Wireless Sensor Networks Testbeds and State-of-the-Art Multimedia Sensor Nodes

Abstract

In recent years, Wireless Sensor Networks (WSNs) attracted researchers’ attention. This has led to advancements made in both hardware design and software components. The research at the hardware level has resulted in a variant of WSN called Wireless Multimedia Sensor Networks (WMSNs). WSNs have application in battlefields, industrial process monitoring, and environmental monitoring, to name but a few. Furthermore,WMSNs have the potential for real-time object detection and recognition. It can be used to locate misplaced items, assisted living, helping people with cognitive impairments, species detection, and providing real-time images for many other surveillance-based applications. Thus, such networks need to operate in a diverse environment. Researchers invariably test their designed algorithms and protocols using simulation tools. Simulators do not accurately model the environment in which such networks are deployed. Therefore, there exists many WSNs and WMSNs testbeds. These testbeds enable researchers and programmers to validate the performance of their algorithms and protocols on a physical network. In this paper, we provide a detailed description of various state-of-the-artWSN testbeds.We have evaluated these state-of-the-art testbeds on a set of metrics. Furthermore, we survey the state-of-the-art in wireless multimedia sensor nodes hardware along with their eminent features.

Full text

Appl. Math. Inf. Sci. 6, No. 1, 29-33 (2012) 29
c
2012 NSP
Natural Sciences Publishing Cor.
Wireless Sensor Networks Testbeds and State-of-the-Art
Multimedia Sensor Nodes
Muhammad Omer Farooq1and Thomas Kunz2
1Institute of Telematics, University of Luebeck, Luebeck, Germany
2Department of Systems and Computer Engineering, Carleton University, Ottawa, Canada
Received: Jul 8, 2011; Revised Oct. 4, 2011; Accepted Oct. 6, 2011
Published online: 1 January 2012
Abstract: In recent years, Wireless Sensor Networks (WSNs) attracted researchers’ attention. This has led to advancements made in
both hardware design and software components. The research at the hardware level has resulted in a variant of WSN called Wireless
Multimedia Sensor Networks (WMSNs). WSNs have application in battlefields, industrial process monitoring, and environmental
monitoring, to name but a few. Furthermore, WMSNs have the potential for real-time object detection and recognition. It can be used to
locate misplaced items, assisted living, helping people with cognitive impairments, species detection, and providing real-time images
for many other surveillance-based applications. Thus, such networks need to operate in a diverse environment. Researchers invariably
test their designed algorithms and protocols using simulation tools. Simulators do not accurately model the environment in which such
networks are deployed. Therefore, there exists many WSNs and WMSNs testbeds. These testbeds enable researchers and programmers
to validate the performance of their algorithms and protocols on a physical network. In this paper, we provide a detailed description of
various state-of-the-art WSN testbeds. We have evaluated these state-of-the-art testbeds on a set of metrics. Furthermore, we survey the
state-of-the-art in wireless multimedia sensor nodes hardware along with their eminent features.
Keywords: Wireless Sensor Networks (WSNs), Wireless Multimedia Sensor Networks (WMSNs), Wireless Sensor Networks
Testbeds, Wireless Multimedia Sensor Nodes
1. Introduction
Recent years have seen tremendous research activity in the
WSN domain. Research in WSNs is resulting in a multi-
tude of new protocols and systems. Invariably, scientists
use simulation tools and mathematical modelling to eval-
uate their proposed protocols. Since WSNs have the po-
tential to operate in diverse environments,simulation tools
and mathematically modeling do not capture physical chan-
nel characteristics accurately. This results in in-accurate
performance evaluations. Therefore, some researchers have
focused on designing WSN testbeds, so that system and
protocol performance can be evaluated in a more realis-
tic setting. Recently research has also resulted in a variant
of WSN called WMSNs. WMSNs are composed of au-
dio/video sensing devices that are capable of retrieving and
transmitting multimedia content such as audio, video and
still images. WMSNs is a multidisciplinary research area
and the advancements made in hardware design, signal
processing techniques, coding theory, wireless network-
ing, and control theory have made WMSNs a reality.
