

AWGN Additive White Gaussian Noise

BAN Body Area Network

BASN Body Area Sensor Network

BER Bit Error Rate

BMAC Berkeley MAC

BSN Body Sensor Network

CCITT International Telegraph and Telephone Consultative Committee

CRC Cyclic Redundancy Check

CTS Clear to Send

dB Decibels

dBm Decibels per Milliwatt

DMAC Distributed MAC

DSDV Destination Sequence Distance Vector

DSMAC Differential Service MAC

Eb/No Energy per Bit to Noise power spectral density ratio

ECG Electrocardiogram

EMT Emergency Medical Technician

g Gravitational constant

GHz Giga Hertz

GPRS General Packet Radio Service

GUI Graphic User Interface

HPF High Pass Filter

I/O Input/Output

ID Identification

ISM Industrial Scientific Medical

ISP In System Programmer

KB Kilobyte

Kbps Kilobits Per Second

LPF Low Pass Filter

LR-WPAN Low-Rate Wireless Personal Area Network

mA Milli-Ampere

MAC Media Access Control

8

MB Mega Byte

MHz Mega Hertz

mm Millimeter

NACK Negative Acknowledgement

nesC Network Embedded System C

O-QPSK Offset Quadrature Phase-Shift Keying

PCB Printed Circuit Board

PDA Personal Digital Assistant

PER Packet Error Rate

PHY Physical Layer

PSK Phase-Shift Keying

QoS Quality of Service

RAM Random Access Memory

RF Radio Frequency

RS-232 Recommended Standard 232

RTS Request to Send

Rx Receiver

SMAC Sensor MAC

SNR Signal-to-Noise Ratio

SPI Serial Peripheral Interface Bus

SpO2 Oxygen Saturation

TCP/IP Transmission Control Protocol/Internet Protocol

TinyOS Tiny Micro Threading Operating System

TMAC Timeout MAC

Tx Transmitter

uA Micro-Ampere

UART Universal Asynchronous Receiver/Transmitter

UbiMon Ubiquitous Monitoring Environment for Wearable and Implantable

Sensors

UMTS Universal Mobile Telecommunication System

USB Universal Serial Bus

USN Ubiquitous Sensor Network

WiFi WLAN/IEEE 802.11

WLAN Wireless Local Area Network

WSN Wireless Sensor Network

9

List of original papers

I Lee Y-D, Lee D-S, Chung W-Y & Myllylä R (2006) A wireless sensor network platform for elderly persons at home. Rare metal materials and engineering 35(3): 95–99.

II Amit P, Lee Y-D & Chung W-Y (2008) Triaxial MEMS accelerometer for activity monitoring of elderly person. Sensor Letters 6(6): 1054–1058.

III Lee Y-D & Chung W-Y (2009) Wireless sensor network based wearable smart shirt for ubiquitous health and activity monitoring. Sensors and Actuators B: Chemical 140(2): 390–395.

IV Walia G, Lee Y-D, Chung W-Y & Myllylä R (2006) Distributed healthcare monitoring in an ad-hoc network using a wireless sensor network. Proceedings of 11th International Meeting on Chemical Sensors, Brescia, Italy.

V Lee Y-D, Lee D-S, Walia G, Myllylä R & Chung W-Y (2006) Query based duplex vital signal monitoring system using wireless sensor network for ubiquitous healthcare. Proceedings of World Congress on Medical Physics and Biomedical Engineering, Seoul, Korea: 396–399.

VI Lee Y-D, Lee D-S, Walia G, Alasaarela E & Chung W-Y (2007) Design and evaluation of query supported healthcare information system using wireless sensor ad-hoc network. Proceedings of International Conference on Convergence Information Technology, Gyeongju, Korea: 997–1002.

VII Lee Y-D, Alasaarela E & Chung W-Y (2007) A wireless health monitoring system in multi-hop body sensor networks. Proceedings of 2nd International Symposium on Medical Information and Communication Technology (ISMICT’'07), Oulu, Finland.

VIII Chung W-Y, Lee Y-D & Jung S-J (2008) A wireless sensor network compatible wearable u-healthcare monitoring system using integrated ECG, accelerometer and SpO2. Proceedings of 30th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC 2008), Vancouver, Canada: 1529–1532.

The objective of paper I was to design and implement a small size ubiquitous

healthcare system for the home care of elderly persons. The system comprised a

wireless sensor network node, base station and server computer for the continuous

monitoring of ECG signals. ECG data, important vital heart signals that are

commonly used in clinical and trauma care, were displayed on a graphical user

interface (GUI) by transferring the data to a PDA or a terminal PC.

Paper II presents an activity classification system using an MEMS

accelerometer and a wireless sensor node. The three axes accelerometer measures

body acceleration signals and transmits measured data with the help of a sensor

node to a base-station. On the PC, real time accelerometer data is processed for

various classifications; rest-fall, running-fall and walking-running.

10

Wearable sensor devices are proposed in paper III that are designed to fit well

into a shirt, which are of small size, low power consumption and offer the patients

wireless sensor network communication based on IEEE 802.15.4. The shirt

mainly consists of sensors for continuous monitoring of health data and

conductive fabrics to get the body signal, functioning as electrodes.

Paper IV describes the development of a distributed health monitoring system

prototype for clinical and trauma patients using a commercially available wireless

sensor network platform. The system has been designed to measure various vital

signs of the patients and transfer the health status wirelessly to a remote base-

station which was connected to a doctor’s PDA and terminal PC.

Paper V discusses the design issue for a query driven healthcare system and

its implementation using a wireless sensor node in an ad-hoc network. The paper

describes the system's architecture, including hardware design, query model and

experimental setup, and results and performance analysis are also presented.

Papers VI and VII present a wireless health monitoring system in multi-hop

body sensor networks using our ubiquitous sensor network platform. A mote

based 3-electrodes ECG monitoring system operates in wireless sensor networks.

The ECG signal from multiple patients was relayed wirelessly using a multi-hop

routing scheme to a base-station. The design and development of a wearable u-

Healthcare monitoring system using integrated ECG, accelerometer and oxygen

saturation sensors is introduced in paper VIII.

The author defined the research plan and carried out the literature review, and

wrote a manuscript presented in papers I-VIII. Experiments and measurements

were done by the author in association with the co-authors. The manuscript for

papers I, III-VIII, was written by the author with the kind help of the co-authors.

The author’s contribution to paper II was to design and fabricate a tri-axial

accelerometer sensor system using wireless sensor nodes. The author then carried

out the accelerometer sensor measurements together with Amit Purwar. He wrote

the first version of the manuscript, after discussions with the co-authors according

to the reviewer’s comments, the author wrote the revision of the manuscript.

11

Contents

Abstract Acknowledgements 5 List of terms, symbols and abbreviations 7 List of original papers 9 Contents 11 1 Introduction 13

1.1 Related works .......................................................................................... 13 1.2 Objectives and outline of the thesis ........................................................ 15

2 Healthcare system requirements 17 2.1 Node lifetime .......................................................................................... 17 2.2 Packet recovery ....................................................................................... 17 2.3 Data collection ........................................................................................ 18 2.4 Dependent and independent waveforms ................................................. 18 2.5 Scalability ............................................................................................... 19

3 Vital signs monitoring system 21 3.1 System architecture of a general u-Healthcare system ............................ 21 3.2 MICAz node and ECG interface circuit .................................................. 22 3.3 USN platform for wireless sensor networking ........................................ 24 3.4 Sensor boards for vital signs monitoring................................................. 26

3.4.1 Single ECG interface circuit for MICAz node ............................. 26 3.4.2 ECG & accelerometer sensor board ............................................. 26 3.4.3 Wearable sensor node with sensor board ...................................... 28

3.5 Sensor node software .............................................................................. 30 3.5.1 Sampling time for the ECG signal ................................................ 33 3.5.2 Packet format ................................................................................ 33 3.5.3 Routing protocol ........................................................................... 34

4 Reliable medical data transmission 35 4.1 Need for reliable medical data transmission ........................................... 35 4.2 Low-rate wireless personal area networks .............................................. 36 4.3 Error rate analysis for reliability ............................................................. 36

5 Results 39 5.1 Experimental setup and ECG signal measurement ................................. 39 5.2 Wearable u-Healthcare monitoring ......................................................... 39 5.3 ECG data on the client display ................................................................ 42 5.4 PDA display ............................................................................................ 43

12

5.5 Reliability analysis .................................................................................. 44 5.5.1 Reliability test in function of distance between sensor

nodes ............................................................................................. 44 5.5.2 Reliability test in the function of the number of the sensor

nodes ............................................................................................. 47 5.5.3 Reliability test for different time intervals .................................... 49

6 Discussion 51 6.1 Multi-hop network for wireless body area network ................................ 51 6.2 Electrocardiogram and accelerometer sensor .......................................... 52 6.3 Wearable system for ubiquitous healthcare ............................................. 53

7 Conclusion 55 References 57 Original Papers 61

13

1 Introduction

An increasingly aging population has raised the concern of many countries about

its impact on health care systems. This has led to an urgent need for devising

cheaper and smarter ways to provide health care for suffers of age related disease.

