REQUIREMENTS AND SYSTEM ARCHITECTURE

FOR A HEALTHCARE WIRELESS BODY AREA NETWORK

Finn Overgaard Hansen and Thomas Skjødeberg Toftegaard

Aarhus School of Engineering (ASE), Aarhus University, Aarhus, Denmark

Keywords: Pervasive Healthcare, Wireless Body Area Network (WBAN), Sensor Networks, Healthcare System

Architecture, Body Gateway, Body Area Network (BAN) Requirements.

Abstract: Wireless body area networks enable new opportunities for personal healthcare monitoring and personal

healthcare applications. This paper presents a comprehensive set of requirements and challenges for build-

ing a wireless body area network to support diverse user groups and a corresponding set of healthcare appli-

cations. Based on the identified requirements, the paper presents an architecture for a wireless body area

network and describes how this architecture is connected to an existing it-infrastructure supporting health-

care at home. Finally the paper presents our on-going research with development of an ASE-BAN test bed.

The major goal for this test bed is to be a platform for research and experiments with development of an ul-

tra-low power body area network including sensor, communication nodes, communication protocols and a

body gateway component.

1 INTRODUCTION

AND BACKGROUND

There is a rising need for personal home healthcare

due to a growing population of elderly people (Patel

et. al., 2010) and (Wagner et. al., 2009). To support

the health problems of the elderly population wire-

less communication technologies have enabled new

types of applications for monitoring and controlling

people’s physical parameters.

The first generation of e-healthcare solutions

were more or less replacement of a wire with a wire-

less communication channel i.e. another set of pro-

tocols on top of the new physical communication

media. In the second generation the devices commu-

nicated wireless with a local system host, which re-

layed alarms and possible also data to remote sites.

In the third generation the measuring devices are

wireless connected to a mobile body area network

along with other sensors and actuator.

The development of successful body area net-

work (BAN) infrastructure, sensors and the support-

ing applications involves specifying a set of re-

quirement based on real-life problems of the elderly.

The current work in this field is mainly dominated

by a technology driven perspective but as the suc-

cess of these body near technologies depends very

much on user acceptance of the technologies we

suggest, to use a much more user focused and user

driven innovation process in the future research in-

cluding practical testing with real users using these

new inventions.

In this paper we will present a comprehensive set

of requirements for a BAN-system and the individ-

ual BAN components.

The proposed requirement are based on our cur-

rent work with architecture and network infrastruc-

ture for home healthcare in scope of the OpenCare

project (Wagner et. al., 2009). The focus has initially

been to create an it-infrastructure for connecting

single wireless healthcare sensors typically based on

Bluetooth. In this paper we extend the OpenCare

project with the description of development of a test

bed for a body area network called ASE-BAN.

The paper will first present a definition of ASE-

BAN requirements and based on these develop a

supporting ASE-BAN architecture.

This paper takes the perspective of the user, in

defining the functional requirement as well as the

non-functional system requirements.

More technical requirements are currently being

defined by the IEEE 802.15 WPAN Task Group 6.

(IEEE P802.15, 2008), which defines the require-

ments for a Wireless Personal Area Networks

(WPAN). A short overview of these requirements,

193

Overgaard Hansen F. and Skjødeberg Toftegaard T..

REQUIREMENTS AND SYSTEM ARCHITECTURE FOR A HEALTHCARE WIRELESS BODY AREA NETWORK.

DOI: 10.5220/0003138601930199

In Proceedings of the International Conference on Health Informatics (HEALTHINF-2011), pages 193-199

ISBN: 978-989-8425-34-8

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

current challenges and wireless technologies for

BANs are presented by Patel (Patel et. al., 2010).

A supplementary proposal for technical require-

ments is presented by Drude (Drude, 2007).

2 APPLICATION SCENARIOS

There are numerous user scenarios for healthcare

applications (Drude, 2007). In this Section we will

present the four user scenarios we found most inter-

esting in relation to development of a BAN.

2.1 Health Monitoring

Health monitoring is about using technology to

monitor a person remotely and is one of the most

common usages of healthcare applications. An ex-

ample could be for heart diseases, where a sensor

will monitor the person’s ECG signal for detecting

heart arrhythmias, heart attach or for monitoring of a

person’s heart rate variability (HRV). Another typi-

cal scenario includes blood pressure measurements

with remote registration of measurements.

2.2 Closed Loop Applications

In a closed loop application an automatic regulation

loop is closed between a sensor and an actuator e.g.

a sensor measuring glucose level and automatically

controlling an insulin pump actuator. These types of

applications are yet not as common, as they require

very reliable systems and also a very strict approval

procedure from healthcare authorities.