Ordinary sensor nodes cannot capture and process mul-
timedia data. Therefore, manufactures and researchers pro-
vide solutions like Cyclops [1], Stargate [2] and Imote2
[2]. To evaluate WMSNs systems, researchers have started
to design and deploy WMSN testbeds, for same reasons
WSN testbeds have become popular. There exist a range
of WMSNs testbed, detailes can be found in [3].
We have organized this paper as follows. State-of-the-
art hardware for WMSNs is analyzed in Section 2. Sec-
tion 3 presents a detailed description of WSNs testbeds.
Finally, we conclude this paper in Section 4.
∗Corresponding author: e-mail: farooq@itm.uni-luebeck.de
c
2012 NSP
Natural Sciences Publishing Cor.

30 M. O. Farooq et al : Wireless Sensor Networks Testbeds andState-of-the-Art Multimedia Sensor Nodes
2. State-of-the-Art WMSN Nodes
In this section, we briefly summarize the state-of-the-art in
WMSN nodes.
2.1. Stargate
Stargate [2] uses an Intel PXA-255 Xscale processor that
operates at 400 MHz. It has 64 MB of SDRAM and 32 MB
of flash storage. Its processor can execute advanced image
processing algorithms. It supports the embedded LINUX
(kernel 2.4.19) operating system. It supports two radios
i.e., the IEEE 802.11 and the IEEE 802.15.4.
2.2. Imote2
Imote2 [2] is an advanced platform especially designed for
sensor network applications requiring high CPU/DSP and
wireless link performance and reliability. The Imote2 con-
tains an Intel XScale processor, PXA271. It comes with
an IEEE 802.15.4 radio (TI CC2420) with an onboard an-
tenna. It has two basic sensor board interfaces, consisting
of two connectors. Moreover, it has an advanced sensor
board interface, consisting of two high density connectors
on the other side of the board. It has 256 KB SRAM, 32
MB Flash, and 32 MB SDRAM and an integrated 2.4 GHz
antenna.
2.3. CMUCam3
The CMUCam3 [4] is an ARM7TDMI based fully pro-
grammable embedded computer sensor. It is equipped with
a Philips LPC2106 processor that is connected to an Omni-
vision CMOS camera sensor module. Features of CMU-
Cam3 include: CIF resolution (352 x 288 pixels) RGB
color sensor, MMC flash slot with FAT16 driver support,
loads images into memory at 26 frames/sec. It contains
LUA (a lightweight scripting programming language) that
allows for rapid prototyping, software-based JPEG com-
pression and a basic image manipulation library. It con-
tains 64 KB RAM and 128 KB of ROM.
2.4. MeshEye
MeshEye [5] is a hybrid resolution smart camera node for
applications in distributed intelligence surveillance. Mesh-
Eye has a unique vision system i.e., a low resolution stereo
vision system continuously determines position, range, and
size of moving objects entering in its field of view. This in-
formation triggers a color camera module to acquire a high
resolution image of the object. It contains an ARM7TDMI-
ARM thumb processor. It is a 32 bit RISC architecture
that can operate at up to 47.92 MHz. MeshEye has 64
KB of SRAM and 256 KB of flash. It uses a CC2420
transceiver and provides an implementation of the IEEE
802.15.4 standard.
2.5. WiCa
WiCa [6] is a smart camera node with high performance
vision system. It uses two VGA camera modules (640 x
480) 24 bit color, which feed video to IC3D, which is a
dedicated parallel processor, running at 80 MHz. WiCa
uses an 8051 based Atmel AT89C51 host processor, which
runs at 24 MHz, through a 128 KB dual port RAM. For
large scale image processing it supports 10 Mbits of RAM.
It provides an implementation of the IEEE 802.15.4 stan-
dard.
2.6. Cyclops
Cyclops [1] is a small camera device developed for WM-
SNs. Cyclops can be interfaced with MICA2 and MICAz
nodes. The Cyclops hardware architecture consists of an
Agilent ADCM-1700 CMOS camera module, a Xilinx FPGA
and an 8 bit RISC ATMega128 micro-controller.
2.7. Summary
There exist a range of wireless multimedia sensing nodes.