Recent advances in sensor, communication and information technologies have

enabled the development of novel vital signs monitoring systems by which

various vital health parameters can be measured, like electrocardiogram (ECG),

body temperature, heart rate, blood pressure and oxygen saturation. In particular,

wireless healthcare related applications using wireless sensor networks [1–4] may

assist residents and caregivers by providing non-invasive and invasive continuous

health monitoring with a minimum interaction of doctors and patients. However,

there are some significant disparities, such as node lifetime, packet recovery and

medical data collection between existing wireless sensor networks and those

required for healthcare [5, 6]. In addition to the basic requirements of being

wireless, a future healthcare system based on wireless sensor networks must have

support for ad-hoc [7] networks, the mobility of patients or elderly persons, wide

ranges of data rates and a high degree of reliability. Also, these systems need to be

miniature in size, and should be easily wearable, thus causing the least

inconvenience to the patients.

The design of a sensor network is application-specific, and different

applications have different reliability [8] requirements. Reliable data

communication is an important factor for the dependability and quality-of-service

(QoS) [9–11] in several applications of wireless sensor networks. In particular,

this is the case for critical care applications of hospitals that allow the continuous

monitoring of a patient’s vital signs. We must consider the reliability of those

sensors, the sampling rate, the number of nodes in the same network and data

packet sizes, as well as the reliability of the medical devices themselves. This is

closely related to the performance of the system in medical applications [12].

1.1 Related works

The MobiHealth [13] project has developed a system for ambulant patient

monitoring over public wireless networks. Based on the body area network (BAN)

[14–17] interconnecting different vital signal sensors and actuators, the

measurements are transmitted using UMTS (or GPRS) to the healthcare center

where they are presented to the medical personnel. In this way, patients can be

14

continuously monitored and receive advice when needed. The MobiHealth

patient/user is equipped with different vital constant sensors, like blood pressure,

pulse rate and ECG, interconnected via the healthcare BAN. The mobile base unit

is the central point of healthcare BAN, acting as a gateway aggregating the vital

sensor measurements (intra-BAN communication based on a wireless network

like Bluetooth [18] and Zigbee [19]) and transmitting them to the back-end

system (extra-BAN communication based on GPRS and UMTS), which can be

located within the health broker premises or be part of a wireless service provider.

From there, the measurements are dispatched to the healthcare broker where the

medical personnel monitor them.

The ubiquitous monitoring environment for wearable and implantable sensors

project (UbiMon) [20] at Imperial College London aims to provide a continuous

and unobtrusive monitoring system for patients in order to capture transient but

life threatening events. The basic framework of the UbiMon is based on the body

sensor network (BSN) design. The system consists of five major components,

namely the BSN nodes, the local processing unit, the central server, the patient

database and the workstation. With the current UbiMon structure, a number of

wireless biosensors including 3-leads ECG, 2-leads ECG strips, and oxygen

saturation (SpO2) sensors have been developed. To facilitate the incorporation of

context information, sensor-based context awareness, including accelerometers,

temperature and skin conductance sensors are also integrated into the BSN node.

Furthermore, a compact flash BSN card is developed for personal digital

assistants (PDAs), where sensor signals can be gathered, displayed and analyzed

by the PDA.

CodeBlue [21] has been designed to operate across a wide range of devices,

including low power motes, PDAs and PCs, and addresses the special robustness

and security requirements of medical care settings. The device consists of a

MICA2 [22] mote with a pulse oximetry signal processing module. The device

transmits periodic packets containing heart rate, SpO2 and photoplethysmogram

waveform data. Vital sign data from multiple patients can be relayed using an

adaptive, multi-hop routing scheme either to a wired base-station (such as a PC or

laptop) or directly to multiple handheld PDA devices carried by emergency

medical technicians (EMT’s), physicians, or nurses. CodeBlue has been designed

to provide routing, naming, discovery, and security for wireless medical sensors,

PDAs, PCs and other devices that may be used to monitor and treat patients in a

range of medical settings.

15

1.2 Objectives and outline of the thesis

The main objective of this thesis project is to implement a wireless vital signs

monitoring system for measuring the ECG of a patient in the home environment.

The research focuses on two specific research objectives: 1) development of a

distributed healthcare system for vital signs monitoring using wireless sensor

network devices and 2) a practical test and performance evaluation for the

reliability for such low-rate wireless technology in ubiquitous health monitoring

applications.

The content of this thesis is summarized as follows:

Chapter 2 addresses the requirements for a ubiquitous healthcare system. The

main requirements for wireless healthcare such as node lifetime, packet recovery,

data collection, waveform and scalability are presented.

Chapter 3 presents the design and implementation of a ubiquitous healthcare

system that applies small-size components for the home care of elderly persons.

Two kinds of health monitoring systems have been developed; a single ECG

interface circuit for a MICAz [23] node and multiple sensors, ECG and

accelerometer sensor systems. In addition to this, a wearable ubiquitous

healthcare system for vital signs monitoring is presented.

Chapter 4 discusses reliable medical data transmission for improving the

quality-of-service, including the need for reliable medical data transmission, low-

rate wireless personal area networks (LR-WPAN) and error rate analysis for

reliability.

Chapter 5 presents the experimental results with a focus on vital signs

monitoring and performance evaluation.

Chapter 6 provides a summary of this work.

16

17

2 Healthcare system requirements

Wireless healthcare should satisfy the main requirements such as node lifetime,

packet recovery, data collection, waveform and scalability.

2.1 Node lifetime

The lifetime of a wireless sensor network is application-specific. With continuous

healthcare application, it is necessary to monitor a patient’s condition

continuously (keep the wakeup sensor active) because an event

(electrocardiogram and heart rate) can occur at any time. An important issue in

wireless sensor networks for continuous healthcare application is to maximize the

node lifetime while maintaining application or other constraints, such as delay

and reliability [24, 25]. In order to maximize node lifetime for continuous

healthcare application, power saving techniques should be considered through

avoiding unnecessary packet recovery, efficient data collection and reducing

packet losses.

2.2 Packet recovery

In continuous healthcare application, it is not necessary to use a packet recovery

mechanism because of packet loss, as it wastes resources. Any lost packet will be

replaced by the most recent update in the next transmission. Also, retransmission

of packets will require more storage memory at the motes, which is generally

limited. Since this is a healthcare scenario, where each node attached to a patient

transfers data independently of other nodes in the network, protocols like SMAC

[26], TMAC [27] and DSMAC [28], which need RTS/CTS packet exchanges for

lost packet recovery and have synchronized sleep and wake periods, are ignored.

SIFT [29] is a MAC protocol for event-driven wireless sensor network

environments. Although SIFT [29] achieves low latency at the cost of some

increase in energy consumption, it needs a system-wide synchronization for a

slotted contention window and would be complex to use with protocols not

utilizing synchronization.

18

2.3 Data collection

While designing wireless sensor networks, the choice of a MAC layer protocol

also depends on the application requirements, topology used and network layer

routing protocol. A healthcare system generally requires a single monitoring base-

station, therefore a convergent type communication pattern would be a wise

choice, where data from the entire sensor network is finally collected at a single

base-station. For convergent type communication, the DMAC [30] protocol is a

good choice as it achieves very good latency compared to other sleep/listen

periods. Although DMAC does not avoid collision detection, the possibilities of

collision are reduced by the fact that in healthcare environments events will not

trigger all nodes simultaneously. The BMAC [31] protocol could also be a

possible choice, as it avoids packet collision and has a power management

scheme via low power listening, but unlike SMAC, it does not perform link level

retransmission or hidden terminal avoidance using RTS/CTS schemes [32]. The

best thing about BMAC is that it offers control to the protocols above it, allowing

the routing and application layers to change control parameters. This cross layer

communication is highly demanded in the designing of wireless sensor

applications, where often each communication layer not only needs to interact

with the intermediate layer, but also with the other layers directly.

2.4 Dependent and independent waveforms

Healthcare data which is waveform independent does not need continuous data

transfer, and therefore data transfer should be initiated only when desired by the

monitoring side, or in periods of finite durations. On the other hand, waveform

dependent data requires long-term medical data transfer in continuous healthcare

application. Therefore, for waveform dependent data, data-centric approaches

would be well suited where a query or command from the base-station initiates

the data transfer. The waveform independent data can use either the data-centric

approach or the event-driven approach where data is transferred when a specific

event is sensed. Generally, waveform independent data changes gradually without

any periodic sequence. So, only the change in amplitude over a specified

threshold is required to detect an abnormal event. To limit the amount of data

transferred, a threshold limit can be set, and if the value of sampled data goes

beyond the threshold, only then is the data transfer initiated. Also, packet loss is

not a major issue for wave independent data because even if a few packets are

19

missed they can be replaced. For waveform dependent data, any packet lost can

be a big problem as it will result in a loss of useful information or can also give a

false impression of abnormality about a patient’s health parameter.

2.5 Scalability

Another important issue is that the routing of message packets from or to moving

nodes is more challenging, since stability becomes an important optimization

factor, in addition to energy and bandwidth etc. Since fast data delivery is a prime

requirement in healthcare, proactive routing protocols will be efficient, because in

this case, the path to be chosen is known in advance, before data transmission.

Also, any change in topology is updated by timely updates. Among all the

proactive protocols, a DSDV [33] based protocol would be better in

implementation and supports a flat network. The key advantage of DSDV is that

it guarantees loop freedom.