2.3 Rehabilitation from Accidents

In a rehabilitation process after an accident, a per-

sonal healthcare system can give instructions for

exercises and monitor the performance and the pro-

gress. Users for this type of application could be of

all ages. A body area network used for computer

assisted rehabilitation is described by Joanov (Joa-

nov et. al., 2005). Computer assisted home training

have the opportunities to getting people quicker back

to a normal life and also to save the users time and

money.

2.4 Empowerment of Elderly People

With the growing number of elderly and active peo-

ple, these types of application could in the future

help this group be active longer, stay in their homes

and thereby prevent hospitalization. It could be in-

teractive computer games or dedicated training pro-

grams which is controlled by feedback from the

BAN sensor nodes.

3 BAN REQUIREMENTS

This section presents requirements for a wireless

body area network consisting of wireless network

nodes (called motes in Figure 1) and a body gateway

node connected to some kind of host server. The

motivation for this architecture will be given in Sec-

tion 4.

Body Area Network

Body

Gateway

Host

Server

Sensor A

Actuator B

Mote

Figure 1: Wireless body area betwork topology and

components: sensors/actuators, body gateway and host

server.

Requirements in this section are mainly require-

ments which will have an influence on the system

architecture for the BAN and thereby not an attempt

to define a complete set of application oriented func-

tional requirements, which normally are defined

using the use case technique. First, the user related

requirements are described, followed by a set of

more general system requirements. Most of these

requirements have an impact on both the hardware

architecture and partly also on the software architec-

ture. A subset of these requirements is also listed by

Shnayder (Shnayder et. al., 2005).

3.1 User Related Requirements

3.1.1 Diverse User Group

We name the user of the BAN a citizen as a more

inclusive term than patient. Users of the BAN can

for example be elderly persons living at home or in a

nursing home, it can be physical disabled persons at

all ages, it can be persons suffering from dementia, it

can be persons with chronicle diseases at all ages

and it can be athletes. Some of these users have sev-

eral of these characteristics e.g. an elderly physical

disabled person with a chronicle disease.

HEALTHINF 2011 - International Conference on Health Informatics

194

We have in this way a very diverse user group

spanning from young to very old and in some cases

people suffering from dementia. These different

types of users have very different needs and differ-

ent skill levels for handling new technology, where

the user group with dementia and disabled people

raise the largest challenge for the healthcare devel-

opers. This leads to the first challenge:

Challenge 1: Dealing with very diverse types of

users, with different application needs and differ-

ent skill levels.

Requirements: Adjustable technology, user

friendly, easy installation and configuration of

software and hardware, easy to add new

functionality and sensors.

Development of a BAN system for this diverse

user group will benefit from using a user driven in-

novation and development process.

3.1.2 Citizen Communication

The BAN should support different ways of commu-

nicating with the user. It could be by messages, LED

lamps and sounds; it could be by speech syntheses or

speech recognition, by activating normal buttons or

soft buttons on a touch screen. Another possibility is

communication with hearing-aids or headphones.

Some of these devices can be used to give reminders

to the citizen e.g. a reminder to take medicine or to

exercise or to measure blood pressure.

Challenge 2: User interface design for a diverse

user group.

Requirements: User friendly and easy to use in-

terfaces.

This could be obtained by conducting usability

studies with different user groups and different types

of interfaces supported by incorporating industrial

designers in the design team and process.

3.1.3 Calling for Help

The BAN should support a “call for help” device so

a citizen can call help at any time. This functionality

could be supplemented with a voice-channel so the

caretakers can communicate with the citizen.

Challenge 3: To offer safety and security to citi-

zens.

Requirements: Physical design of a reliable call-

device and a reliable system for transferring this

event, as this could be an emergency call.

3.1.4 GPS Outdoor Positioning

The BAN should allow the connection of a GPS-

device for locating people in case of an accident. It

could for example be demented people who left the

nursing home without supervision or a citizen get-

ting a heart attack outside the home. As a GPS-

receiver is a power demanding device, the receiver

should be controlled by the BAN and the connected

system so it only works on demand and therefore

only use power in a short time frame.

Challenge 4: To locate a citizen in case of an ac-

cident.

Requirement: Outdoor navigation using GPS.

Ideally, indoor positioning is also relevant, how-

ever this is currently much more challenging and

not part of our current research scope.

