RESURE

Wireless Sensor Network for Health–Care Monitoring

Enrique A. Vargas Cabral, Vicente A. Gonzalez and Jean A. Guevara

Departamento de Electr

onica e Inform

atica, Facultad de Ciencias y Tecnolog

ıa

Univ. Cat

olica, Campus Univ.–Tte. Cantalupi y Villalon, Asunci

on, -Paraguay

Keywords:

Wireless Personal Area Network (WPAN), IEEE 802.15.4, Zigbee, Medical Care, Physiological signals.

Abstract:

This work presents a sensor system to monitoring a patient vital signs remotely. The proposed sensor network

architecture is based on IEEE 802.15.4 PAN technology with start topology. An experimental set-up was

implemented including the development of some biological sensors. In the implementation stage we use Xbee

modules, the remote sensors was conﬁgured as RFD devices and the network coordinator as FFD device. As

central device a smart phone was used. From the results, we conclude that the IEEE 802.15.4 technology is

appropriate for medical WPAN sensor network implementation.

1 INTRODUCTION

The World Health Organization (WHO) estimates that

by the year 2025, more than 1,000 million people will

be older than 60 years (WHO, 2004), in other words,

the average age of the world-wide population is in-

creasing. Normally, the aging comes accompanied

by non-transmissible chronic diseases, such as car-

diovascular disease and others that need a continuous

monitoring. This scenario presents many challenges

to modern societies; one of them is the design of sen-

sor systems that can continuously record physiolog-

ical variables, from these variables the health of the

people can be inferred (Woodward and Rasid, 2003).

Nowadays, the technological advances in micro-

electronics and instrumentation ease the development

of sophisticated, small size, low power consumption

and relative low cost monitoring systems. In addi-

tion, the WPAN technology allows easy and fast in-

terconnection of different kind of elements, i.e. ear-

pieces for mobile phones, video, camera, audio re-

producer, computers, etc. Therefore, the new tech-

nologies offer the possibility of transmitting physio-

logical signs for processing or visualization devices

(PCs, PDAs, screen, etc.) or to processing center

without limiting the patient mobility (Rodr

ıguez, J.

et al., 2005). Due to medical and economical rea-

sons, shorter periods of inpatient stay are desirable.

Patients’ monitoring at home environments is one of

the most attractive areas for WPAN applications (Karl

and Willig, 2005), because this technology can en-

able seamless connectivity in hospital and home en-

vironment. A WPAN central device can be imple-

mented as a wristwatch or a small belt-worn box.The

device could be able to communicate with the hos-

pital monitoring system using the home network ac-

cess points via wireless indoor communication sys-

tems like WLAN, DECT or Bluetooth. WPAN moni-

toring is also adequate for patients with chronic dis-

eases such as diabetes, asthma and cardiovascular dis-

eases. The employing of WPAN in the patient’ home

is useful for rehabilitation and post operational care.

A central WPAN device with modules for accessing

global wireless networks like GSM or UMTS can pro-

vide potentially worldwide mobility for WPAN users

(Istepanian, R. et al., 2001). Summarizing, we can

say that quality of life can be signiﬁcantly improved.

2 REQUERIMENTS OF

A MEDICAL SENSOR

NETWORKS

A medical sensor network design greatly depends on

the application and deployment environment. A sen-

sor network designed for ad hoc deployment in an

emergency site has very different requirements than

one deployed permanently in a hospital. For exam-

ple, the latter can make use of ﬁxed, wall powered

gateway nodes, which provide access to a wired net-

work infrastructure. In general, we can identify seve-

101

A. Vargas Cabral E., A. Gonzalez V. and A. Guevara J. (2008).

RESURE - Wireless Sensor Network for Health–Care Monitoring.

In Proceedings of the First International Conference on Biomedical Electronics and Devices, pages 101-106

DOI: 10.5220/0001050101010106

 SciTePress

ral characteristics that nearly all medical sensor net-

works should have (Shnayder, V. et al., 2005).

• Scalability: is the capability of a system to in-

crease performance when new nodes are added.

• Wearable Sensor Networks (WSN):Medical ap-

plications require very small, lightweight, and

wearable sensors. Power consumption in a WSN

is important since most or all devices must be bat-

tery powered. Replacing or recharging in short in-

tervals will be impractical, so power consumption

is of signiﬁcant concern.

