FRAMEWORK FOR QOS PERFORMANCE ASSESSMENT

ON BIOMEDICAL WIRELESS SENSOR NETWORKS

Carlos Abreu

1,2,3

, Manuel Ricardo

and Paulo Mendes

Polytechnic Institute of Viana do Castelo, Av. Atl

antico, Viana do Castelo, Portugal

INESC Porto, Faculty of Engineering, University of Porto, Porto, Portugal

Centro Algoritmi, University of Minho, Guimar

aes, Portugal

Keywords:

Quality of service, Biomedical wireless sensor networks.

Abstract:

A Biomedical Wireless Sensor Network (BWSN) is a special Wireless Sensor Network (WSN) with a small

number of nodes designed for medical applications. These networks must ensure that medical data is delivered

reliably and efﬁciently, in order to fulﬁl a set of pre-established Quality of Service (QoS) requirements. In this

way, the research community have been proposing new solutions to improve QoS in WSN, namely in routing

protocols and power consumption efﬁciency. However, there still a need for appropriate QoS guaranties in

BWSN. In this paper, possible QoS requirements of BWSN are discussed, together with a framework to

automatically evaluate the performance of such QoS techniques. That framework was used together with

simulators and operating systems appropriate for WSN, COOJA and Contiki OS, and proved to be a valuable

tool for a proper evaluation of QoS parameters and metrics.

1 INTRODUCTION

The fast development of low-power wireless

communication technologies and devices enable the

application of WSN in several areas, e.g, ambient

monitoring, catastrophe response, industrial and

home automation, or healthcare services (Hof,

2007). Regarding the healthcare application of WSN,

BWSN can be used to develop new applications

and services in different scenarios (see ﬁgure 1),

e.g., ambient assisted living (AAL), emergency

response or patient monitoring (Ren et al., 2005).

In any of these scenarios, BWSN are presented as

a key technology to ensure high quality levels in

the healthcare services that are provided to citizens.

However, since such networks support some degree

of medical care, they must guarantee appropriated

levels of QoS, namely with respect to the intrinsic

characteristics of medical data and their applications

(Abreu et al., 2011). Then, the study of QoS in

BWSN emerges as a very important area of research,

being a very challenging task due to the interaction

of the different phases of hardware and/or software

development (Bhuyan, 2010).

Software development and debugging on

BWSN is morose and difﬁcult work to carry

out, with many interactions and testing phases

Network

Router

Sensor Node

WiFi Router

Emergency Sensor Network

Coordinator / Gateway

ECG

EMG

ECG

EMG

ECG

EMG

Patient Monitoring Network

GPRS

Ambient Assisted Living Network

Figure 1: BWSN applied to different healthcare scenarios:

ambient assisted living (AAL), emergence response and

patient monitoring.

(Huber et al., 2011). Performance evaluation of

the different QoS parameters requires comparing

different implementations and, based on data

retrieved from tests, verify which proves to be

the best implementation. To address this issue, a

platform-independent framework was designed and

implemented to evaluate the QoS and performance of

BWSN. This framework can be used for simulated

or real networks and, thus, be used to help in the

initial phase of applications development, as well

as at later design stages, to evaluate real scenario

implementations.

Abreu C., Ricardo M. and Mendes P..

FRAMEWORK FOR QOS PERFORMANCE ASSESSMENT ON BIOMEDICAL WIRELESS SENSOR NETWORKS.

DOI: 10.5220/0003706800990103

In Proceedings of the International Conference on Biomedical Electronics and Devices (BIODEVICES-2012), pages 99-103

ISBN: 978-989-8425-91-1

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

First, the QoS requirements are brieﬂy discussed,

as well as the parameters and metrics of BWSN.

Then, the framework that automatically evaluates the

performance of QoS techniques applied to BWSN is

presented.

2 QOS REQUIREMENTS OF

BWSN

As discussed in (Gama et al., 2007) and (Abreu

et al., 2011), BWSN need to ensure an appropriate

level of QoS. The QoS requirements of BWSN are

imposed to the network and depend on the target

applications. They don’t depend only on the intrinsic

characteristics of data to be transmitted, but also on

its purpose.

According to (Ruiz, 2006) and (Marchese, 2007),

from the network point of view, the most important

parameters used to measure and ensure the QoS

required for a given application, in traditional IP

networks for telemedicine applications and services,

are presented in table 1.

Table 1: QoS parameters, to traditional IP networks,

concerning the network point of view.