In a healthcare network, as data generated from different sensors will be at

different rates, subject to different quality of service constraints and following a

multiple data delivery model, traditional methods of data aggregation cannot be

applied. Scalability is also major issue because single gateway architecture is not

scalable for a large set of sensors [34].

20

21

3 Vital signs monitoring system

3.1 System architecture of a general u-Healthcare system

Ubiquitous healthcare is built around the concept of placing unobtrusive sensors

on a person’s body to form a wirelessly connected network. Personal area sensor

networks [35–37] make it possible to continually monitor patients almost

anywhere and immediately notify clinicians, the nearest hospital or an emergency

service of any critical change in status. Such networks can quickly transfer

biomedical data from sensors deployed on the body to a server computer for

processing.

Fig. 1 illustrates the system architecture of a general ubiquitous healthcare

system for monitoring patient status. The system consists of three major

components, namely, a body area sensor network (BASN) node, local processing

unit and central server.

Fig. 1. System architecture of ubiquitous healthcare system for vital signs monitoring

[I, published by permission of Rare Metal Materials and Engineering].

Health parameters obtained from the sensor nodes attached on the body of a

patient can be monitored by a PC or a PDA terminal using standard TCP/IP or

IEEE 802.11 (WLAN) connections. Integrated to the BSN node there are several

wireless biological sensors, including ECG, SpO2, and accelerometers, as well as

temperature and skin conductance sensors, used to monitor the patient’s vital

signs and context information. The base-station is a stand-alone unit, capable of

22

receiving data from the wireless sensor node using an RF receiver, as well as

storing and transferring the data to a mobile unit using TCP/IP or IEEE 802.11.

An example of ubiquitous healthcare is provided by the ubiquitous monitoring

environment for wearable and implantable sensors (UbiMon) project [20] at

Imperial College, London, UK. This project aims to provide a continuous and

unobtrusive monitoring system for patient care in order to capture transient, but

life threatening events. To this end, a compact flash BSN card has been developed

for PDAs, allowing sensor signals to be gathered, displayed and analyzed by the

PDA. Apart from acting as the local processor, the PDA can also serve as the

router between the BSN node and a central server to which all collected sensor

data will be transmitted through a WiFi/GPRS network for long-term storage and

trend analysis. Monitoring terminals are utilized to allow clinicians to analyze

patient data using such portable handheld devices as laptops, PDAs or cellular

phones. In addition to real-time sensor information, historical data can also be

retrieved and played back to assist the diagnosis. [Paper I]

3.2 MICAz node and ECG interface circuit

Fig. 2 shows the system architecture of the wireless sensor network platform

discussed in this thesis, consisting of four major parts: sensor node, data

acquisition board, ECG interface circuit and base-station. As a sensor node, this

study used the commercially available MICAz (Crossbow Technology, Inc.,

USA). The MICAz has a Chipcon CC2420 radio chip set operating in the IEEE

802.15.4 standard (Zigbee compatible), an Atmega 128L microprocessor (Atmel

Corporation, USA) with a data throughput of 250 kbps. The data acquisition

board has a maximum of 11 input channels, each with its own 12-bit analog-to-

digital converter (ADC).

23

Fig. 2. Hardware configuration of the wireless sensor network [I, published by

permission of Rare Metal Materials and Engineering].

One of most commonly monitored vital signs in clinical and trauma care are ECG

signals. A two- or three-electrode ECG is used to evaluate cardiac activity for an

extended period. Physicians, clinicians or the patients themselves can then

continuously watch for signs of deterioration in the patient’s condition and of

cardiac problems such as arrhythmia that occur intermittently, maybe only once or

twice a day. In this thesis, an ECG signal generator was used to get a more

accurate ECG signal for our preliminary experiment. In the rest of the

experiments, the ECG signal was obtained from sensors attached to a real human

body.

After processing in an interface circuit, signals produced by the ECG signal

generator were recorded by the MICAz wireless sensor node. MICAz nodes run a

custom operating system called TinyOS [38], which supports multiple threads,

allowing the performance of a variety of tasks tailored specifically for wireless

sensor network applications. Each node is capable of sending and receiving

specific wireless network packets called active messages. A similar MICAz node

was used as the base-station with an MIB510 serial PC interface to allow the

sensor node to communicate with a PC. Fig. 3 presents an inside block diagram of

a MICAz node consisting of five major modules: a micro-controller with internal

flash program memory, Zigbee compliant chipsets (CC2420, Chipcon Ltd.,

Norway), 512 KB of external serial flash memory, an ECG interface circuit and

an I/O connector, functioning as the input port of the sensor’s interface circuit.

[Paper I]

24

Fig. 3. Block diagram of a MICAz sensor node [I, published by permission of Rare

Metal Materials and Engineering].

3.3 USN platform for wireless sensor networking

Fig. 4 shows a block diagram of the USN platform developed by our laboratory.

The specifications of the USN platform are summarized in Table 1.

The USN platform features an ultra low power Texas Instruments MSP430

micro-controller [39] with 10KB RAM, 48KB flash memory and 12-bit A/D

converter. It supports several low power operating modes and consumes as low as

5.1uA in sleep mode and 1.8mA in active mode. The CC2420 wireless transceiver

[40] is IEEE 802.15.4 Zigbee compliant and has programmable output power, a

maximum data rate of 250 kbps and hardware that provides PHY and some MAC

layer functions. The CC2420 is controlled by the MSP430F1611 through a SPI

port and a series of digital I/O lines. The M25P80 is an 8 MB serial flash memory

with a write protection mechanism, accessible from a SPI bus. To minimize the

size of the USN platform, the USB programming board was made as a separate

module which is needed only when nodes are connected to the PC either for

application download or when the node acts as a base-station. [Paper VII]

25

Fig. 4. Block diagram of the USN platform.

Recent research has focused on the development of wireless sensor networks and

pervasive monitoring systems that can monitor the status of a patient (ECG, body

temperature, blood pressure, SpO2, accelerometer, etc.) for ubiquitous healthcare

applications. Therefore, a low power operating ECG sensor board was developed

for patient health monitoring, and the output of the ECG sensor board was

connected to a USN platform.

Table 1. Specifications of USN platform.

Item Specification Remark

Platform

Processor MSP430F1611

Memory 48 KB/10 KB Program flash/data RAM

Current draw 1.8 mA/5.1 uA Active/sleep mode

ADC 12 bit resolution 8 channels

Power 3V Battery

Connector Expansion/USB type

RF transceiver

Frequency band 2.4–2.485 GHz

Sensitivity −95 dBm

Transfer rate 250 Kbps

RF power −25 dBm–0 dBm

Range 150 m/20–30 m Outdoor/indoor

Current draw 18.8/17.4 mA/1 uA Rx/Tx/Sleep

Antenna Ceramic PCB antenaa

26

3.4 Sensor boards for vital signs monitoring

3.4.1 Single ECG interface circuit for MICAz node

An ECG signal generator with an interface circuit was attached to the MICAz

node for the purpose of continuously supplying ECG signals identical to those

produced by an ECG sensor on a patient’s body.

Fig. 5 shows a block diagram of the ECG interface circuit. The ECG interface

circuit is required for the amplification of the output signal produced by the ECG

signal generator, signal conditioning and analog to digital conversion.

Fig. 5. Block diagram of an ECG interface circuit [I, published by permission of Rare

Metal Materials and Engineering].

3.4.2 ECG & accelerometer sensor board

Fig. 6 illustrates the block diagram of the sensor board which consists of an ECG

interface and a three-axis accelerometer sensor [Paper II] circuit. An ECG is a

bioelectric signal which records the heart’s electrical activity versus time;

therefore it is a very important and basic diagnostic tool for assessing heart

functions. An electrocardiogram is obtained by measuring electrical potential

between various points of the body using a biomedical instrumentation amplifier.

The electrical activity of the heart can be measured on the body surface by

attaching a set of electrodes to the skin. A standard 12-lead ECG is normally

acquired by using ten electrodes placed at specific locations on the patient’s body.

It is possible to reduce the number of leads to be used, due to the redundancy in

the cardiac information present in each of the standard ECG leads. ECG signals

27

from at least two leads are generally used in order to generate an

electrocardiogram representation of the cardiac electrical activity. Another

important part of home healthcare is to monitor behaviour and physical activity in

daily life. There are many exisitng research studies [41–43] to develop fall

detection systems using a 3-axis accelerometer. The system can collect the

accelerometer signals to determine whether the person with the device attached

has fallen or not.

ECG signals from the electrodes are amplified with a gain of 300 (24.8 dB)

and filtered with the cut-off frequencies of 0.05 Hz and 123 Hz in the sensor

board. An ECG electrode has two conductive fabric electrodes (Polar Electro Oy,

Finland) which are woven into the fabric. In addition, the sensor board also has a

three-axis accelerometer sensor (MMA7260Q, Freescale) [Paper II] to measure

acceleration signals for the activity monitoring of a patient. The shape of a sensor

board with a wireless sensor node is designed with a contoured shape for

comfortable wear and convenience. [Paper III]

Fig. 6. Block diagram of a sensor board [III, published by permission of Elsevier].

Table 2. Sensor board specification.