• Reliable communications:In medical settings, a

great emphasis is placed on data availability, al-

though intermittent packet loss due to interference

may be acceptable. However, persistent packet

loss would be a problem. Depending on the used

sensors, sampling rates may vary between 1 to

1000 Hz or more, placing heavy demands on the

wireless channel.

• Device mobility:Both patients and caregivers are

mobile, requiring that the communication layer

adapt rapidly to changes in link quality. For ex-

ample, if a multihop routing protocol is in use,

it should quickly ﬁnd new routes when a doctor

moves from room to room during his rounds.

• Security:The security in wireless networks is al-

ways of great importance. In sensor networks, it is

especially important to have integrity and authen-

tication. Integrity means that data should not be

altered or destroyed in its way from the sender to

the receiver and the authentication should ensure

the identity of the sender and the receiver.

• Association:The service used to establish a de-

vice’s membership in a WPAN. How a WPAN

adds a new node is of interest in this kind of net-

works. As well as how one sensor is associated

with the correct output at the central monitor side.

• Cost:To be able to compete in the international

market it is essential that the components be at

the lowest possible price. This is most important

when the product shall be mass-produced.

3 SHORT RANGE WIRELESS

TECHNOLOGIES

Based on the above mentioned requirements of med-

ical WPAN systems, following the technical charac-

teristics of some wireless technologies are presented,

which according to the bibliographical review and our

criteria, can be used to design a medical WPAN.

3.1 IEEE 802.15.1 / Bluetooth

The IEEE 802.15.1 standard is derived from the Blue-

tooth speciﬁcation (version 1.1). In fact, the IEEE

standard has added two clauses to the existing speciﬁ-

cation; WPAN architecture overview and Service Ac-

cess Points (SAPs). In other words, the 802.15.1 stan-

dard presents a wireless personal area network that

utilizes the Bluetooth wireless technology. A PAN

is deﬁned as a computer network used for communi-

cation among computer devices close to one person.

The Bluetooth WPAN operates in the unlicensed 2.4

GHz industrial, scientiﬁc and medical (ISM) band.

The Bluetooth 1.0 data rates include an asymmetric

data rate of 721 kbit/s while permitting 57.6 kbit/s

in the return direction; and a symmetric data rate of

432.6 kbit/s (IEEE Standard, 2002).

3.2 IEEE 802.15.4 /Low-Rate WPAN

IEEE 802.15.4 is a standard deﬁned for low-rate (LR)

WPANs. A LRWPAN is a simple, low cost com-

munication network that allows wireless connectiv-

ity in applications with limited power and relaxed

throughput requirements. The main objectives of a

LR-WPAN are ease of installation, reliable data trans-

fer, short-range operation, extremely low cost, and a

reasonable battery life (IEEE Standard, 2003). Like

all IEEE 802 standards, the IEEE 802.15.4 standard

encompasses only those layers up to and including

portions of the data link layer (DLL). I.e. the standard

802.15.4 deﬁnes only the PHY and the medium ac-

cess layers (MAC). In particular, it deﬁnes two PHYs

representing three license-free frequency bands that

include sixteen channels at 2.4 GHz, ten channels at

902 to 928 MHz, and one channel at 868 to 870 MHz.

The maximum data rates for each band are 250 kbps,

40 kbps and 20 kbps, respectively. A WPAN consists

of several components; the most basic is the device.

There are two different device types, which can par-

ticipate in an LR-WPAN; a full-function device (FFD)

and a reduced-function device (RFD). A FFD can talk

to RFDs or other FFDs, while an RFD can talk only

to an FFD. An RFD is intended for applications that

are extremely simple, such as light switches or pa-

ssive infrared sensors; they do not have the need to

send large amounts of data and may only be associa-

ted with a single FFD at a time. Because of that, the

RFD can be implemented using minimal hardware re-

sources (Thraning, 2005).

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

102

Table 1: Comparison of short–range low power wireless

technologies.

Bluetooth 802.15.4 ZigBee

Data rate(Mbit/s) 1 0.250 0.250

Range(m) 10-100 10-100 10-100

Power supply Medium Very low Low

Security High Medium High

Scalability 7 255 255

Cost Medium Very low Low

3.3 Zigbee

ZigBee was created to address a speciﬁc market need

for an industrial standard to support automation and

remote control applications. The ZigBee Alliance

decided to use the IEEE 802.15.4 standard as the

Physical layer and Media Access Control sub-layer,

while the ZigBee Alliance deﬁnes the upper layers.