3.1.5 Fall Detection

The BAN should support a fall detection device

node with the purpose of sending an automatic call

for help. It could be in situations where the citizen is

unconscious after a fall or it could be a person with

dementia, who couldn’t operate a call button or a

call device.

Challenge 5: Reliable detection of a fall.

Requirement: Physical design of a tiny and reli-

able fall detector node integrated on the person e.g.

in the cloth or in a belt.

3.1.6 Mobility

The citizen should be allowed to move freely around

that means for example a heart ECG monitoring

should take place indoors in a private home or at

work as well as outdoors and in public places.

Challenge 6: To be continuous and everywhere

connected.

Requirement: Seamless connectivity over hetero-

geneous networks with automatic roaming sup-

porting indoor as well as outdoor communication

over Wireless-LAN (WLAN) and WWAN.

3.1.7 Physical Constraints

for BAN Components

All the BAN components are connected with wire-

less technology and should be as integrated as possi-

ble in the person’s daily life. This raises specific

requirements for the physical design i.e. it should

have a very small form factor and be lightweight and

have a smart design.

REQUIREMENTS AND SYSTEM ARCHITECTURE FOR A HEALTHCARE WIRELESS BODY AREA NETWORK

195

Some of these devices requires skin contact and

could be integrated in a plaster; some could be inte-

grated in the cloth as an intelligent textile and some

should be visible e.g. a device with user interaction

for example integrated in the body gateway.

Challenge 7: Obtaining user acceptance of health-

care technology devices and wearing.

Requirements: Low form factor, low weight and

easy installation, wearing and a nice-looking de-

sign.

3.1.8 Power Consumption

With the diverse user group in mind it is difficult for

these users to handle battery exchange and charging

of a number of sensor nodes. For general conven-

ience the devices should be developed as ultra-low

powered devices with either long battery life or util-

izing some kind of energy harvesting energy neutral

technique. This leads to the architectural design with

only a single demanding unit – the body gateway.

Challenge 8: Low powered devices with low

power communication cost.

Requirements: There is a demand for ultra-low

powered devices (nodes) and communication pro-

tocols.

The body gateway requires more power and

could be charged e.g. by an inductive way or by a

normal power charger with the inconvenience for the

user and problems with being offline.

3.1.9 Economics for a BAN

The technology can help reducing the workload with

care giving, but of course with a price – the cost of

the new healthcare technology. With the high vol-

ume of users there are strict requirements to the so-

lutions to be as cheap as possible both in buying,

installation and operation.

Challenge 9: Obtaining low total system cost and

operation cost.

Requirements: Low system costs, low communi-

cation costs, specific for the mobile communica-

tion part, which currently can be quite expensive.

3.2 General System Requirements

3.2.1 Security and Safety Issues

It is very important that both the BAN and the rest of

the infrastructure are both safe and secure. Personal

related information is normally regulated by national

laws and should be transferred in a safe and secure

manner. Another problem could be external hackers

which could threaten for example a close-looped

application connected to a medicine injection pump.

Personal related information is very sensitive and

confidential and a BAN sets strict requirements to

the handling of this information.

Challenges 10: Obtaining a safe and secure sys-

tem.

Requirements: Use of standard encryption tech-

niques and authentication protocols.

3.2.2 Healthcare Application Flexibility

The BAN should support the possibility to place the

application or business logic code on different places

in the architecture. It could be on a sensor node, on

the gateway or on one of the connected servers. Im-

plementing an application on a sensor node doing

pre-processing of the signal can reduce the commu-

nication bandwidth and thereby save power, but at

the cost of a more expensive sensor node.

Challenge 11: Obtaining a flexible software and

hardware architecture with different processing

capabilities.

Requirements: An adjustable SW-framework or

structure for application code and a flexible com-

ponent oriented hardware architecture.

An automatic configuration of the application

and sensor node software is a clear goal.

3.2.3. Monitoring Data Types

Data types can be real-time life-critical application

data: ECG data as well as sporadic event data for

example alarms and emergency call for help.

Challenge 12: Very diverse requirements for sig-

nal monitoring.

Requirements: Support for continuous real-time

monitoring as well as for sporadic events. See

(Patel et.al., 2010) and (IEEE P802.15, 2008) for a

list of technical requirements for different applica-

tions with bitrates from less than 1 kBit/s for drug

dosage and up to 10 MBit/s for video imaging.

3.2.4. Citizen Identification

The BAN should support an identification mean so

the citizen can be unambiguous identified by sup-

porting systems and the identification can be send

with the collected data to remote servers.