PTD Packet Transfer Delay

PDV Packet Delay Variation

PLR Packet Loss Ratio

PER Packet Error Ratio

BW Bandwidth

However, in the context of WSN, due to its unique

characteristics such as: extremely low resources

(memory and computational power) constrained

nodes or limited radio capabilities and battery, the

QoS parameters presented earlier are insufﬁcient. In

(Chen and Varshney, 2004) the authors identify new

QoS parameters, presented in table 2, that reﬂect the

collective effort of all network nodes to perform a

given task.

Table 2: Collective QoS parameters to WSN concerning the

network point of view.

CPTD Collective Packet Transfer Delay

CPLR Collective Packet Loss Ratio

CDR Collective Data Rate

IT Information Throughput

BWSN can be used to support several healthcare

applications and services. Therefore, they have

to provide QoS support to multi-applications.

Concerning, the application point of view, there are

different QoS requirements that must be satisﬁed.

However, despite the different speciﬁcations, it is

possible to identify some requirements that are

common to most applications or services.

Below, those requirements are identiﬁed and their

relevancy and applicability will be brieﬂy discussed.

The authors of (Xie and Wang, 2010) identiﬁed

some important application-centric QoS requirements

that can be applied also to BWSN. Below, those

requirements are identiﬁed, and their relevance and

applicability will be outlined.

Ability to Provide Valid Data, in BWSN the data

integrity is a major requirement. The decisions

made by healthcare providers depend on the data

provided by the network. In this way, this

is a requirement that must be satisﬁed on all

BWSN applications and services. This parameter

is inﬂuenced by topology, coverage and trafﬁc

capacity, connectivity, and energy consumption;

Delay Management, the time spent to transmit data

and detect events is a very important issue

in BWSN. Different applications have distinct

delay requirements, however, all of them have

some limit that must be respected. This is a

requirement that must be applied with different

weights to different applications. This parameter

is affected by collective packet transfer delay,

information throughput, network topology and

decision algorithm;

Network Lifetime, is the period of effective service.

This is a major issue especially in implanted

BWSN nodes. It is affected by network topology,

connectivity and energy consumption;

Network Survivability, is the ability to

automatically restore. It is more evident in

BWSN with mobile nodes or in scenarios with

constant changes, like in a hospital where patients

are being monitored. It depends on deploying

density of nodes and routing protocols;

Decision Accuracy, the quality of the decision made

by healthcare providers depend on the quality

of the data collected, transmission delay and

decision algorithm.

Developing QoS mechanisms to guarantee

appropriate levels of QoS is important but not

sufﬁcient (Asokan, 2010). To prevent QoS

degradation, due to network limited capacity, it

is necessary to develop tools to supervise and

manage the network. This work contributes, with

the development of a framework, to help in the

development and evaluation of strategies to supervise

and manage the QoS on BWSN.

BIODEVICES 2012 - International Conference on Biomedical Electronics and Devices

100

3 QOS ASSESSMENT

FRAMEWORK FOR BWSN

As discussed previously, BWSN have to fulﬁl

strong QoS requirements each of which have to

be systematically tested to guarantee high levels of

conﬁdence and reliability. This is true, not only in

research, but also during the development of speciﬁc

applications. BWSN have to pass very demanding

tests to be accepted by its users and the healthcare

providers. They have to be tested in both simulated

and real scenarios.

The development of applications and protocols for

BWSN is a challenging, time consuming and error

prone task (Mozumdar et al., 2010). To avoid this

situation various software and hardware tools have

been developed in the last few years (Huber et al.,

2011).

The use of simulators for BWSN software

development proved to be a great help, and with

widespread use within the academic and industrial

community (Liang, 2009). However, the raw data

generated by simulators needs to be analysed in order

to extract results, as well as for reporting purposes.

This analysis may be laborious and time consuming.

In fact, the data analysis and report generation is one

of the bottlenecks of software development processes.

To mitigate this problem, a framework was developed,

based on open source tools, for automatic data

analysis and QoS performance assessment.

The developed QoS assessment framework is

platform-independent and might be used both for

simulated and real networks. In this way,

it’s possible to evaluate the network along the

development process as well as when the network

is ﬁnally deployed in a real application. This

is an important feature because it enables easier

performance comparison between simulated and real

BWSN.

Figure 2: WSN hardware platform architecture.

In order to implement the real BWSN, a hardware

platform (WSN mote) was developed and included

on the QoS assessment framework. The mote was

developed to be modular, thus, it can be used in

several scenarios. Its functionalities can be extended

via daughter boards, providing different sensing

capabilities. Depending on the daughter sensor

board, the mote can be used in different healthcare

applications. Figures 2 and 3 show its architecture

and implementation, respectively.