The ZigBee network layer supports multiple network

topologies including star, cluster tree, and mesh. The

application layer consists of the application support

layer (APS), the ZigBee device object (ZDO) and the

manufacturer-deﬁned application objects. The APS is

responsible for maintaining tables for binding and for-

warding messages between bound devices. A binding

is the ability to match two devices together based on

their services and their needs. The binding include the

device discovery, which is the procedure to discover

other devices that are operating in the same area (Zig-

Bee Alliance, 2005).

The different IEEE 802.15 standards were de-

signed with different purposes. Bluetooth was de-

signed to replace cabling connections between de-

vices, while 802.15.4 and ZigBee for network sen-

sor targeted to home automation. Therefore, none of

them was speciﬁcally designed to be used in medical

sensor network.

4 PROPOSED ARCHITECTURE

FOR THE SENSOR SYSTEM

In the absent of a speciﬁc standard, in this work we

propose the use of the IEEE 802.15.4 and Zigbee

standard as the more adequate solution for medical

WPAN sensor networks. Most Bluetooth modules are

not appropriate because its power requirement is high

for battery-operated devices that cannot be charged or

changed regularly. It is also restricted by its scalabil-

ity capacity; it can only have seven active nodes in one

network. 802.15.4 and ZigBee looks like a perfect ﬁt,

except for its low data rate. This technology presents

low power requirement because the nodes are in sleep

mode most of the time.

Figure 1: Proposed network architecture for biological

signs monitoring.

By using ZigBee mesh topology supports (Thran-

ing, 2005) there is no need of a master node in the

WPAN, in this conﬁguration every node/sensor can

communicate directly and each node acts as a router.

This way a data ﬂow from one node can take multiple

routes to its destination, making it very resilient; if a

node drops out the ﬂow, it is redirected through other

nodes. A routing algorithm is used to ensure that the

data takes the fastest possible route.

A drawback of mesh networking is synchroniza-

tion. Two or more data ﬂows that need to be time syn-

chronized at the receiver may take different routes, re-

sulting in different time delays. If this delay is greater

than what is possible to buffer, data lost occur. To

overcome the above mentioned problem, in this work

we propose the use of IEEE 802.15.4 technology with

a star network topology. With respect to power con-

sumption, this technology has suitable characteristics

for medical WPAN application since it was originally

developed to have low power consumption; i.e. idea-

lly a node should work for months or years without

batteries changes. In addition, the standard was devel-

oped for the communication between sensors, which

reduces its complexity, allowing lower implementa-

tion costs than Bluetooth technologies.

5 EXPERIMENTAL SET-UP

In order to validate the proposed network an experi-

mental set-up was implemented, including the devel-

opment of some biological sensors. Figure 1 presents

the proposed network architecture. In the proposed

topology the sensors are RFD devices which commu-

nicate to a central node, conﬁgured as FFD.

RESURE - Wireless Sensor Network for Health–Care Monitoring

103

5.1 Hardware Components

There are many chips or modules manufacturers’

following this standard, including Texas Instruments,

Maxstream, FreeScale, among others. In this work,

we select an OEM module from Maxstream Com-

pany, the XBee. This module has a 100 meters range

for outdoor and 30 meters indoor, with a 1 mW of

power in antenna, consuming 45 mA in transmission

and 50 mA in reception modes. The XBee module

uses the MC13193 transceiver chip working in the

2,4GHz frequency band and a MC9S08GB60 micro-

controller, both from Freescale. In addition, the mi-

crocontroller has 8-channel analog to digital converter

with 10 bits resolution and an UART interface (Digi

International, 2006).

The WPAN central device is a mobile phone

(Nokia 6620), which collects and processes the sen-

sors signals, transmitting them to a central monitoring

system through the mobile phone network using SMS

and MMS protocols. Due to the novelty of IEEE

802.15.4 technology, the mobile phone does not sup-

port this standard, reason why we implemented an

IEEE 802.15.4-to-Bluetooth bridge for integrate the

phone to the sensor network. The Nokia 6620 was

selected because it has Bluetooth interface and runs

Symbian operating system, which allows the devel-

opment of applications in C++ or Java programming

languages.

In addition, a cardiac pulse sensor was imple-

mented. Actually, we are developing others biolog-

ical sensors such as ECG and pulse oximeter.The

cardiac pulse sensor is implemented using a piezo-

electric transducer that obtains the blood pressure

pulse waveform across an artery. The SDT1-028KD

PVDF piezoelectric transducer from MSI (Measure-

ment Specialties, 2006) was selected.