Specifications

ECG sensor specification

ECG type Chest-belt type

Gain 300 (24.8)dB

Bandwidth 0.05–123 Hz

Supply voltage 3.3 V

Accelerometer sensor specification

Acclerometer type Chest-belt type

Axis 3-Axis (X, Y, Z)

Selective sensitivity 1.5/2/4/6 g

Supply voltage 3.3 V

28

3.4.3 Wearable sensor node with sensor board

Fig. 7 shows the overall system architecture of the wearable smart shirt for

ubiquitous health and activity monitoring, which consists of a shirt with

integrated wireless sensor nodes, a base-station and server PC for remote

monitoring. The wireless sensor network consists of a large number of small

nodes, which have built-in computing, power, sensors to acquire physiological

and activity data from the human body and wireless transmission and reception

capability. The smart shirt is compatible with the wireless sensor network, thus

the individual physiological data from each smart shirt is transmitted in ad-hoc

wireless communication for further processing using a wireless link. [Paper III]

The wearable sensor nodes are responsible for acquiring the physiological

data and transmitting it to the base-station. The sensor nodes are designed to be

tiny in size and consume low operating power to reduce battery size, which can

also last for longer durations. The sensor node has a limited battery power, as well

as computing and communication capability, due to its physical structure.

Fig. 7. System architecture of the u-Healthcare system with wearable smart shirt [III,

published by permission of Elsevier].

The measured ECG and accelerometer data are transmitted to a server PC in a

wireless sensor network. The ECG signal is one of the most important vital signs

for knowing the health status of a patient or elderly person, and three axis

accelerometer signals are used in a medical device for detecting that a patient or

elderly person has fallen. If both signals, that is, ECG and accelerometer data, are

measured simultaneously, the resolution of the diagnosis can be improved. Fig. 8

29

shows the architecture of the designed wireless sensor node. The wireless sensor

node is round in shape and 40 mm in diameter.

Fig. 8. Wireless sensor node (a) front side, (b) back side [III, published by permission

of Elsevier].

(a)

(b)

30

The smart shirt consists of a wireless sensor node, an ECG & accelerometer

sensor board and conductive fabric electrodes for ECG measurement in a normal

shirt design. To reduce the size of the integrated wearable sensor node, the

structure is made up of two round shaped PCB boards which are composed of a

wireless sensor node plate for communication in a wireless sensor network and a

sensor board plate with an ECG interface & accelerometer. The wireless sensor

node and sensor board are combined together as shown in Fig. 9. The wireless

sensor node is placed in the top position and the sensor board with ECG interface

and accelerometer circuits is placed in the bottom position of the wearable sensor

node structure. To obtain physiological ECG data, two conductive fabric

electrodes are extended from the ECG interface circuit of the sensor board and are

knitted into the shirt. The smart shirt is a wearable T-shirt designed to collect

ECG and acceleration signals from the human body continuously in daily life.

The shirt contains ECG and accelerometer sensors that can be used to monitor

vital signs such as heart rate, ECG and acceleration. The double layer structure of

the wearable sensor node reduces the width of the normal single layer wireless

sensor node structure with sensors and gives convenience in wear with two AAA

size batteries. [Paper III]

Fig. 9. An integrated wearable sensor node combined with a wireless sensor node and

a sensor board in double layer structure.

3.5 Sensor node software

The sensor node software that runs on the USN platform takes samples, collects

physiological data and does multi-hop routing in real-time, and transmits the

results wirelessly to the base-station. The ‘uHealth’ application for the USN

31

platform is implemented based on TinyOS [38]. TinyOS is a small, open source

embedded operating system developed by UC Berkeley. A stand-alone, power-

constrained operating system, TinyOS is application specific and comprises

selected system components and custom components needed for a single

application. In the TinyOS operating system, libraries and applications are written

in nesC [40], which is a structured component-based language. NesC has a C like

syntax, but supports the TinyOS concurrency model, as well as mechanisms for

structuring, naming and linking together software components into robust

network embedded systems. TinyOS defines a number of important concepts that

are expressed in nesC: (1) nesC applications are built out of components with

well-defined, bidirectional interfaces, (2) nesC defines a concurrency model,

based on tasks and hardware event handlers and detects data-races at compile

time. Races are avoided either by accessing a shared state only in tasks, or only

within atomic statements.

The component based architecture and event driven execution model of

TinyOS enables fine-grained power management while minimizing code size,

keeping in view the memory constraints in wireless sensor networks. Our

architecture for software is based on the Active Message communication model.

The Healthcare application consists of several components like ‘uHealth’, ‘ECG’,

‘GenericComm’ and ‘Routing’. These components have specific functions and

services that they provide/use to/from other connected components. Interfacing

techniques of these have been developed, and some of the components used are

shown in Fig. 10. [Paper IV, Paper V]

32

Fig. 10. The uHealth software architecture.

33

The main application level component is ‘uHealth’ which controls the event

handling of various hardware and software events. The component ‘ECG’

samples the ECG signal from the ECG sensor board. ‘GenericComm’ provides

generic packet handling and a basic ‘SendMsg’ and ‘ReceiveMsg’ interface using

TinyOS messages. ‘GenericComm’ will also interface to low level platform

specific TinyOS hardware drivers. The ‘Routing’ component provides the

function of routing data and topology updates to the application. [Paper VI, Paper

VII]

3.5.1 Sampling time for the ECG signal

The default value of the TIME_SCALE (local clocks of nodes) is fixed to a

resolution of 100, indicating a 0.1 second sampling time in TinyOS. In our study,

TIME_SCALE could be modified to 10, corresponding to a sampling time of 0.01

s. Thus, the ECG signal produced by the ECG signal generator passed through the

signal interface circuit and entered the MICAz. The analog ECG signal was

collected at a sampling rate of 125Hz and converted into a digital signal as 12 bit

data. [Paper I]

3.5.2 Packet format

Digitized ECG data can be transmitted to other sensor nodes or the base station

connected to the server PC by an RS-232 serial interface in the IEEE 802.15.4

(Zigbee compatible) standard format. The overall packet structure for the

developed system is composed of 36 bytes, as shown in Table 3. Similar to the

structure of tinyos message structure (TOS_MSG), the header part of the message

is comprised of a destination address, an active message handler, a group ID and

message length. This is followed by a data part of 29 bytes, which is composed of

7 bytes of multi-hop message attributes and data of 22 bytes. A 1 byte channel

was included for indication of types of sensors. The sensor reading is 16 bytes

and the rest are for internal management. The header and data part is followed by

a 2 byte CRC-CCITT, which is used to detect error that may occur in a message

during reception. [Paper I]

34

Table 3. Packet structure for communication between sensor nodes [I, published by

permission of Rare Metal Materials and Engineering].

Header Data Part Payload Data Part (26bytes)

Destination

address

Active

Message

handler

Group ID Message

Length

Mote ID Counter ADC

Channel

ADC

Data

reading

2Bytes 1Byte 1Byte 1Byte 2Bytes 2Bytes 2Bytes 20Bytes

3.5.3 Routing protocol

For constructing our ad-hoc network, a variant of the Destination-Sequenced

Distance Vector (DSDV) algorithm [33] has been used in which the final

destination of data from all nodes is a single base station. The base station

periodically broadcasts its identity, which is used by receiving nodes for updating

its routing information table. They then rebroadcast the new routing update to any

nodes in their range. Thus, a hierarchy of nodes is formed with the base station at

the top level, and all nodes target their data to the nodes which are just above

them. Any change in topology is conveyed to the parent nodes, which can update

their tables.

35

4 Reliable medical data transmission

4.1 Need for reliable medical data transmission

Quality of service is an important part of healthcare application. A critical factor

is reliable medical data transmission, by which medical information can be

monitored correctly by healthcare professionals. In a reliable medical data

transmission system, the receiver side BER may be affected by transmission

channel noise, interference, distortion, attenuation, wireless multipath fading, etc.

A healthcare application has a dynamic duty cycle compared to habitat and

environmental monitoring applications. For example, environmental monitoring

applications need a low duty cycle to collect environmental parameters like

temperature, humidity etc. These applications have lower packet loss

requirements than the high duty cycle applications (healthcare applications). For

many applications, a fundamental problem is providing efficient and reliable data

transmission. In wireless sensor networks, the data transmission is unreliable due

to several factors such as [45]:

– The wireless links can be unreliable, which leads to packet losses.

– The network congestions lead to packet losses as packets are dropped by

intermediate/routing nodes in the routing paths.

– Interference, attenuation and fading are the main factors in unreliable data

transmission

– Channel bit errors during wireless data transmissions.

In particular, ubiquitous health monitoring applications are related to human life.

So, health management, precise measurement and reliable data transmission

methods should be analysed before the designing of healthcare monitoring

applications. So far, a few research studies have been proposed [45, 54] for

reliable data transmission using positive acknowledgement (ACK) and negative

acknowledgement (NACK) methods. The data retransmission approach is the

most commonly used method for recovering packet error and losses. A packet loss

recovery through retransmission is not needed, because medical data is real time

data, as in the case of real-time u-Healthcare monitoring applications. If the data

collected needs to be retransmitted, then packet delay and significant energy

waste may occur because any lost packet will be replaced by the most recent

update in the next transmission. Furthermore, the retransmission of packets will

require extra storage memory at the motes, which is generally limited.