For transmitting multimedia data over low bandwidth links
these nodes have capabilities to run advanced coding algo-
rithms. Furthermore, sufficient resources in terms of mem-
ory and processing capabilities are available to run ad-
vanced image processing algorithms.
3. Wireless Sensor Networks Testbeds
In this section, we elaborate on the state-of-the-art WSN
testbeds. We evaluate various existing WSN testbeds based
on metrics such as: number of deployed nodes, hetero-
geneity in terms of supported hardware and software, avail-
ability to the general public, and deployment scale.
3.1. WISBED
The WISEBED [12] is a large-scale WSN testbed and it is
a joint effort of nine European Universities and Research
Institutes. The main goals of WISEBED are: heterogeneous
WSN testbed, testbed virtualization, facilitate end users
and application through variety of interfaces, and unified
algorithmic and software environment.
WISEBED follows a hierarchical architecture that re-
volves around four main entities: wireless sensor nodes,
gateways, portal server,and overlay network. Wireless sen-
sor nodes comprise the WSN and they are at the lowest
level of the hierarchy. A set of wireless sensor nodes is
connected to a gateway that allows access to the attached
sensor nodes. The gateways are connected to a portal server
c
2012 NSP
Natural Sciences Publishing Cor.

Appl. Math. Inf. Sci. 6, No. 1, 29-33 (2012) / www.naturalspublishing.com/Journals.asp 31
Table 1 WISEBED Summary
University/Institute Total Nodes Node Type Micro-processor Radio RAM/FLASH
University of Luebeck 162 iSense [7] Jennic 32-bit (JN5139R1 and JN5148) IEEE 802.15.4 128KB/512KB
TelosB [8] Texas Instruments 16-bit (TI MSP 43) IEEE 802.15.4 10KB/1MB
Freie University 100 DES-node [9] AMD Geode LX800 and ARM7 IEEE 802.11a/b/g 98KB/512KB
Braunschweig Institute
of Technology 30 iSense [7] Jennic 32-bit (JN5139R1) IEEE 802.15.4 128KB/512KB
Computer Technology
Institute 154 iSense [7] Jennic 32-bit (JN5139R1 and JN5148) IEEE 802.15.4 128KB/512KB
TelosB [8] Texas Instruments 16-bit (TI MSP 43) IEEE 802.15.4 10KB/1MB
Technical University
Catalonia 12 iSense [7] Jennic 32-bit (JN5139R1) IEEE 802.15.4 128KB/512KB
University of Bern 47 MSB-430 [10] (TIMSP430F1612) Chipcon CC1020 5KB/55KB
TelosB [8] Texas Instruments 16-bit (TI MSP 43) IEEE 802.15.4 10KB/1MB
University of Geneva 50 iSense [7] Jennic 32-bit (JN5139R1) IEEE 802.15.4 128KB/512KB
Delft University of
Technology 140 G-node [11] Texas Instruments (MSP430F2418) TI CC1101 8KB/8Mb
T-node [11] Atmel Atmega 128L TI CC1101 4KB/4Mb
Tmote Sky [2] Texas Instruments (MSP430F2418) TI CC2420 10KB/8Mb
Lancaster University 16 TelosB [8] Texas Instruments 16-bit (TI MSP 43) IEEE 802.15.4 10KB/1MB
that not only manages the WSN but also enables user in-
teraction with the testbed. Each WISBED site maintains a
separate portal server. The portal servers at different WIS-
BED sites are interconnected to form an overlay network
that can provide virtual access to each connected testbed
located at geographically different locations. Portal servers
provide services to enable communication among different
portal servers so that the goal of testbed virtualization can
be fulfilled.
In order to accomplish the heterogeneous WSN testbed
goal, the testbed deployed at different sites contain wire-
less sensor nodes from different vendors that differ in ca-
pabilities and hardware. Furthermore, it is possible that
each testbed supports different software/firmware but the
middleware makes it possible to keep interoperability among
these testbeds.
The operating systems supported by the testbed are
Contiki [13], iSense [7], and TinyOS [14]. The WISEBED
testbed provides a C++ library called WISELIB, which
contains implementation of different WSN related algo-
rithms, moreover applications written using WISELIB can
be compiled to any of the supported operating systems. In
order to evaluate the performance of different algorithms
for WSN, the WISEBED testbed provides a WSN simula-
tor called Shawn.