Figure 3: WSN hardware platform implementation.

The core element of the mote is the low-power RF

module deRFmega128. This module is a 2.4 GHz

radio transceiver and a 8Bit microcontroller single

chip solution (SoC) (Dresden, 2011). To interact with

the outside, the mote contains three interfaces: 1) a

JTAG interface to program and debugging purposes;

2) a mini USB port; 3) a 60-Pin header to connect to

daughter sensor boards. The mote power supply can

be from different sources, e.g: battery, USB port or

energy harvesting modules.

3.1 Framework Architecture

The QoS assessment framework follows a two layer

architecture: Data Analysis (DA) and visualisation,

as can be seen in ﬁgure 4. The DA layer is

responsible for results extraction from raw data,

generated by simulated or real networks, and these

will be represented in such a way that will be easily

interpreted. The visualisation layer is used for report

generation purposes.

The software used to extract results from raw data

and the report template depends on the study that

is being carried out. However, once developed, the

software can be reused in future assessments. The

results representation and report generation software

is independent of the application. In this way,

the network QoS evaluation is improved between

different tests of the same study.

FRAMEWORK FOR QOS PERFORMANCE ASSESSMENT ON BIOMEDICAL WIRELESS SENSOR NETWORKS

101

Figure 4: QoS assessment framework architecture.

3.2 Framework Evaluation

To evaluate the implementation of the QoS

assessment framework, a BWSN case study

was deﬁned for simulation. The BWSN was

simulated using COOJA that, with Contiki Operating

System (Contiki OS), can be considered a complete

development platform for WSN. COOJA allows

simultaneous simulations at different levels, namely:

application, network, operating system and machine

code instruction level (Osterlind et al., 2006)

(Dunkels et al., 2004). One of the major advantages

in using Contiki OS and COOJA is the code reuse

and hardware platform independence. It is possible

to develop the application software and then use it

in different hardware platforms, as well as inside the

COOJA simulator.

There are several applications to BWSN.

However, in this case study, the focus is on the

monitoring of vital and biological signals, such as,

electrocardiogram (ECG), electroencephalogram

(EEG), blood pressure (BP) and temperature (T).

According to (Ruiz, 2006) and (Varshney, 2009), the

default value for the PTD in real-time telemedicine

applications and services is 400ms. In this context,

the case study aims to evaluate the end-to-end PTD

regarding the number of hops in the network and

which must be less than 400ms.

Each case study consists in: a scenario to deploy

the network; a real or simulated WSN; a test script

to extract data from the network; a script to extract

results from data generated by network; and a report

template to easily evaluate the results (see ﬁgure 5).

After the establishment of the scenario in which

the BWSN will be used, the application is developed

on the Contiki OS and, then, compiled to the hardware

in use. The simulations are performed on COOJA. At

the end of all the simulations, the generated raw data

is analysed and the results extracted. Finally, a report

is generated to present the results and compare them

with the previous. Figure 5 illustrates the framework

operation.

Figure 5: QoS assessment framework operation.

To study the end-to-end delay of a package

regarding the number of hops in the network, 9

simulations were mounted in COOJA (1 to 10

hops). Figure 6 represents the simulation scenario to

measure the end-to-end delay in the case of 5 hops.

The data from each simulation were extracted with

a script written in java script using the Contiki Test

Editor plugin of COOJA.

The extraction of results from data generated by

simulation was done through a script that uses the

Linux Shell Scripting capabilities. To represent the

results, the Gnuplot graphing utility was used, and the

report was automatically generated using the pdﬂatex

library.

To perform the simulations experience, COOJA

was used in batch mode, which made it possible to

automatically perform all the simulations and carry

out the results, see ﬁgure 6.

100

2 3 4 5 6 7 8 9 10

Average delay (ms)

# Hops

Figure 6: Mean end-to-end delay of a package regarding the

number of hops in the network.

As shown in ﬁgure 6, in this test case scenario,

the end-to-end PTD is less than 400ms. However,

BIODEVICES 2012 - International Conference on Biomedical Electronics and Devices

102

if necessary, it is possible to adjust the software,

and automatically perform new evaluations without

human interaction streamlining the process of

software development.