36

4.2 Low-rate wireless personal area networks

Wireless technology for medical applications like IEEE 802.11 based WLANs

was used in this study The IEEE 802.15.4 standard was developed to address a

demand for low power and low-cost in low-rate wireless personal area networks

(LR-WPAN). IEEE 802.15.4 deals with a low data rate, but has a very long

battery life (months or even years) and a very low complexity. The IEEE standard,

802.15.4, defines the physical layer (PHY) and medium access control (MAC)

sub-layer specifications for low-rate wireless personal area networks. IEEE

802.15.4 defines two PHY layers; in the 2.4 GHz and 868/915 MHz bands. The

license free industrial scientific medical (ISM) 2.4 GHz band is available

worldwide, while the ISM 868 MHz and 915 MHz bands are available in Europe

and North America respectively. A total of 27 channels with three different over-

the-air data rates are allocated in 802.15.4: 16 channels with a data rate of 250

kbps in the 2.4 GHz band, 10 channels with a data rate of 40 kb/s in the 915 MHz

band, and 1 channel with a data rate of 20 kb/s in the 868 MHz band [46].

4.3 Error rate analysis for reliability

The bit error rate (BER) is the number of received binary bits that have been

altered due to noise and interference, divided by the total number of transferred

bits during a studied time interval. The BER can be considered as an approximate

estimate of the bit error probability ( bP ).

The packet error rate (PER) is the number of incorrectly transferred data packets

divided by the number of transferred packets. A packet is assumed to be incorrect

if at least one bit is incorrect. Denote that the packet error probability pP , is the

expectation value of PER, which can be expressed for a data packet length of N

bits as

1 (1 )Np bP P= − − . (1)

The BER is often expressed as a function of the normalized signal-to-noise ratio

(SNR) denoted by 0/bE N . The 0/bE N is the ratio of the average energy per

information bit to the noise power spectral density at the receiver input. For

example, in the case of an additive white Gaussian noise (AWGN) channel, the bit

probability of symbol error, eP , as a function of the 0/bE N is given by [47]

37

0

2 be

EP Q

N

= ⋅

, (2)

where ( )Q ⋅ is the error function, bE is energy per bit, and 0N is the noise power

spectral density ratio. The SNR as a function of transmission distance is

expressed as

0

bE RSNRN B

= , (3)

where R is the channel data rate, and B is the channel bandwidth.

Fig. 11 shows the theoretical bit error rates of O-QPSK, 8-PSK and 16-PSK.

The curves present that O-QPSK is better than 8-PSK and 16-PSK. It is seen that

higher-order modulations exhibit higher error-rates in exchange and, however,

they deliver a higher raw data-rate.

Fig. 11. Theoretical bit error rate of O-QPSK, 8-PSK and 16-PSK.

-10 -5 0 5 10 15 2010

-8

10-7

10-6

10-5

10-4

10-3

10-2

10-1

100

Eb/N

0 (dB)

BE

R

O-QPSK

8-PSK16-PSK

38

39

5 Results

5.1 Experimental setup and ECG signal measurement

In our tests, ECG data was obtained from the ECG signal generator via an ECG

interface circuit, instead of an ECG sensor on the patient’s body. The ECG signal

generator generates a synthesized ECG signal with a user-settable heart rate,

signal duration, sampling frequency, QRS waves, etc. Fig. 12 shows the

oscilloscope output of the interface. This analog ECG signal, supplied to the

MICAz wireless sensor node, is nearly identical to a theoretically calculated ECG

signal. The deflections in the tracing of the ECG comprise the Q, R, and S waves,

that represent the ventricular activity of the heart. The interval between two R-

peaks indicates the time between heartbeats. [Paper I]

Fig. 12. (a) ECG data from the ECG signal generator, (b) ECG signal generator,

interface circuit board, MICAz sensor node and base station with an RS-232 interface

[I, published by permission of Rare Metal Materials and Engineering].

5.2 Wearable u-Healthcare monitoring

A wearable system must be comfortable and easy to use for the patient. The

wearable sensor node system was designed to fit perfectly into a shirt. Fig. 13 (a)

shows the wearable smart shirt which consists of conductive fabric electrode pairs

and the wearable sensor node system. The wearable sensor node is attached on the

wearer’s chest as shown in Fig. 13 (b). The smart shirt provides an extremely

(a) (b)

40

versatile framework for the incorporation of sensing, monitoring and information

processing devices. Moreover, the smart shirt can be used in a variety of

applications such as battlefield, public safety, health monitoring and sports and

fitness, among others [48, 49]. [Paper III]

Fig. 13. Smart shirt with an integrated wearable sensor node [III, published by

permission of Elsevier].

The monitoring of vital signs during the wearing of the designed smart shirt was

tested by real time monitoring of the ECG and acceleration signals of a wearer.

For the simultaneous monitoring of physiological ECG data and the activity of the

smart shirt wearer, a treadmill was used. The speed of the treadmill was

controlled at a speed of 5km/h for walking and 8km/h for running. The system

test on the treadmill was performed with the wearer standing motionless (resting),

walking and running on the treadmill as shown in Fig. 14. Fig. 15 shows the test

results for walking, running and resting after the exercise.

41

Fig. 14. ECG signal variations during walking, running and resting on the treadmill [III,

published by permission of Elsevier].

The raw acceleration signals from a 3-axes accelerometer in the wearable sensor

node of a smart shirt during the exercise of a person on a treadmill were measured

in a wireless sensor network, as shown in Fig. 15. Acceleration signals provide

valuable information for the wearer’s activity classification, such as walking,

running and resting (standing). [Paper III]

42

Fig. 15. Acceleration signals form a three-axes accelerometer during walking, running

and resting of a person on a treadmill [III, published by permission of Elsevier].

5.3 ECG data on the client display

Attached to the ECG signal generator and interface circuit, the MICAz sensor

node collected the ECG signal at a sampling rate of 125 Hz. Thus, the client PC

also collected data at a 125 Hz sampling rate. Fig. 16 shows such ECG data,

sampled at a 12 bit resolution, in the terminal PC’s GUI. A comparison of the

ECG data obtained from the ECG signal generator via the interface circuit in Fig.

12 (b) and the ECG data in the terminal PC’s GUI in Fig. 16, indicates that the

sensor data were accurately transferred to the terminal PC via wireless IEEE

802.15.4 communication between the sensor node and the base station. The same

is true concerning the transmission of data via TCP/IP between the server and the

terminal PC in real time. [Paper I]

43

Fig. 16. ECG data in the terminal PC’s GUI [IV, published by permission of IMCS2006].

5.4 PDA display

A PDA (Personal Digital Assistant) was also used as a client terminal. Fig. 17 (a)

shows a PDA GUI snapshot of the PDA emulator, which displays real-time ECG

data sent from the server. The graphic user interface was developed to enable the

PDA terminal to display the recorded ECG data in real time on the PDA emulator

and PDA screen, as shown in Fig.17 (a) and (b). [Paper I]

Fig. 17. Screenshot from a PDA emulator (a) and PDA GUI (b) [I, published by

permission of Rare Metal Materials and Engineering].

(a) (b)

44

5.5 Reliability analysis

The parameters used in the experiments are summarized in Table 4. In every

measurement, a total of 500 packets were sent to the base-station with a fixed

packet size of 46 bytes in each packet. The measurements for PER [50, 51] were

made from a 1 to 100 meter distance between sensor nodes and base-station. The

system operates at a 2.4 GHz band and offers a bit rate of 250 kbps [52–54].

Table 4. Table 1 Reliability analysis parameters.

Parameters Value

Number of sent packets 500 (ECG data)

Number of packet size 46 bytes

Number of nodes 1~5

Topology configuration Randomized

Channel data rate 250 kbps

Ground distance 1~100 m

RF frequency 2.45 GHz

RF power −25~0 dBm

5.5.1 Reliability test in function of distance between sensor nodes

For empirical measurements, three scenarios were considered, such as; (1) the

distance between sensor nodes and the base-station, (2) the number of the

deployed sensor nodes in a network, (3) data transmission by different time

intervals. They were measured to analyse the reliability of the technology in low-

rate wireless health monitoring applications. The PER is calculated from the

received data that is analyzed to evaluate performance according to the scenarios.

All of the experiments were conducted using ECG sensor data in an indoor

environment. The first experiment is about the measurement of the PER, as a

function of distance, in a short communication range. The experimental setup was

built according to Table 4. The experiment scenario of different distances between

sensor node and base-station is shown in Fig. 18. For this experiment, a distance

of 1–100 meters was applied. The distances between sensor node and base-station

were varied to measure the PER.

45

Fig. 18. Measurement setup of network reliability of the distance between sensor

nodes and base-station.

In Fig. 19, the measured PER is shown as a function of the distance between

sensor node and base-station. The average packet loss rate was obtained by setting

RF signal strength to 0 dBm and 500 packets were sent to the base-station. The

PER was calculated using the received packet loss rate. The x-axis shows the

distance between sensor node and base-station. The PER increased when the

distance between sensor node and base-station was increased. The result shows

that the minimum and maximum packet losses are not much different. However, it

was found that the PER in the function of the distance jumped up suddenly after

50 meters and was almost zero when node 1 and node 2 were deployed at the

same time in a network.

The results of the measured PER versus SNR of different distance values

such as D=10, 60 and 90 meter are shown in Fig. 20. The PER is 0.0033 where

SNR is 8 dB.