The analog circuit for transducer signal condition-

ing consists basically of a differential charge ampliﬁer

which uses three op-amps in the classic instrumenta-

tion ampliﬁers conﬁguration. The differential topol-

ogy reduces line-noise pickup, which is a problem

with high-gain circuits. Besides, an active guard has

been used to compensate unbalances in the transducer

terminals and a shield around the transducer behaves

like a Faraday cage, connected to the active guard.

The next ampliﬁers and ﬁlters stages are implemented

as single ended devices. In the Figure 2 a simpli-

ﬁed schematic of the charge ampliﬁer is shown. The

small pulses of charges are integrated and converted

to differential voltages pulses at the output of the ﬁrst

stage. The second stage is a differential–to–single–

ended ampliﬁer. The gain of the differential stage is

given by the ratio between C1 and C2, where C2 cor-

Figure 2: Charge ampliﬁer schematic for the PVDF trans-

ducer.

Figure 3: Photo of the sensor node and the smart phone used

as the central device.

responds to the equivalent capacitance of piezoﬁlm

(Maxim Inc., 2002). The circuit has a good perfor-

mance, with an S/N ratio of approximately 40 dB.

Figure 3 is a photo of the designed sensor node. It

can be seen the Xbee OEM module, the analog signal

conditioning circuit for the PVDF transducer and the

smart phone used as the central device.

5.2 Software Modules

On the Nokia 6620, both Java and C language can be

used for applications development. While both lan-

guages are capable of connecting to the Bluetooth

stack and communicate via GPRS, SMS or MMS;

Java applications are halted if the user receives an in-

coming call or loads another Java application. Writ-

ing code in C language allows our program to con-

tinue capturing data from the sensor network in the

background regardless of the task the mobile phone is

doing on the foreground. The used development en-

vironment was the Carbide V1.0 with the Nokia Se-

ries 60 SDK 2nd Edition. This platform integrates

a phone simulator for debugging purpose. Compiled

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

104

Figure 4: Measurements taken by two cardiac pulse sensors

located on two different positions of the neck.

code was packaged into a Symbian .SIS installation

ﬁle and downloaded to the phone via Bluetooth.

The Figure 4 shows some measurements taken by

two cardiac pulse sensors located on two different po-

sitions in the neck. The signals are displayed on the

screen of a test program developed using LabWin-

dows environment. It can be observed that the shapes

of the cardiac pulses are well deﬁned and the sig-

nals have good signal–to–noise ratio. If analyzed by

a physician, these waveforms could give information

about the arteries’ state.

6 NETWORK PERFORMANCE

ANALYSIS

In order to validate the proposed network architecture

performance, we implement two cardiac pulse sen-

sors. From this test bench, we obtain some network

parameters and acquired two pressure signals simul-

taneously from two different positions of the body.

Following, some results are presented:

• Security: The 802.15.4 Default Security protocol

offers AES encryption with a 128-bit key. How-

ever, the use of this feature causes a packet over-

head reducing the available bandwidth. That’s

why, the AES encryption is disable in the carried

out experiments.

• Battery lifetime: With a 2400 mAh battery and a

transmitter consumption of 45 mA with duty cycle

of 100% we have a lifetime of 2400 mAh/45mA,

approximately 53 hours. The calculation assump-

tion of 100 % duty cycle is not realistic; it could

be a lot less in this kind of networks. The ques-

tion is how often do the sensors need to trans-

mit? To achieve best possible effective data rate

it is important to use the largest packet size pos-

sible. The Maxstream module can send up to

106 bytes per packet with a maximum 92 bytes

payload. Each sample has 10 bit resolution, this

means that 46 samples ﬁt in each packet. The

sensor, with 500 Hz sampling rate, takes a sam-

ple every 2 ms (1/500 Hz). If the sensor does not

transmit until the packet is full (46 samples), the

sensor transmits every 46*2 ms = 92 ms. Trans-

mitting 106 bytes at 250 kbit/s takes 3.36 ms, to

wake up the transmitter from sleep mode takes 2

ms and to access the channel 10 ms. An ACK

frame of 11 bytes which takes 0.35 ms to be trans-

mitted. Summarizing, the radio activity time is 2

ms + 10 ms + 3.36 ms + 0.35 ms =15.71ms. Con-

sidering that the radio can sleep when it is not in

use and if no error occurs, the transmitter is in use

during 15.71ms per each 92 ms which gives a 17%

duty cycle. Beacon frames from the coordinator,

for time synchronization, is not take into account.