Challenge 13: To obtain an unambiguous and se-

cure identification.

Requirement: A secure identification of the citi-

zen is required for the BAN system

HEALTHINF 2011 - International Conference on Health Informatics

196

3.2.5 Node and Person Matching

The BAN should support a mean for unambiguous

identification of sensor- and actuator nodes on a

given person and connect these devices with the

citizen’s identification code. In this way the sensor

data can be linked to a given person. The problem

occurs when a sensor node connects to nodes on

other persons BAN in near vicinity of the person.

Challenge 14: Matching nodes with the person

wearing the wireless node.

Requirement: For a secure and easy identification

method.

This could for example be obtained by using

Body-coupled communication (BCC) where the

BCC is used to discover an identify sensor nodes on

the same body as presented in (Falck et. al, 2007).

3.2.6 Open Standards and Open Source

The BAN should be based on open international

standards for supporting as many BAN devices as

possible from different vendors and with different

types of functionality. The Continua Health Alli-

ance, a non profit coalition of more than 200 mem-

ber companies, has defined interoperability goals for

wireless systems (Continua Alliance, 2010) and the

(IEEE P802.15, 2008) group is working on an IEEE

standard for wireless personal area networks. The

Continua Alliance material and software are mainly

openly available for members of the alliance.

Challenge 15: Development of open standards for

the BAN.

Requirements: Open standards free to everyone

and optionally also open source software solutions

for BAN components.

3.2.7 Network Topology & Communication

The BAN should work with any kind of network

topology from a star network with bidirectional

communication between gateway and sensors or

actuators, to a meshed network that allows commu-

nication between all nodes. It is critical to have a

network infrastructure and related communication

protocols that minimize the power consumption of

this part of the BAN too.

Challenge 16: Design a network with ultra-low

power, secure and reliable communication.

Requirements: Support for star and mesh topol-

ogy.

4 SYSTEM ARCHITECTURE

AND ON-GOING RESEARCH

The overall design guideline for the ASE-BAN is to

have a powerful body gateway node acting as the

link to external systems i.e. the host server in Figure

1. and the Central Server or Home Base Station in

Figure 2. This body gateway should be the only

power demanding component with a longer commu-

nication range supporting both wireless LAN and

WWAN communication and with a seamless hand-

over between the two.

The other BAN-nodes should be ultra-low power

sensor or actuator nodes with a limited communica-

tion range less than one meter, where the communi-

cation power level can be adjusted to the minimum

required for getting a reliable on-body communica-

tion.

Figure 2 shows a proposal for a domain model

for a complete healthcare system showing the BAN

system which is mounted on the indicated citizen.

When the citizen is at home the communication will

be over WLAN from the body gateway and it can

typically send both alarms and monitored data from

the BAN to the home base station component, e.g. a

touch screen based computer.

If the citizen leaves his or her home the BAN

will automatically stop sending real-time monitoring

data and store them locally on the BAN gateway

component and only communicate alarms and keep

alive signals over the WWAN (e.g. GSM or UMTS).

Figure 2: System domain model including BAN.

This solution is previously proposed in

(Saadaoui et.al, 2007), as it saves communication

cost i.e. both power and money. The principle of

having a Central Server and a Home Base Station is

REQUIREMENTS AND SYSTEM ARCHITECTURE FOR A HEALTHCARE WIRELESS BODY AREA NETWORK

197

implemented in the OpenCare project described in

(Wagner et.al., 2009), where the BAN is described

as a Mobile Tier component for communicating a

single physical value from a citizen and not as being

a part of a body area network.

The idea of having a powerful gateway for the

body area network is also described in the work by

Janov (Janov et. al., 2005) and Otto (Otto et.al.,

2006), where they describe a three tier system con-

sisting of tier 1. WBAN nodes, tier 2. a Personal

Server and tier 3. Central Systems. On their WBAN

each node communicates in a star network with the

personal server i.e. the gateway. In our work we

have both a star and a mesh network as possible so-

lutions as a mesh configuration enables ultra-low

power communication. Another important differ-

ence, in relation to the work described in (Saadaoui

et. al., 2007), is the introduction of the home base

station component, which gives another level of ser-

vice to the citizens living in a private home; for eld-

erly people normally one or two persons. The home

base station collects monitoring data from the BANs

for the people living in the house and it also supports

shared and non-personal related healthcare devices

in the home, which assist the residents with staying

healthy. This could be a medicine dispenser automa-

ton, a blood pressure meter or a weight, which can

have one or more users. Using a home base station

enables development of healthcare applications

which takes decisions based on inputs from several

different input sources i.e. BAN sensors or from the

shared devices.