WISEBED is an open testbed. A user can access the
testbed by creating an account at any of the WISEBED’s
site, afterwards the user can freely use the WSN testbed.
To access the testbed, WISEBED provides three methods:
a web-based API, web-based clients, and desktop-based
clients. Hence, user friendliness is achieved through sup-
porting different hardware/software platforms along with a
range of methods to access the testbed. Table 1 summaries
WISEBED.
3.2. SensLAB
SensLAB [15] is a very large-scale WSN testbed with a
total of 1000 nodes deployed at four sites in the France.
The main goal of SensLAB is to offer an accurate and effi-
cient scientific tool to help in the design and development
of large scale WSN. SensLAB is currently deployed at: (i)
INRIA Grenoble, (ii)INRIA Lille, (iii)INRIA Rennes,
and (iv)University of Strasbourg.
The hardware used in the SensLAB testbed includes
different versions of WSN430 nodes [15]. It includes WSN430
heart rate daughter board, WSN430 GPS-Accelerometer
daughter board, WSN430 testbed daughter board, WSN430
daughter board, WSN430 daughter board, WSN430 strain
daughter board, WSN430 Bluetooth daughter board, and
WSN430 motion capture daughter board.
From the software perspective SensLAB provides sup-
port for three operating systems: Contiki [13], FreeRTOS
[16], and TinyOS [14]. SensLAB provides a software pro-
gramming library that includes implementation of various
OS free Medium Access Control (MAC) protocols i.e.,
Carrier Sense Multiple Access (CSMA), Time Division
Multiple Access (TDMA), and XMAC. Furthermore, it
provides an implementation of a localization protocol and
a gradient based routing algorithm. SensLAB provides sup-
port for WSN and embedded system simulations in the
form of WSNet [15], and Wsim [15]. SensLAB is open
to researchers of the host institutes and outside users need
to request permission in order to use the testbed. SensLAB
provides a web-based system to access the testbed.
3.3. MoteLab
MoteLab [17] is a WSN testbed developed at the Electrical
and Computer Engineering department of Harvard Univer-
sity. It is a public testbed, users can run their WSN appli-
c
2012 NSP
Natural Sciences Publishing Cor.

32 M. O. Farooq et al : Wireless Sensor Networks Testbeds andState-of-the-Art Multimedia Sensor Nodes
Table 2 WSN Testbeds Comaprison
Testbed No. of Nodes Hardware
Heterogeneity Software Heterogeneity Availability Deployment Scale
WISEBED
[12] 711 Yes
Supports multiple operating
systems, provides
implementation of network
simulator, and software API
library for application
development
Public Developed in multiple
countries of Europe
SensLAB [15] 1000 No
Supports multiple operating
systems, provides
implementation of network
simulator, and a software
programming library
Through
request Four locations in
France
moteLab [17] 190 Only Tmote Only TinyOS Public Maxwell Dworkin
Laboratory, Harvard
University
CitySense [18] 100 No Linux-based Public City-wide
Sensei [19] 20 Only TelosB TinyOS and Contiki Public Lab. Level deployment
at Uppsala University
cations using a web-based interface. Registered users can
schedule their applications on wireless sensor nodes, up-
load binary images, and visualize wireless sensor nodes
output through the web-based interface. The testbed stores
the output of wireless sensor nodes in a database and the
detailed output is presented to the user once the job is com-
pleted.
The design of the MoteLab testbed revolves around
the server, a MySQL database backend, a web interface,
a DBLogger, and a Job Daemon. The server hosts the web
interface, different databases, and provides connectivity to
the wireless sensor nodes. The web interface enables the
registered users to create a new job i.e., to upload binary
images, and to assign binary images to different wireless
sensor nodes. Moreover, the web interface can be used
to modify or delete already created jobs. The DBLogger
stores the messages corresponding to the active job(s) and
on the completion of the job(s) the messages are presented
to the user. The job daemon is responsible for reprogram-
ming the nodes, killing processes pertaining to finished
jobs, and allocating and de-allocating other system com-
ponents.