4 CONCLUSIONS

This work presents a framework to support the

software development for WSN that was developed

envisioning QoS deployment and evaluation of

BWSN. It has proved to be a very useful and

complementary tool for WSN simulators and

operating systems. As a following step, this

framework will be used as a tool to evaluate and test

protocols, targeting QoS requirements of BWSN,

exploring the best solutions to achieve the QoS

parameters presented in tables 1 and 2 in simulated

and real BWSN. As it allows a straightforward

network deployment after simulations, this

framework has the potential to reduce the debugging

time after a real network deployment.

ACKNOWLEDGEMENTS

This work has been ﬁnancially supported by the

PhD grant of Portuguese Foundation for Science and

Technology, FCT, SFRH/BD/61278/2009.

REFERENCES

Abreu, C., Ricardo, M., and Mendes, P. (2011). Monitoring

qos over wireless sensor networks for medical

applications. In Biodevices 2011 - International

Conference on Biomedical Electronics and Devices,

Rome Italy.

Asokan, R. (2010). A review of quality of service

(qos) routing protocols for mobile ad hoc networks.

In Wireless Communication and Sensor Computing,

2010. ICWCSC 2010. International Conference on,

pages 1 –6.

Bhuyan, B. (2010). Quality of service (QoS) provisions

in wireless sensor networks and related challenges.

Wireless Sensor Network, 02(11):861–868.

Chen, D. and Varshney, P. K. (2004). Qos support in

wireless sensor networks a survey. In Proc. of the

2004 International Conference on Wireless Networks

(ICWN 2004), Las Vegas, Nevada, USA.

Dresden, E. (2011). Dresden elektronik - 2.4GHz radio

module. http://www.dresden-elektronik.de.

Dunkels, A., Gronvall, B., and Voigt, T. (2004). Contiki -

a lightweight and ﬂexible operating system for tiny

networked sensors. In Local Computer Networks,

2004. 29th Annual IEEE International Conference on,

pages 455 – 462.

Gama, O., Figueiredo, C., Carvalho, P., and Mendes, P. M.

(2007). Towards a reconﬁgurable wireless sensor

network for biomedical applications. In Proceedings

of the 2007 International Conference on Sensor

Technologies and Applications, SENSORCOMM.

IEEE Computer Society.

Hof, H.-J. (2007). Applications of sensor networks. In

Wagner, D. and Wattenhofer, R., editors, Algorithms

for Sensor and Ad Hoc Networks, volume 4621 of

Lecture Notes in Computer Science, pages 1–20.

Springer.

Huber, R., Sommer, P., and Wattenhofer, R. (2011).

Demo abstract: Debugging wireless sensor network

simulations with yeti and cooja. In Information

Processing in Sensor Networks (IPSN), 2011 10th

International Conference on, pages 141 –142.

Liang, X. (2009). QoS Provisioning for Wireless Sensor

Networks: Algorithms, Protocols and Modeling. PhD

thesis, Interventional Center, University of Oslo.

Marchese, M. (2007). QoS Over Heterogeneous Networks.

John Wiley & Sons Ltd.

Mozumdar, M., Lavagno, L., Vanzago, L., and

Sangiovanni-Vincentelli, A. (2010). Hilac: A

framework for hardware in the loop simulation and

multi-platform automatic code generation of wsn

applications. In Industrial Embedded Systems (SIES),

2010 International Symposium on, pages 88 –97.

Osterlind, F., Dunkels, A., Eriksson, J., Finne, N.,

and Voigt, T. (2006). Cross-level sensor network

simulation with cooja. In Local Computer Networks,

Proceedings 2006 31st IEEE Conference on, pages

641 –648.

Ren, H., Meng, M.-H., and Chen, X. (2005). Physiological

information acquisition through wireless biomedical

sensor networks. In Information Acquisition, 2005

IEEE International Conference on, page 6 pp.

Ruiz, I. M. (2006). Contribuciones a Modelos de Tr

aﬁco

y Control de QoS en los Nuevos Servicios Sanitarios

Basados en Telemedicina. PhD thesis, Instituto

de Investigaci

on en Ingenier

ıa de Arag

on, Dpto.

de Ingenier

ıa Electr

onica y Comunicaciones, Centro

Polit

ecnico Superior, Universidad de Zaragoza.

Varshney, U. (2009). Pervasive Healthcare Computing:

EMR/EHR, Wireless and Health Monitoring.

Springer, 1 edition.

Xie, D. and Wang, F. (2010). A self-adjustment

qos architecture for wireless sensor networks.

The Journal of China Universities of Posts and

Telecommunications, 17(Supplement 2):79 – 83.

FRAMEWORK FOR QOS PERFORMANCE ASSESSMENT ON BIOMEDICAL WIRELESS SENSOR NETWORKS

103