46

Fig. 19. Measured PER in the function of the distance between sensor node and base-

station.

10 20 30 40 50 60 70 80 900

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Distance [m]

PE

R

Distance

47

Fig. 20. Measured PER versus SNR for different distance values, D=10, 60 and 90

meter.

5.5.2 Reliability test in the function of the number of the sensor nodes

The second set of experiments was to measure the PER as a function of the

number of the sensor nodes, as shown in Fig. 21. In this experiment, a

randomized topology was used and four sensor nodes and one base-station were

deployed in a network. It was assumed that each sensor node is about 20 meter

away from a base-station and RF signal strength was set to −5 dBm.

The PER, by deployment of the number of sensor nodes, is shown in Fig. 22.

The measured PER is almost 0 when two sensor nodes are deployed in a network.

However, it was found that the PER increases from the third sensor node. X axis

represents the number of sensor nodes deployed in a network and the curves

represent each sensor node with ID numbers. The PER is almost 0 (100% success

rate) when node 1 and node 2 deployed at the same time in a network. The PERs

of each sensor nodes are not much different as shown in Table 5. However, the

-15 -10 -5 0 5 1010

-3

10-2

10-1

100

Eb/N0 (dB)

BE

R

D=10

D=60D=90

48

PER increases with an increase in the number of nodes, due to collision of the

packets.

Fig. 21. Measurement setup for deployment of the number of sensor nodes.

Fig. 22. Measured PER of number of transmission nodes.

1 1.5 2 2.5 3 3.5 4 4.5 50

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Number of nodes

PE

R

1 node2 node3 node4 node5 node

49

Table 5. End-to-end success rate of number of transmission nodes.

Node ID Number of nodes Success rate (%)

1 1 99.56

2 1 99.32

2 99.44

3 1 72.4

2 71.64

3 72.52

4 1 59.32

2 50.52

3 59.96

4 43.8

5 1 39.72

2 47.76

3 41.56

4 35.56

5 37.96

5.5.3 Reliability test for different time intervals

For the last experiment, a scenario was considered for different time intervals on

different sensors using ECG and temperature sensors. Fig. 23 shows the results of

the PER for different sampling time intervals, when ECG data packets were sent

to the base-station at a rate of every 0.01 sec. On the other hand, the time interval

of the temperature sensor was set to 1 sec when sending the data to the base-

station. The PER of the temperature sensor is almost 0, when seven sensor nodes

are deployed in a network. On the other hand, in the ECG experiment, the PER

increased sharply for three or more sensor nodes in the network. As the number of

ECG sensor nodes increases, the PER also increases because of contention among

multiple sensor nodes and data packet overflow.

50

Fig. 23. Measured PER for different sampling time intervals, ECG (0.01 sec) and

temperature sensors (1 sec).

1 2 3 4 5 6 7 8 9 100

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Number of nodes

PE

R

ECG : 0.01s

Temperature : 1s

51

6 Discussion

The main objective of this thesis project was to study the use of wireless sensor

networks in a healthcare system as a low power consumption, low cost and ad-

hoc network. This thesis project focused on two specific research topics: 1) the

development of a distributed healthcare system for vital signs monitoring using

wireless sensor network devices and 2) the empirical analysis of the reliability of

such low-rate wireless technology ubiquitous health monitoring applications.

6.1 Multi-hop network for wireless body area network

This study proposes a ubiquitous healthcare system for vital signs monitoring in

Chapter 3. Since remote sensing techniques are to use wireless health monitoring,

this thesis project sought to provide a reliable ECG signal for professional

healthcare monitoring. While technological advancements have resulted in

changes in many aspects of daily life, there is still a significant gap between the

existing solutions and the needs in the medical field.

This thesis project is an endeavour to suggest a solution for utilizing wireless

technologies and to provide a remote health monitoring system designed for the

medical environment. This system provides continuous and unobtrusive

monitoring for patient care in order to capture transient behaviour. This system

consists of three major components; a body sensor network node, local processing

unit and a central server. Body sensor network nodes are attached on the body of a

patient based on wireless sensor networks. The gathered data is transmitted

wirelessly over radio to the receiving base-station. Recent research has also

focused on the development of wireless sensor networks and health monitoring

systems. For example, a number of wearable systems have been proposed with

integrated wireless transmission and local processing. The vital signs monitoring

systems worked fine in single-hop (ad-hoc network) data communications, but

revealed some scheduling problems in a multi-hop network. One of the more

difficult constraints was that several wireless biological data sensors, including

ECG, SpO2, accelerometers etc., reached the limits of using a multi-hop

environment. Our best guess is that the star topology and one hop connection to

the base-station is preferred for body sensor networks. The star topology network

is a popular option for a wireless body area network where a base-station

(gateway node) gathers the sensing information from sensor nodes.

52

6.2 Electrocardiogram and accelerometer sensor

The electrocardiogram (ECG) is currently one of the most widely prescribed

medical diagnostic procedures. ECG is usually used in a diagnostic mode and a

monitoring mode. In the monitoring mode, the frequency filter band is set at

0.5Hz to 40Hz. In the diagnostic mode, the high-pass filter is set at 0.05Hz and

the low-pass filter at 150Hz. The ECG sampling rate was 125Hz in this thesis

project, which allowed the bandwidth of the ECG (0.05 to 123Hz) to be captured.

The ECG signal was amplified with a gain of 200 (24.8dB). In TinyOS, the

default value of the sampling rate is fixed at a resolution of 100, indicating 0.1

seconds (10Hz). In this thesis project, we tried to apply a resolution of 4,

indicating 250Hz, with which an ECG signal cannot be clearly obtained.

Therefore, we were interested in obtaining better performances in terms of the

sampling rate, because according to well-known sampling theories, the sampling

rate must be more than two times the highest frequency in the ECG signal.

However, a bandpass-filter from 0.05 to 123Hz was used to limit analog-to-digital

converter (ADC) saturation and for anti-aliasing when the analog outputs of the

playback unit were digitized. Perhaps there are not very many signal components

above 62Hz. The ADC was unipolar, with a 12-bit resolution over a 5mV range.

Sample values thus range from 0 to 2047 inclusive, with a value of 1024

corresponding to zero volts. So, to convert a sample back into physical units

(millivolts), the ADC Zero 1024) is subtracted from the sample value and then

divided by the gain.

Accelerometer sensors have been widely used to monitor human posture and

movement parameters. We attached on the chest a three-axes accelerometer to

measure body acceleration as the reference signal. The subject was waking and

running on the treadmill. Acceleration signals provide valuable information, such

as walking, running and resting (standing) in our experiments. A three-axes

acceleration signal, together with an ECG signal which included motion artefacts,

are measured as shown in Fig. 24. As seen in Fig. 24, Y-axis acceleration data is

the most sensitive to the motion artefact contained in the ECG signal, even though

X-axis acceleration data varies much more widely in our experimental result.

53

Fig. 24. ECG signal with motion artefact and three-axes accelerometer signals.

6.3 Wearable system for ubiquitous healthcare

A wearable system for ubiquitous healthcare should provide stable signals during

the patient’s daily activities. Resistance to motion artefact remains a significant

challenge for wearable systems. Efforts are ongoing to eliminate artefacts by

improving hardware design and applying signal processing algorithms. Various

studies have been devoted to the development of ring, watch, band, or garment

based wearable monitors, to improve the patient’s tolerance. In our smart shirts, it

is still not very clear how to reduce the motion artefacts. Further improvements in

artefact reduction will remain critical for improvements in the wearability, patient

acceptance and medical value of existing wearable systems. These contributions

are limited by factors including hardware design constraints associated with the

sensing of human status, power consumption, data storage and signal processing.

54

55

7 Conclusion

The main objective of this thesis project is to implement a wireless vital signs

monitoring system for measuring ECG from a patient in the home environment.

The research focuses on two specific research objectives: 1) the development of a

distributed healthcare system for vital signs monitoring using wireless sensor

network devices and 2) a practical test and performance evaluation of the

reliability for such low-rate wireless technology in ubiquitous health monitoring

applications.

The first section of the thesis describes the design and implementation of a

ubiquitous healthcare system with tiny devices for the home care of elderly

persons. The system has been designed using a commercially available wireless

sensor network platform (MICAz, Crossbow Technology Inc., USA). The system

is capable of obtaining physiological data from a sensor attached to the patient’s

body and transfers the data wirelessly to a remote base-station in an ad-hoc

network following the IEEE 802.15.4 standard for wireless personal area

networks. The system concept can be used for ad-hoc routing of vital signs

monitoring signals to base-stations within the hospital premises, as well as in

applications that require monitoring from the patient’s home. A wireless health

monitoring system has potential to provide an improved service to patients,

doctors and caregivers through continuous health monitoring. In addition, a smart

shirt with wireless sensor network compatibility is designed and fabricated for the

continuous monitoring of physiological ECG signals and physical activity signals

from an accelerometer simultaneously. To collect physiological ECG data and

activity of the smart shirt wearer simultaneously, the performance test is done on

a treadmill by a wearer who is resting, walking and running on the treadmill at

various speeds. Thus, the developed smart shirt system can be capable of

improving the accuracy of a patient's diagnosis by continuous monitoring with a

non-invasive, comfortable and convenient shirt to wear in a wireless sensor

network environment.