Moreover, the MoteLab testbed imposes a quota on
each registered user in terms of total duration of pending
jobs i.e., if the user quota allows 40 minutes of testbed us-
age then the user cannot run jobs that exceed the 40 min-
utes duration.
MoteLab only supports the TinyOS operating system,
therefore users need to know the NesC programming lan-
guage. Secondly, the MoteLab team does not provide a
software-based API to facilitate the application develop-
ment process on the MoteLab testbed.
The MoteLab testbed consists of 190 Tmote Sky [2]
wireless sensor nodes. The wireless sensor node specifica-
tions are a TI MSP430 processor, 10 KB RAM, 1Mb flash,
and Chipcon CC2420 radio. Each node is also connected
to the Ethernet.
3.4. CitySense
CitySense [18] is an open wireless mesh and sensor net-
working testbed that spans the entire city of Cambridge,
Massachusetts, USA. The goals of CitySense include: an
evaluation platform for WSN applications, support for new
mesh routing algorithms, sensor networking at the urban-
scale, new distributed algorithms for in-network data pro-
cessing and aggregation, novel programming abstractions,
and to enable users to reprogram the nodes using the Inter-
net.The key features of CitySense are: city-wide deploy-
ment and monitoring of the physical world through sen-
sors. The CitySense aim was to deploy a city-wide wire-
less mesh and sensor network testbed. Therefore, it was
not feasible to use sensor nodes that communicate at a
range of less than 100 meters (primarily due to the cost
factor). Furthermore, the CitySense team envisions that
future WSN applications will demand high data rate and
complex processing at the nodes. Therefore, they decided
to develop sensor nodes that can communicate using the
IEEE 802.11 standard. This meets the high data rate re-
quirements and reduces the cost of deployment.
The CitySense testbed includes wireless sensors nodes,
wire-line gateways, back-end servers, and a web-based in-
terface. The wire-line gateways link the wireless mesh to
the Internet and to the back-end servers. The server hosts
databases pertaining to the testbed, registered users, and
users experimental data. Furthermore, the web interface
is hosted on the server that enables users interaction with
c
2012 NSP
Natural Sciences Publishing Cor.

Appl. Math. Inf. Sci. 6, No. 1, 29-33 (2012) / www.naturalspublishing.com/Journals.asp 33
the testbed. The wireless nodes are mounted on streetlights
and are powered through these streetlights as well.
3.5. Sensei
Sensei [19] is a nomadic WSN testbed developed at Up-
psala University. The distinguishing feature of Sensei is
that it provides support for mobile nodes and its nomadic
nature makes it possible to evaluate WSN application per-
formance in a range of environments.
Wireless sensor nodes, sensor hosts, site manager, and
monitors constitute the design of the Sensei testbed. A
group of sensor nodes is connected to a sensor host and
there are a number of sensor hosts present in the testbed.
Wireless sensor nodes connect to the sensor host through
the USB interface. Since the testbed supports mobile nodes,
sensor hosts keep track of mobile nodes and inform the
site manager about the location of these nodes periodically.
Communication among sensor hosts takes place using the
IEEE802.11 standard. For communication between sensor
hosts and the site manger, a control channel is setup using
the IEEE802.11 standard. The site manager controls the
WSN testbed. The site manger has a wire-line connection
with the monitor. The site manager acts as the gateway
to the testbed and logs events. The monitor presents the
output of the sensor nodes along with other control infor-
mation to the users. A Java-based desktop client is used to
enable users to interact with the testbed.
The testbed uses a Linux based Asus WL-500G wire-
less access point as sensor hosts and TelosB [8] sensor
nodes with CC2420 radio transceiver. The testbed provides
support for both TinyOS and Contiki operating systems.
4. Conclusion
There exist a range of powerful multimedia sensing nodes.
A lot of effort was made in past few years in setting up
large, shared, heterogeneous WSN testbeds. Little work
yet to integrate these new multimedia sensors into these
testbeds. The latter is necessary as new nodal capabili-
ties require new protocols and pose challenges: Quality
of Service, transmitting multimedia data over low band-
width links, and memory management to name but a few.