Another advantage with the WWAN enabled

body gateway is the extra security obtained by hav-

ing a backup channel for alarms in case of malfunc-

tions in the normal data flow from BAN, to home

base station to central server.

Figure 3: The ASE-BAN ECG Sensor module. The size is

13 mm x 18 mm x 30 mm. The weight is 6 g.

Figure 3 shows an example of a sensor node

used in the ASE-BAN test bed (Madsen et. al.,

2010) consisting of a sensor print with a digital sig-

nal processer connected in a sandwich structure with

a wireless radio print. This gives the flexibility to

experiment with different radio technologies and

components. In our further work we plan to develop

these modules in even smaller scales. Other exam-

ples of ASE-BAN sensor nodes under development

are a fluid balance sensor node for detecting dehy-

dration of elderly people and a fall detector node.

BAN nodes can be body worn sensors or actuators

as well as implantable medical devices (IMD) e.g. a

pacemaker.

5 FUTURE WORK

AND CONCLUSIONS

Our future research will experiment with different

Wireless radio technologies, mesh networks with

power efficient communication protocols and build-

ing a flexible hardware and software architecture for

both the body gateway and the sensor nodes. Cur-

rently we are developing a generic and flexible sen-

sor node platform, where different types of sensors

can be mounted together with different processor

types and radios. We are currently investigating the

body gateway platform, where we are looking on the

possibility of using a standard smartphone or alter-

natively develop an embedded body gateway.

This paper has defined a set of research chal-

lenges and corresponding set of relevant require-

ments for a BAN system to be used in healthcare

application. Based on these requirements a system

architecture proposal for an ASE-BAN has been

presented which is integrated with an infrastructure

in the home and with central servers.

REFERENCES

ContinuaAlliance, 2010: http://www.continuaalliance.org

(on line, 2010-07-05)

Drude, S. 2007: Requirements and Application Scenarios

for Body Area Networks. 16th IST Mobile and Wire-

less Communications Summit, Budapest, 2007, pp.1-5.

Falck, T. et. al., 2007: Plug´n play simplicity for wireless

medical body sensors. Mobile Network Applications,

vol. 12. no.2-3, 2007, pp. 143-153,.

IEEE P802.15, 2008: Wireless Personal Area Networks,

TG6 Technical Requirements Document, IEEE P802.

15-08-0644-09-0006. https://mentor.ieee.org/802.15/

Joanov, E., Milenkovic, A., Otto, C., C de Groen, P. 2005:

A wireless body area network of intelligent motion

sensors for computer assisted rehabilitation. Journal of

NeuroEngineering and Rehabilitation, 2005, Vol. 2,

pp. 6-16.

HEALTHINF 2011 - International Conference on Health Informatics

198

Madsen J. K., Karstoft H., Hansen F. O., Toftegaard T. S.,

2010: ASE-BAN – a Wireless Body Area Network

Testbed. EMERGING 2010, Florence, Italy, 2010, pp.

1-4.

Otto, C., Milenkovic, A., Sanders, C., Jovanov, E. 2006:

System Architecture of a wireless body area sensor

network for ubiquitous health monitoring. Journal of

Mobile Multimedia. Vol. 1. No. 4, 2006, pp. 307-326.

Patel, P., Wang, J. 2010: Applications, Challenges, and

Prospective in Emerging Body Area Networking

Technologies. IEEE Wireless Communications,

Feruar 2010, pp. 80-88.

Shnayder, V., Chen, B., Lorincz, K. Fulford-Jones, T. and

Welsh, M., 2005: Sensor Networks for Medical Care.

Proceedings of the 3rd international conference on

Embedded networked sensor system. San Diego, USA,

2005, pp. 314-327.

Saadaoui, S., Wolf L. 2007: Architecture Concept of a

Wireless Body Area Sensor Network for Health Moni-

toring of Elderly People. 4th IEEE Consumer Com-

munications and Networking Conference, 2007.

CCNC 2007, pp. 722-726.

Wagner, S., Nielsen, C. 2009: OpenCare Project: An

Open, Flexible and Easily Extendible Infrastructure

for Pervasive Healthcare Assisted Living Solutions.

Proceedings of the 4th International ICST Conference

on Pervasive Computing Technologies for Healthcare,

London, 2010, pp.1-10.

REQUIREMENTS AND SYSTEM ARCHITECTURE FOR A HEALTHCARE WIRELESS BODY AREA NETWORK

199