In the experimental tests, ECG data was displayed on a Graphical User

Interface (GUI) by transferring the data to a PDA or a PC through a base-station

connected via the UART. To prove the capability of the wireless sensor network

platform for ubiquitous healthcare applications, the performance of the TinyOS-

based sensor network platform was tested with positive results.

In the second section, the thesis project produced an empirical analysis of the

reliability of low-rate wireless ubiquitous healthcare monitoring applications.

56

Three scenarios of performance studies were considered for empirical

measurement: 1) the effects of the distance between sensor nodes and base-station,

2) the number of the deployed sensor nodes in a network and 3) data transmission

by different time intervals. They were measured to analyse the reliability of the

technology in low-rate wireless u-Healthcare monitoring applications.

In this thesis project, three scenarios were considered for studying the

reliability of low-rate wireless u-Healthcare monitoring. In medical applications

using ECG data, the PER is higher (about 10−4) than the environment data (10−6)

coming from a temperature sensor. Moreover, reliability based on wireless sensor

networks, in particular, low-rate wireless medical monitoring applications, the

distance between sensor node and base-station, the number of sensor nodes in the

network and different time intervals affect reliability performance, as shown in

experimental results. It was found that the PER of distance jumped up suddenly

after 50 meters and was almost zero when node 1 and node 2 deployed at the

same time in a network. To apply low-rate wireless technology for wireless

ubiquitous health monitoring applications, the sampling rate, number of nodes

and data packet sizes, as well as reliable medical devices, are closely related to

the overall reliability of the system.

57

References

1. Culler D, Estrin D & Srivastava M (2004) Overview of sensor networks. Computer 37: 41–49.

2. Akyildiz IF, Su W, Sankarasubramaniam Y & Cayirci E (2002) A survey on sensor networks. IEEE Communications Magazine 40(8): 102–114.

3. Szewczyk R, Osterweil E, Polastre J, Hamilton M, Mainwaring A & Estrin D (2004) Habitat monitoring with sensor networks. Communications of the ACM 47(6): 34–40.

4. Kim S (2005) Wireless sensor networks for structural health monitoring. Master thesis, University of California at Berkeley.

5. Puccinelli D & Haenggi M (2005) Wireless sensor networks-applications and challenges of ubiquitous sensing. IEEE Circuits and Systems Magazine 5: 19–29.

6. Otto C, Milenkovic A, Sanders A & Jovanov E (2006) System architecture of a wireless body area sensor network for ubiquitous health monitoring. Journal of Mobile Multimedia 1(4): 307–326.

7. Makki SK, Li XY, Pissinou N, Makki S, Karimi M & Makki K (2008) Sensor and ad-hoc networks: theoretical and algorithmic aspects. Lecture Notes in Electrical Engineering, Springer.

8. Varshney U & Sneha S (2006) Patient monitoring using ad hoc wireless networks: reliability and power management. IEEE Communications Magazine 44(4): 49–55.

9. Zhou G, Lu J, Wan CY, Yarvis MD & Stankovic JA (2008) BodyQoS: adaptive and radio-agnostic QoS for body sensor networks. Proceedings of the IEEE INFOCOM: 565–573.

10. Liu Y, Elhanany I & Qi H (2005) An energy-efficient QoS-aware media access control protocol for wireless sensor networks. Proceedings of the IEEE MASS 2005, Washington DC, USA.

11. Zhao Q & Tong L (2003) QoS specific medium access control for wireless sensor networks with fading, In Proceedings of the 8th International Workshop on Signal Processing for Space Communications (SPSC ’03), Catania, Italy.

12. Golmie N, Cypher D & Rebala D (2005) Performance analysis of low rate wireless technologies for medical applications. Computer Communications 28(10): 1255–1275.

13. Halteren VA, Bults R, Wac K, Dokovsky N, Koprinkov G, Widya I, Konstantas D & Jones V (2004) Wireless body area networks for healthcare: the MobiHealth project. Studies in Health Technology and Informatics 108: 181–93.

14. Jovanov E, Milenkovic A, Otto C & de Groen PC (2005) A wireless body area network of intelligent motion sensors for computer assisted physical rehabilitation. Journal of NeuroEngineering and Rehabilitation 2(6).

15. Li H-B, Takizawa K-I, Zhen Bin & Kohno R (2007) Body area network and its standardization at IEEE 802.15.MBAN. In Mobile and Wireless Communications Summit. 16th IST, Budapest, Hungary: 1–5.

16. Zasowski T (2007) A system concept for Ultra Wideband (UWB) body area networks. PhD thesis, ETH Zurich.

58

17. Schmidt R, Norgall T, Morsdorf J, Bernhard J & Grun T (2002) Body area network BAN-a key infrastructure element for patient-centered medical applications. Biomed Tech (Berl) 47: 365–368.

18. Rosener D (2007) Introduction to Bluetooth engineering. John Wiley & Sons Inc. 19. Zigbee Alliance. URI: http://www.zigbee.org. 20. Ng JWP, Lo BPL, Wells O, Sloman M, Toumazou C, Peters N, Darzi A & Yang GZ

(2004) Ubiquitous monitoring environment for wearable and implantable sensor. International Conference on Ubiquitous Computing (Ubicomp).

21. Malan D, Fulford-Jones T, Welsh M & Moulton S (2004) CodeBlue: An ad hoc sensor network infrastructure for emergency medical care. In Proceedings of the Mobisys 2004 Workshop on Applications of Mobile Embedded Systems (WAMES 2004), Boston, MA.

22. Landsiedel O, Wehrle K, Gotz S (2005) Accurate prediction of power consumption in sensor networks. Proceedings of the Second IEEE Workshop on Embedded Networked Sensors (EmNetSII), Sydney, Australia.

23. XBOW Mote Specifications. URI: http://www.xbow.com. 24. Feeney LM (2001) An energy-consumption model for performance analysis of routing

protocols for mobile ad hoc networks. Mobile Networks and Applications 3(6): 239–249.

25. Younis M, Youssef M & Arisha K (2002) Energy-aware routing in cluster-based sensor networks. Proceedings of the 10th IEEE Int’l Symp. on Modeling, Analysis & Simulation of Computer & Telecommunications Systems (MASCOTS ’02).

26. Ye W, Heidemann J & Estrin D (2002) An energy-efficient MAC protocol for wireless sensor networks. Proceeding of the INFOCOM, New York, USA: 1567–1576.

27. Dam TV & Langendoen K (2003) An adaptive energy-efficient MAC protocol for wireless sensor networks. Proceedings of the First ACM Conference on Embedded Networked Sensors Systems, New York, USA: 1567–1576.

28. Lin P, Qiao C & Wang X (2004) Medium access control with a dynamic duty cycle for sensor network. Proceedings of the IEEE Wireless Communications and Networking Conference (WCNC) 3: 1534–1539.

29. Jamieson K, Balakrishnan H & Tay YC (2006) Sift: a MAC protocol for event-driven wireless sensor networks. Proceedings of Third European Workshop on Wireless Sensor Networks (EWSN): 260–275.

30. Lu G, Krishnamachari B & Raghavendra CS (2004) An adaptive energy-efficient and low-latency MAC for data gathering in sensor networks. Proceedings of the IEEE International Workshop on Algorithms for Wireless, Mobile, Ad Hoc and Sensor Networks, Santa Fe, New Mexico.

31. Polastre J, Hill J & Culler D (2004) Versatile low power media access for wireless sensor networks. Proceedings of 2nd ACM Conference on Embedded Network Sensor System (SenSys), Baltimore, USA: 95–107.

32. Malesci U (2005) A measurement-based analysis of the interaction among MAC, network and application layers in wireless sensor networks. Master Thesis, MIT.

59

33. Perkins CE & Bhagwat P (1994) Highly dynamic Destination-Sequenced Distance-Vector Routing (DSDV) for mobile computers. Proceedings of SIGCOMM ’94 Conference on Communication Architecture, Protocols and Applications, London, UK: 234–244.

34. Pakzad SN, Fenves GL, Kim S & Culler D (2008) Design and implementation of scalable wireless sensor network for structural monitoring. Journal of Infrastructure Systems 14(1): 89–101.

35. Callaway E, Gorday P, Hester L, Gutierrez JA, Naeve M, Heile B & Bahl V (2002) Home networking with IEEE 802.15.4: a developing standard for low-rate wireless personal area networks. IEEE Communications Magazine 40(8): 70–77.

36. Gutierrez JA, Naeve M, Callaway E, Bourgeois M, Mitter V & Heile B (2001) IEEE 802.15.4: a developing standard for low-power low cost wireless personal area networks. IEEE Network 15(5): 12–19.

37. IEEE 802.15.4–2003: IEEE Standard for Information Technology - Part 15.4: Wireless Medium Access Control and Physical Layer specifications for Low Rate Wireless Personal Area Networks.

38. Levis P, Madden S, Polastre J, Szewczyk R, Whitehouse K, Woo A, Gay D, Hill J, Welsh M, Brewer E & Culler D (2004) TinyOS: an operating Ssystem for sensor networks. Ambient Intelligence: 115–148.