There is an opportunity for new types of applications that
require new information coding techniques, network re-
source management, and algorithms and protocols to trans-
fer multimedia data over low bandwidth links.
References
[1] M. Rahimi, R. Baer, O. I. Iroezi, J. C. Garcia, J. Warrior,
E. Estrin, and M. Srivastav. Cyclops: in situ Image Sensing
and Interpretation in Wireless Sensor Networks. In the proc.
of the 3rd International Conference on Embedded Network
Sensor Systems, 2005.
[2] MEMSIC. [online], http://www.memsic.com.
[3] M. O. Farooq and T. Kunz. Wireless Multimedia Sensor
Networks Testbeds and State-of-the-Art Hardware: A Sur-
vey. In FGIT-FGCN (1), pages 1–14, 2011.
[4] A. Rowe, A. Goode, G. Dhiraj, and N. Illah. CMUcam3: An
Open Programmable Embedded Vision Sensor. Technical
report, Carnegie Mellon Robotics Institute, 2007.
[5] S. Hengstler, D. Prashanth, S. Fong, and H. Aghajan. Mesh-
Eye: A Hybrid-Resolution Smart Camera Mote for Applica-
tion in Distributed Intelligent Surveillance. In the proc. of
the 6th International Symposium on Information Processing
in Sensor Networks, 2007.
[6] R. Kleihorst, A. Abbo, B. Schueler, and A. Danilin. Camera
mote with High-Performance Parallel Processor for Real-
Time Frame-based Video Processing. In the proc. of the
IEEE Conference on Advanced Video and Signal based
Surveillance, 2007.
[7] iSense. [online], http://www.coalesenses.com/
index.php?page=software-system.
[8] TelosB. [online], http://www.willow.co.uk/
html/telosb_mote_platform.html.
[9] DES-nodes. [online], http://www.
des-testbed.net/sites/default/
files/DES-Testbed-Information_
Flyer-June-2010.pdf.
[10] Modular-Sensor-Board-430. [online], http://cst.mi.
fu-berlin.de/Projects/ScatterWeb/mod_
MSB-430.html.
[11] SWONet-Technologies. [online], http://www.
sownet.nl/index.php/products.
[12] I. Chatzigiannakis, C. Koninis, G. Mylonas, S. Fischer, and
D. Pfister. WISEBED: an Open Large-scale Wireless Sensor
Network Testbed. In the proc. of the 1st International Con-
ference on Sensor Network Applications, Experimentation,
and Logistics, 2009.
[13] A. Dunkels, B. Gronvall, and T. Vogit. Contiki A
Lightweight and Flexible Operating System for Tiny Net-
worked Sensors. In the proc. of the 29th Annual Interna-
tional Conference on Local Computer Networks, 2004.
[14] P. Levis, S. Madden, J. Polastre, R. Szewcyzk, K. White-
house, A. Woo, D. Gay, J. Hill, M. Welsh, E. Brewer,
and D. Culler. TinyOS: An Operating System for
Sensor Networks, http://dx.doi.org/10.1007/
3-540-27139-2_7.
[15] Senslab. [online], http://www.senslab.info.
[16] FreeRTOS. [online], http://www.freertos.org.
[17] G.W. Allen, P. Swieskowsli, and M. Welsh. MoteLab: A
Wireless Sensor Network Testbed. In the proc. of the 4th
International Symposium on Information Processing in Sen-
sor Networks, 2005.
[18] G.W. Allen, P. Swieskowsli, and M. Welsh. CitySense: An
Urban-Scale Wireless Sensor Network and Testbed. In the
proc. of the IEEE Conference on Technologies for Homeland
Security, 2008.
[19] O. Rensfelt, F. Hermans, C. Ferm, P. Gunningberg, and L.A.
Larzon. Sensei-UU: A Nomadic Sensor Network Testbed
Supporting Mobile Nodes. In the proc. of the 4th ACM In-
ternational Workshop on Wireless Networks Testbeds, Ex-
perimental Evaluation, and Characterization, 2010.
c
2012 NSP
Natural Sciences Publishing Cor.