With the above considerations, the battery life-

time is approximately 2400mAh / (45mAh*0.17)

= 313.72 hours. This is an optimistic result; ac-

tual battery lifetime will likely be somewhat be-

low this. Also, note that sleep mode current is not

taken into account; it is approximately 50 uA and

his impact on the battery life will be minimal.

• Scalability: With a 500 sampling per second rate

and a 10 bit resolution, 5000 bits needs to be trans-

ferred each second. Each packet in the network

has 14 bytes overhead and 92 bytes payload. The

required data rate to transmit 5000 bits per second

is ((500/46)*106)*8 9.2 kbits/s; a very low data

rate considering the 250 kbits/s bandwidth avail-

able between each sensor and the PAN coordina-

tor. Theoretically, up to 25 nodes can be added

to the network, but this is not a true number be-

cause of the IEEE 802.15.4 medium access con-

trol mechanism.

From the above analysis we can conclude that the

IEEE 802.15.4 standard is appropriate for medical

WPAN sensor network.

7 CONCLUSIONS

In this work a wireless sensor network for medical ap-

plications is proposed based on IEEE 802.15.4 stan-

dard. To validate the proposal a prototype, composed

by two wireless blood pressure sensors, was build.

From the theoretical analysis and experimental

tests, we can conclude that the IEEE 802.15.4 stan-

dard is appropriate for medical WPAN sensor network

with respect to the following performance parameters:

battery lifetime, scalability and security.

Actually, we are working on the development of

other sensors, speciﬁcally an ECG and a oximeter

sensors, which will be added as nodes in the WPAN

sensors network. Also, we are working on a program

RESURE - Wireless Sensor Network for Health–Care Monitoring

105

for the mobile device that will analyze the different

signals in order to obtain useful information to be pre-

sented to the user.

REFERENCES

Digi International (2006). 802.15.4 Protocol-

Maxstream (version 1.xAx). Digi International,

www.maxstream.net.

IEEE Standard (2002). IEEE Standard for Infor-

mation technology - Telecommunications and in-

formation exchange between systems - Local and

metropolitan area networks - Speciﬁc requirements

Part 15.1: Wireless Medium Access Control (MAC)

and Physical Layer (PHY) Speciﬁcations for Wire-

less Personal Area Networks (WPANs). IEEE,

http://standards.ieee.org/.

IEEE Standard (2003). IEEE Standard for Infor-

mation technology - Telecommunications and in-

formation exchange between systems - Local and

metropolitan area networks - Speciﬁc requirements

Part 15.1: Wireless Medium Access Control (MAC)

and Physical Layer (PHY) Speciﬁcations for Wire-

less Personal Area Networks (WPANs). IEEE,

http://standards.ieee.org/.

Istepanian, R. et al. (2001). Advances in telemedicine using

mobile communications. Proceedings of the 23rd An-

nual EMBS International Conference,Istanbul,Turkey.

Karl, H. and Willig, A. (2005). Protocols and Architectures

for Wireless Sensor Networks. Wiley.

Maxim Inc. (2002). Ampliﬁer provides signal con-

ditioning for piezoﬁlm sensor. Maxim Inc.,

http://www.maxim.com/.

Measurement Specialties (2006). Sdt shielded piezo sensors

element with shielded cable. Technical report, MEA-

SUREMENT SPECIALTIES, www.meas-spec.com.

Rodr

ıguez, J. et al. (2005). Real-time classiﬁcation of ECGs

on a PDA. IEEE Transactions on information technol-

ogy in Biomedicine, 9(1):23–34.

Shnayder, V. et al. (2005). Sensor networks for medi-

cal care. Division of Engineering and Applied Sci-

ences TR-08-05, Harvard University. Available at

http://www.eecs.harvard.edu/mdw/proj/codeblue.

Thraning, B. M. (2005). The Impact of ZigBee in a BioMed-

ical Environment. Agder University College.

WHO (2004). http://www.who.int/mediacentre/news/

releases/2004/pr60/es/.

Woodward, B. and Rasid, M. (2003). Wireless

telemedicine: The next step? Proc. of the 4th Annual

IEEE Conf. on Information Technology Applications

in Biomedicine.

ZigBee Alliance (2005). ZigBee Speciﬁcation (Version 1.0).

ZigBee Alliance, www.zigbee.org.

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

106