39. Lee Y-D, Jung S-J, Seo Y-S & Chung W-Y (2008) Measurement of motion activity during ambulatory using pulse oximeter and triaxial accelerometer. Proceedings of the International Conference on Convergence and Hybrid Information Technology, Busan, Korea: 436–441.

40. Lee Y-D, Jung S-J, Seo Y-S & Chung W-Y (2008) Design of a low-power consumption wearable reflectance oximetry for ubiquitous healthcare system. Proceedings of the International Conference on Control, Automation and Systems, Seoul, Korea: 526–529.

41. Bourke AK, O’Brien JV & Lyons GM (2007) Evaluation of a threshold-based tri-axial accelerometer fall detection algorithm. Gait & Posture 26: 194–199.

42. Lindemann U, Hock A, Stuber M, Keck W & Becker C (2005) Evaluation of a fall detector based on accelerometers: A pilot study. Medical and Biological Engineering and Computing 43(5): 548–551.

43. Chen J, Karric Kwong, Chang D, Luk J & Bajcsy R (2006) Wearable sensors for reliable fall detection. IEEE Engineering in Medicine and Biology Society. 3351–3554.

44. Gay D, Levis P, Behren R, Welsh M, Brewer E & Culler D (2003) The nesC language: a holistic approach to networked embedded systems. Proceedings of Programming Language Design and Implementation, San Diego, USA.

45. Shamanna RG (2006) Study of reliable data communication in wireless sensor networks. Master Thesis. Louisiana State University.

46. IEEE 802.15.4 Std: Wireless Medium Access Control (MAC) and Physical Layer (PHY) Specifications for Low-Rate Wireless Personal Area Networks (LR-WPANs). IEEE Computer Society Press, 2003: 1–679.

60

47. Tamilselvan GM & Sahnmugam A (2009) Probability analysis of channel collision between IEEE 802.15.4 and IEEE 802.11b using Qualnet simulation for various topologies. International Journal of Computer Theory and Engineering 1(1): 59–64.

48. Park S, Gopalsamy C, Rajamanickam R & Jayaraman S (1999) The wearable motherboard: an information infrastructure or sensate liner for medical applications, Stud. Health Technol. Informatics 62: 252–258.

49. Park S & Jayaraman S (2003) Enhancing the quality of life through wearable technology. IEEE Eng. Med. Biol. Mag 22(3): 41–48.

50. Petrova M, Riihijarvi J, Mahonen P & Labella S (2006) Performance study of IEEE 802.15.4 using measurements and simulations. Wireless Communications and Networking Conference (WCNC 2006) 1: 487–492.

51. Shin SY, Park HS & Kwon WH (2007) Mutual interference analysis of IEEE 802.15.4 and IEEE 802.11b. Computer Networks 51: 3338–3353.

52. Suhonen J, Hamalainen TD & Hannikainen M (2009) Availability and end-to-end reliability in low duty cycle multihop wireless sensor networks. Sensors 9: 2088–2116.

53. Shuaib K, Alnuaimi M, Boulmalf M, Jawhar I, Sallabi F & Lakas A (2007) Performance evaluation of IEEE 802.15.4: experimental and simulation results. Journal of Communications 2(4): 29–37.

54. Ferrari G, Medagliani P, Piazza SD & Martalo M (2007) Wireless sensor networks: performance analysis in indoor scenarios. EURASIP Journal on Wireless Communications and Networking 2007(1): 41.

61

Original Papers

I Lee Y-D, Lee D-S, Chung W-Y & Myllylä R (2006) A wireless sensor network platform for elderly persons at home. Rare metal materials and engineering 35(3): 95–99.

II Amit P, Lee Y-D & Chung W-Y (2008) Triaxial MEMS accelerometer for activity monitoring of elderly person. Sensor Letters 6(6): 1054–1058.

III Lee Y-D & Chung W-Y (2009) Wireless sensor network based wearable smart shirt for ubiquitous health and activity monitoring. Sensors and Actuators B: Chemical 140(2): 390–395.

IV Walia G, Lee Y-D, Chung W-Y & Myllylä R (2006) Distributed healthcare monitoring in an ad-hoc network using a wireless sensor network. Proceedings of 11th International Meeting on Chemical Sensors, Brescia, Italy.

V Lee Y-D, Lee D-S, Walia G, Myllylä R & Chung W-Y (2006) Query based duplex vital signal monitoring system using wireless sensor network for ubiquitous healthcare. Proceedings of World Congress on Medical Physics and Biomedical Engineering, Seoul, Korea: 396–399.

VI Lee Y-D, Lee D-S, Walia G, Alasaarela E & Chung W-Y (2007) Design and evaluation of query supported healthcare information system using wireless sensor ad-hoc network. Proceedings of International Conference on Convergence Information Technology, Gyeongju, Korea: 997–1002.

VII Lee Y-D, Alasaarela E & Chung W-Y (2007) A wireless health monitoring system in multi-hop body sensor networks. Proceedings of 2nd International Symposium on Medical Information and Communication Technology (ISMICT’'07), Oulu, Finland.

VIII Chung W-Y, Lee Y-D & Jung S-J (2008) A wireless sensor network compatible wearable u-healthcare monitoring system using integrated ECG, accelerometer and SpO2. Proceedings of 30th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC 2008), Vancouver, Canada: 1529–1532.

Reprinted with permission from Rare Metal Materials and Engineering (I), Sensor

Letters (II), Elsevier (III), IEEE (VI, VIII), Springer (V), IMCS 11 (IV) and

ISMICT2007 (VII).

Original publications are not included in the electronic version of the dissertation.

62

A C T A U N I V E R S I T A T I S O U L U E N S I S

Book orders:Granum: Virtual book storehttp://granum.uta.fi/granum/

S E R I E S C T E C H N I C A

355. Ronkainen, Sami (2010) Designing for ultra-mobile interaction : experiences and amethod

356. Loikkanen, Mikko (2010) Design and compensation of high performance class ABamplifiers

357. Puikko, Janne (2010) An exact management method for demand driven, industrialoperations

358. Ylioinas, Jari (2010) Iterative detection, decoding, and channel estimation inMIMO-OFDM

359. Tervonen, Pekka (2010) Integrated ESSQ management : as a part of excellentoperational and business management—a framework, integration and maturity

360. Putkonen, Ari (2010) Macroergonomic approach applied to work systemmodelling in product development contexts

361. Ou, Zhonghong (2010) Structured peer-to-peer networks: Hierarchicalarchitecture and performance evaluation

362. Sahlman, Kari (2010) Elements of strategic technology management

363. Isokangas, Ari (2010) Analysis and management of wood room

364. Väänänen, Mirja (2010) Communication in high technology product developmentprojects : project personnel’s viewpoint for improvement

365. Korhonen, Esa (2010) On-chip testing of A/D and D/A converters : static linearitytesting without statistically known stimulus

366. Palukuru, Vamsi Krishna (2010) Electrically tunable microwave devices using BST-LTCC thick films

367. Saarenpää, Ensio (2010) Good construction quality : the realisation of goodconstruction quality in Finnish construction regulations

368. Vartiainen, Johanna (2010) Concentrated signal extraction using consecutivemean excision algorithms

369. Nousiainen, Olli (2010) Characterization of second-level lead-free BGAinterconnections in thermomechanically loaded LTCC/PWB assemblies

370. Taskila, Sanna (2010) Improved enrichment cultivation of selected food-contaminating bacteria

371. Haapala, Antti (2010) Paper machine white water treatment in channel flow :integration of passive deaeration and selective flotation

C373etukansi.kesken.fm Page 2 Tuesday, November 16, 2010 4:02 PM

ABCDEFG

UNIVERS ITY OF OULU P.O.B . 7500 F I -90014 UNIVERS ITY OF OULU F INLAND

A C T A U N I V E R S I T A T I S O U L U E N S I S

S E R I E S E D I T O R S

SCIENTIAE RERUM NATURALIUM

HUMANIORA

TECHNICA

MEDICA

SCIENTIAE RERUM SOCIALIUM

SCRIPTA ACADEMICA

OECONOMICA

EDITOR IN CHIEF

PUBLICATIONS EDITOR

Professor Mikko Siponen

University Lecturer Elise Kärkkäinen

Professor Hannu Heusala

Professor Olli Vuolteenaho

Senior Researcher Eila Estola

Information officer Tiina Pistokoski

University Lecturer Seppo Eriksson

Professor Olli Vuolteenaho

Publications Editor Kirsti Nurkkala

ISBN 978-951-42-6387-3 (Paperback)ISBN 978-951-42-6388-0 (PDF)ISSN 0355-3213 (Print)ISSN 1796-2226 (Online)

U N I V E R S I TAT I S O U L U E N S I SACTAC

TECHNICA

U N I V E R S I TAT I S O U L U E N S I SACTAC

TECHNICA

OULU 2010

C 373

Young-Dong Lee

WIRELESS VITAL SIGNS MONITORING SYSTEM FOR UBIQUITOUS HEALTHCARE WITH PRACTICAL TESTS AND RELIABILITY ANALYSIS

UNIVERSITY OF OULU,FACULTY OF TECHNOLOGY,DEPARTMENT OF ELECTRICAL AND INFORMATION ENGINEERING

C 373

ACTA

Young-Dong Lee

C373etukansi.kesken.fm Page 1 Tuesday, November 16, 2010 4:02 PM